MakerBot Print is our newest print-prepration software, which supports native CAD files and STL assemblies,
allows you to interact with all your printers via the Cloud, and many other exciting new features.

Download Now

Hey! This thing is still a Work in Progress. Files, instructions, and other stuff might change!

Lyman / Mulier Filament Extruder V5

by hlyman Jul 1, 2014
Download All Files

Thing Apps Enabled

Please Login to Comment

"link the motor velocities so that the puller motor is a percentage of the extruder RPM - complete"
How can i disable this feature in the firmware?

Hi, in summary, I need overall help in changing the parameters of "Lyman / Mulier Filament Extruder V5" firmware, Straight forward steps given in the manual do not apply here(details mentioned below). I want 75 to 100 rpm of extruder, reasonable fast puller and independent control of both extruder and puller motors. Plz see below for details.

I spent almost 500$ on Nema-42 motor, 20:1 reduction gear box for driving 445mm long 20mm dia extruder screw. Installed Lyman / Mulier Filament Extruder V5 firmware on Mega-2560, everything runs fine on stock settings. I want to modify my setup, since i am using bigger screw and bigger motors, hence i want to increase the both stepper motor speeds to achieve reasonable production of filament. My problems are:
1-If change the speed of extruder to like 30RPM, the stepper motor does not run smoothly, checked on oscilloscope the clock pulses were not regular, the reason i found is that for 30RPM at 1/16 step,the stepper driver clock freq. is gets too high, more than 10kHz and i think Marlin can not handle clock of more than 10kHz. So to solve this, i some how tried to limit the clock frequency to 10kHz and increased the speed of my motor by changing the step size from 1/16 to 1/4. I got my desired max rpm of 75, but I was forced to change acceleration from default 200 to 500m/sec.sq. if not changed, then, on scope i saw a lot much variation on the stepper drive clock pulses which make the stepper very unsmooth, the bad thing is that the acceleration now is very fast, and motor may skip or not even spin under loaded conditions.
2-Another problem, if I increase the puller motor speed by rotating the knob on LCD, then a time would come when the clock pulses at both stepper motor drivers would vary in frequency. i mean the clock pulses would not hold still on oscilloscope screen rather fluctuate slightly, this cause slight rpm variation in stepper motors (not suitable for consistant filament dia.) but if the extruder RPM is set a bit high like 40RPM and the puller feed rate is kept low, like 12, then this problem is eliminated, and i dont understand y, just a lot of hit and try...
3-After a lot of fuss and by changing few values on hit and trial basis, like increasing "max_extruder_rpm", etc., i managed to get around the above mentioned problem, now when i thought everything was good and final, i discovered that if i change the extruder motor RPM, the puller motor speed is also changed but the new speed is not reflected on the screen. I want independent operation of both extruder and puller stepper motors. Kindly help, i am not very good at programming, can any one tell me which line of the firmware i will have to edit to solve this problem?

The firmware has very good features, and as per my personal opinion this V5 extruder is the best of all the versions, but it needs some debugging. some unsolved issues are present when we change the parameters to our requirements.

What can use instead of the Urethane Driver ?

Thank you for this thing! Thumbs up to you!

  • Posted with 3D Geeks Thingiverse Browser App

Dear friend it is possible to send link arduino code

Good morning.
To improve this project, you only need to chamfer the ears of the part "U " which holds the top roller of the tractor so that it does not get into the mounting screws of the motor support with the roller holder.

Very good this project congratulations. What I did add 10 pulleys to work with the filaments before the tractor.

Where can I find the instructions?

It is one of the downloadable files.

Hello Everyone,

I'm currently looking for some advices to make the same kind of machines but for PET (as you are saying this works only for ABS), what are the main features to change ?

Thank you very much for your help and support

qual a lista de componentes ?

Anybody know where the best place to source large amounts of ABS pellets for cheap? Just trying to save money on large prints; it's almost not worth the effort to build this when you can find cheap ABS spools off ebay.

Good job, was looking at buying one of theses but they were way over priced, thanks for the design

Hi hlyman, thats a great job. Thank you. I want build it. Can you send me auger bit dimentions. I upload a picture. I cant find it anymore. Please help.
Thank you again.

The auger drawing in the picture is a drawing of a custom auger I never used. I is to expensive.

The auger is a 5/8" x 17" and is cut to length to fit the assembly. I have used several different augers as they are not all the same diameter. Sometimes I have had to ream the 1/2" nipple for the auger to fit as it has a weld joint on the ID. I try to keep the end of the auger about 1/2" from the nozzle. I have used Irwin, Bosh and others.

Hylman this has my attention, I am seriously in need of such a device , nicely done. Just a quick question. Whats material gives best result in the filament making process.

I am inspired.

Only ABS or very similar characteristics resin is practically extrudable with this setup.

I only used ABS. Check out my V6 extruder.

1 question do you use a screen before the dye or dose it extruder with outit

Hi, everybody. I have a nema 23, steppers, arduino and ramps, also an auger bit and the metal parts. I assembled my extruder entirely. But I don't have a sensor neither I'll use the spooler. I just want the ramps/arduino/driver to control the extruder motor and PID.

Taking this setting in account, I have to change the software mackerel somewhere? What do I have to do? Someone could please kill this doubt?

Thanks by advance good people!

Is the nema motor driven by the A4988 drivers? As far as I know your nema requires 2.8A. At the same time, the drivers of the A4988 issue no more than 2A. It works!? I want to replace Arduino with MKS Base. Do you think this is possible !?

He has v6 available, check that design. Far as using current and Steppers go, If the motor says 2.8A thats a MAX rating before the windings overheat. 2A is fine on a 2.8A stepper. More current = more holding torque, simple.

Great Work Sir....Can I make 1.75 mm filament easily with this machine or it only works for 3 mm filament...????

You can make any diameter you wish. I make mine 2.65mm. Just drill the correct hole in the nozzle for the correct diameter.
You may want to look at my V6 extruder, it works just as well at a lot less cost.

Thanks... I'm interested to make this or v6....

Is it possible to use a REPRAP_DISCOUNT_FULL_GRAPHIC_SMART_CONTROLLER? I get a compile error "fanspeed was not declared in this scope". Has anyone used this graphics controller?

ı have same error. I tryin to fix but ı can't.

I haven't got back on the project. But instead of spending the time on modifying the code I just got the display identified in the parts list.

I am forwarding your comment to Filip.

You would have to make some coding changes to get the graphic smart controller LCD to display the right data. The existing code that feeds the display is still unmodified from Marlin. My guess is that I removed fanspeed from the rest of the code, since its not used in the filament extruder, so now when you enable the code for the graphical LCD, it still expects fanspeed to be around and complains. To make the change, you would need to find the code that handles outputs to the graphic LCD and make changes to model it after the character LCD code (make it show the same variables, using the graphical lcd constructs).

I just ran into this same problem. Have y'all corrected it in some patch... if so where is it?

If not, I'll see about going thru and taking care of it ASAP; and, will try to remember to post it.

In the meantime, thank y'all SOOOOOO much for sharing this project! I'm actually using it in part of a research project for one of my last classes before I grad with my BS!

Hi, I'm building one as we speak but i have a little difficulty understanding the wiring plans i can't figure it out completely can someone help me with it. and is there someone who can help with the build of the sensor?

regards Danny

This is all great but where do I get the PLA pellets?

It would be interesting to know if you succeeded extruding PLA with Mr. Lymans extruder. Because I had difficulties with PLA and this setup. For ABS it works well but PLA resins have different properties and apparently need more intricate cooling sequence.
would really appreciate comments on PLA extrusion issue.

It would be interesting to know if you succeeded extruding PLA with Mr. Lymans extruder. Because I had difficulties with PLA and this setup. For ABS it works well but PLA resins have different properties and apparently need more intricate cooling sequence.
would really appreciate comments on PLA extrusion issue.

A University in Colorado is trying to extrude flexible plastic with my V4 extruder. The problem they were having was getting it to cool fast enough to enter the winder. I suggested to them to move the winder farther from the extruder nozzle, add a water bath. I have NOT heard if they have had success.

If they ever get back in touch with you.... cannibalizing a humidifier might do the trick. Use it in conjunction with the fan, and they could cool it a ton faster.

Hello sir,
Is it possible for making flexible material by this machine?
How to contact you?

Folks, I am not supporting any other Lyman Extruder now except V6 (Lyman Filament Extruder V6). I have three working one for white, one for black and the third for other colors. My last 1 kg spool tested at +/- .03mm. It is vertically mounted and the spool wind system is separate.

Hey Mate,

I know your not supporting the extruder, but my question relates to the winder, I don't have access to a Geared motor but I do have a spare nema23 stepper motor, how do i go about using this instead of the geared motor to wind the spool using the mackerel firmware

you may be able to hook it up to the arduino. Check with flipper. The firmware will need coding.

Comments deleted.

what would be the total price to make one now +-?

the auger drawn and pictured here is not available as I could not get and quote to make it.

Where can I get a plastic screw auger?

hardware store or ebay. 5/8" dia, read the manual. Check out my V6 Extruder.

Did you have any success with this. I have made the extruder and assembled. Would love to add this to it. Care to share your files?? :)

Is your interface working.
I will be interest now that my setup is working fine.

Yes I'm using it for some time now. It replaces the RAMPS and Arduino Mega board with a Touchscreen LCD and Arduino Due

Is there any way to get a copy of the programm

It's not a drop-in replacement. Rewrote everything for my own hardware like MAX31855, other motor drivers and a new TLS1401 based filament sensor (Teensy based)

Still working on it as it's not 100% bug free yet :-)

I convert my V5 to used it verticaly like version 6 and changed from 1.75 to 3 mm filament extrusion and I am very surprise of the quality of extrusion.

I published V6 Extruder yesterday.

Any update on the V6 version files? Maybe a website with download files for a fee? :D Cause I sure would pay for it since I am not smart enough to be a ingenious engineer!

Thanks you're awesome

The version 6 youtube video's I posted turned out to be an one-off special project for someone who wanted to run a special kickstarter. For now the V5 is still the latest lyman design as far as I know.

Thanks CFT, Thats good to know. I have all the parts to make the V5 but I have been waiting on new files for V6 before building the Lyman extruder but I guess I will start building the V5 then. Appreciate the comment, otherwise I would've been waiting an awful long time. lol

Hello! Thanks a lot for sharing your project. I'm new in the 3D printing world and I would like to ask you if it is possible to print all the pieces of the Filament Extruder V5 with a "bq Prusa i3 Hephestos ". Your answer it's a big help for me, thank you!

I printed all pieces on a standard i3 (20x20cm)

Do anyone know where can i get the new profiled auger ? i dont have access to CNC machines, is it available in some online store ?

First of all thanks for sharing your project and inspire others.

The project we have in mind is use a robotic arm, we would like to transform it in a huge 3D Printer. we have been thinking to adapt a Extruder Assembly subsystem with a nozzle of a 1 cm. that have arise many questions:

  • Is way for controlling the process of extrusion itself, knowing the volume of plastic extrude ?

    • Is possible to have retractions in the extruder ? The people who have done something similar usually make continues geometries spiral vase. make me wondering if is difficult to have a retraction in a system like this one.


  • How to recalculate the Geometry of the Screw?, because I must be required much more pressure to extrude 1 cm of material.

Seems like you're looking for something like this: http://www.herz-gmbh.com/index.php?id=204&svartpin=VJVXmwXsGPx6wEes&action=31 (6kg extrusion / hour)

Sorry, I am not smart enough to help you on your request.

I was wondering how you came up with the auger screw design.
Having access to CNC equipment, I am thinking of machining the screw for optimal output.

Hi guys,
I'm rookie in this, some one have the full material list and the step by step on assemble
thank you

The thing files section contains the manual (pdf), in it you will find the BOM, assembly and other instructions

I made it!!! It works nice!!! It took me about two months to find all details and assemble it toheather, but finnaly it works!! https://youtu.be/kGBzgypZBDo

Looks good! Did see however that the sensor data (filament width) on the LCD was not showing the correct data. It showed 0.01mm with an average of 0.10mm

Yeah I had some problems with the filament sensor in the beginning, but now it works good. The problem was with microcontroller, I programmed it incorrectly)))

hello .With respect i wanted to know what is it the material of the auger bit in this extruder and the other steel part .(the code)

Stl 30 - shaft collar is missing from the files.

It appears as though STL - 21 , the Motor mount base for the puller assembly are meant for a NEMA 17 stepper motor WITHOUT a gearbox. Do you have the STL file of the mount and coupler for a NEMA 17 WITH gearbox? The gearbox has a tighter bolt pattern than the nema 17 stepper.

Additionally, STL - 22 the shaft coupler, appears to be oversized, Compared to the geared NEMA 17 you suggest, my gearbox shaft diameter is ~7.7mm

Just checked and I think you are correct. However I did not like printed couplers so used a brass one and I did not like the puller motor mount so designed an other one, also not happy with that one but it has to do for the moment :)

Seems that version 6 is in de make. Arduino controlled parts are gone and setup is vertical and a filawinder style winder? No filament diameter sensor, no puller.

Where can I get a nozzle? BOM Page 16 Extruder Assembly Part #1

And I just got all the parts for the V5, minus a few screws I need to get yet and was gonna start assembling.

I was going to make an adjustment to the base because I am thinking of adding bolt holes to the legs and make a bit longer where the Auger wont need to be cut down on the shaft end.

I got a filament sensor board but also made the pcbs and got the parts to make my own. I want to try my hand at putting a board together.

I wonder if I should just wait and see if the V6 is better. I am sure the parts I have would work fine for it.

hmmm...decisions....decisions. lol

Comments deleted.

I want to make the filament puller, what parts do i print and what other parts do i need.

Have a look at the manual, it explains each part of the puller

I can't control the temperature and make it keep at 185 degree C. Is this problem related to the circuit or the thermistor? Give me a hand, please!

Page 55 of the manual "Heater Setup"

Thanks for your reply and help!
I have tried PID autotune for many times, but the temperature still raises up to 300 degree C.
I found that if the switch was turned on, the temperature will keep raising.
So I think that maybe it is the problem about circuit?
I am not sure about the connection from the D10 to power supply 12V+- in the the wiring diagram in page 26.

Combine that diagram with the one on page 46 maybe will make it more clearer? I do not understand your remark about the " D10 to power supply 12V" ??? D10 goes to the SSR and to the illuminated 12v switch.

I think the 12v switch is the key point of the problem. It ever cause smoking due to a short circuit. I am not exactly know how to connect the 12V switch with the wire from D10. Could you show me a picture about this part of circuit? Thank you for all your assistance.

Depends on your switch. You need to look at the datasheet for your switch. Most have 1 of the three pins in a different color and that normally is the gnd pin. The idea is that it's only illuminated when the heat is turned on. (page 8, Part 11)

I think I can finish the work! Thank you again for everything you’ve done.

Added commando "M701 - Clear Statistics" to Mackerel and included this command in the client. Allowing to send this commando before starting to record the log file.

If tests show no errors will be releasing it this week.

I created a new version of the mackerel software where the "smoothening" of the filament width measurements have been removed. This will allow for cleaner data to be analyzed as analyzing filtered data has not my preference. If others would like this version as well let me know and I will release it. Might make it an option that can be turned on/of via de config file.

Just a note: Without these routines (there were two) your calibration rod (1.57mm) measurements will fluctuate +/- 0.04mm. Measuring a piece of 3.00mm filament gave a +/-0.07mm fluctuation.

I released my mackerel version having support for the M701 command. Client will follow

Client version 1.2 includes the option to clear the stats (M701) before starting the log file. You can download it from here : http://www.clubfitter.eu/tiki-download_file.php?fileId=80

I've almost finished the completion of work
Thank Lyman

But I have a problem ABS pellets jam over and over
same times work 3 min , same time 1 min ,
i use ABS MAGNUM 3404

see :https://www.youtube.com/watch?v=JYhPfh6CnGQ
Are there any ideas to solve the problem?
sorry for my english

Put a fan to cool the Stepper drive extruderr and use heatsink, I had same problem.

the extruder driver is too cold ..

and i add fan and heatsink in the extruder driver ..

but it same issue ..

check this image 4 http://www.thingiverse.com/make:127686

when i remove all ABS and keep extruder working, it work 1 hour without stop !

thanks for your help ..

Lyman / Mulier Filament Extruder V5
by ahmedx
Comments deleted.

try a higher temperature. Are you sue you have the correct extruder motor and the stepstick set to the correct voltage?

i have step motor :

it same site Lyman he order from
check this image 3 http://www.thingiverse.com/make:127686

i tried with 200C to 240C it better but it still stop .

i tried change tune extruder to 6 rpm and it work ...
but it have same clicks and stop for 1 sec and Continue
see : https://youtu.be/X2Wk7Lri6fg

thanks for your help ..

Lyman / Mulier Filament Extruder V5
by ahmedx

Motor looks good.

Did you try to increase the vref on the stepstick? (http://reprap.org/wiki/Pololu_stepper_driver_board) If that did not help then you are just driving the motor to hard\fast.

If the motor is very hot then cooling the motor might also help a bit but lower extruder speeds might be the only solution.

Is it possible to put some parts source files for download?
I have a display PCB which is smaller and printed parts 18 and 19 are not fit.
If not, could you make a special stl files for my dimensions? Some displays from China on the market are the same as mine, it would be helpful for other users too. I like this design of display ped, but drawing it from scratch will be time consuming.

P.S. If somebody from Russia is looking for sensor boards, I'll have some number of them to spare in a month or two. I've ordered a batch of 10 boards from SeedStudio and all BOM positions x10. I can arrange delivery to Moscow or SPb only, your delivery for anywhere else. Sorry, if such ads are not welcome here, it's not a business, just don't know what to do with unneeded 7-8 sensors.

Have found display module which fits as is.

P.S. SeedStudio sent me 12 pcbs instead of 10, now I have 10 unneeded sensor boards :-).

Here is my latest video, this is with the double sensor setup.



check out photo number 5!


very nice!

currently at 15,000 mm and still going! hands free in auto!

Photo #6 75,000 mm - check it out! I am so pleased! 3 hours and 18 minutes so far of non stop sweet extrusion!

Lyman / Mulier Filament Extruder V5

photo 5: The min max values of that run are 1.45mm and 1.95mm (+/- 0.25mm), the current width in that screenshot has a lowest value of 1.65 and highest of 1.85 (+/- 0.1mm)

photo 6: In there the current width looks much better between 1.64 and 1.78 (+/- 0.07mm) but the min and max values of the run are 1.31mm and 1.99mm (+/- 0.34)

ya there was some swing when I first started the run, but once it settled down it worked very well, I was able to print a head for a hexapod last night, a 10 hour print, a very big improvement from before. It is not perfect but it is the best I have gotten.

Look good.
That mean I have to build two sensors now.

With this setup, it will be possible to do direct printing from pellets to finish parts.
The puller will become the extruding motor ( cold End ) on Bowden style.
The printer controller will ajust the speed of the puller (extruder) and the sensor will ajust the feeder motor speed.
Sorry for my english, I am french speaking.

would a clamato juice jug fit in the hopper hole ?? awesome looking setup!

Buying the motors is more or less the last thing I need to do to have the parts in hand.

It seems to me that the second shaft plays no role in the machine, so seems to me this motor could be used which is a bit cheaper on the extruder assembly?

From the same vendor it would also seem possible to buy this motor for the puller assembly?


Comments and remarks appreciated!

from omc-stepperonline.com
Nema17 : ($30.90) + Nema23 ($56.93)+Shipping to US ($35.10) = $122.93
from phidgets.com
Nema17 : ($42) + Nema23 ($72)+Shipping to US ($8.39) = $122.39

Agreed, but if you're not in the US it can make sense.

Do the motors seem appropriate?

I'm assembling the parts for the extruder now and have read through the comments several times. Can someone tell me if the all/some/part of the trouble with variations in the filament can be solved by the use of the expensive Filabot augur or perhaps the customized augur placed here?

FYI: My extruder makes me a 2 lb (150,000mm) spool of ABS filament in +/- 4 hours, running at 17 rpm. That is 10 to 11mm per second. My tolerance is mostly +/- .10mm with occasional blips between +/- .15 to .25mm. I extrude a spool at least twice a week and I am printing everyday on two printers that are each equipped with Filip's Sensor, which does wonders for handling the variations. I have tried the screened nozzle with no luck. Further I extrude 2.6mm Ave filament, so a 0.10 difference is 0.038 wherein a 1.75 filament it would be 0.057 difference. So put your second sensor on your printer and try that with Filip's Marlin coding. Also I use Enpol 8433 ABS pellets.

I have tried without the screen and it seems you are correct, it is easier to control without a screen. Another step closer, thanks

My posts prolly make it sound worse than it is, I do print a lot of parts from the filament I make. My problem is the Hexapods I agreed to make for the Boy Scouts, there are two parts that are 10 hour prints and 5 parts that are 6 hour prints and it is hard to not have a fail over a 10 hour print. To get 9 hours in and have a fat spot jam my printer or a thin spot cause a delaminate not only sucks but eats into a very tight printing schedule. Leg parts that take 1 1/2 hours, to have 1 in 10 fail is not a game breaker. I do not mean to dish your extruder, it is amazing. I just want to have almost perfect filament for when I start in on the 6 and 10 hour prints. I am using up over one kg of filament made on my V5 every day. I have always used a screen due to the crap in my first runs, but the screens never plug anymore, it is interesting to hear you have had bad luck with them, I will try removing it and see how it works. Maybe I should just break down and install Filip sensors on my three printers to get me through this Boy Scout project. My understanding is they speed up extrusion when the filament is thin and slow it down when it is thick? Is this an easy add to the firmware?

Could you try the client software during some of those production runs and send me the logfile? Would really like to see the behavior of the original lyman.

With regards to .1 being 0.038 for 2.6mm and 0.057 for 1.75mm I beg to differ. 0.1 for 2.6 Is a ~8% value difference (or 0.08) and for 1.75 it is ~ 11% (0.11) However with 0.1 on 1.75 hotend blocking is very likely. The 1.75 should therefor be extruded at 1.65 making the volume fluctuation ~13%

Hi CFTechno,

I am still getting some swings every 8,000 or 10,000 mm or so, they are well within my limits but it would still be nice to get rid of them. I can hear the extruder changing RPM but it is too hard to see a correlation if there is one. So I don't know if these swings are caused by the extruder or the filament sensor. I am wondering how hard it would be to add extruder RPM to the graph software so I could see if there is any relationship between them. My extruder RPM runs between 3 and 8 RPM, it would be nice to have it show under the filament on the graph.

Yes, adding both extruder and puller speed in that same graph as an option is a nice addition. Will do so and post a message when ready for download.

Thank you very much : )

Mackerel Client software for Windows. version 1.1.0 http://www.clubfitter.eu/tiki-download_file.php?fileId=79

Wow, thank you so much!

I did not even ask for the puller speed to be added, but as usual you are much more knowledgeable about this than me, thank goodness ;-P

The puller speed has been the most helpful, I can now see the delay in response and make much more educated changes to the PID numbers. The delay inherent in the system does not seem to be as much of a problem has I thought before. The puller speed follows changes in filament size much faster than I thought they were, so now I am no longer blaming that delay for changes that are obviously not being caused by it, very helpful.

I can't thank you enough : )

check out my progress up top ;-P

Here is my attempt at a profile die, so far so good but it still needs some work:

new sensor case - http://www.thingiverse.com/thing:736633

video - https://www.youtube.com/watch?v=B3xawERwFqM&feature=youtu.be

I am using a modified firmware from Filip that has the sensor controlling extruder RPM rather than feed roller speed.

Lyman V5 profile die experiment

Thanks you!

Is the puller really running 1 mm/s?? 1kg@1.75mm would then take 4-5 days?.....

It says 1 on the screen, it takes about 12 hours to fill a spool at this rate, 40,000 mm to fill a spool. 11.1 hours by math

Maybe it is reading cm?

1000mm in a meter so 111,11 hours. I'm sure it's not cm but mm in the display. Is your step setting for the puller correct? [Filip] had the default set for 100:1 geared puller. I changed that in my version but maybe you still have the old values?

you can measure the 60seconds length and then compare that to the FR value, just to be sure.


1000 mm at 1 mm per second at 60 seconds per minute = 1000/60=16.6 minutes per 1000mm or 16.6 minutes per meter

40,000 mm fills my 1 kg spool so 40 X 16.6 minutes = 640 minute per spool or 640 minutes/60 minutes per hour = 10.6 hours

what am I missing?

And I have filled a dozen spools and can confirm it takes about 11 hours to fill a 1 kg spool. The first six or so spools i filled took longer as my screen would start to plug and slow things down. Now it seems all the crap has been worked out and I have made 4 spools and I see no slow down.

1kg @ 1.75mm equals 400meters of filament. With 1000 millimeters per meter would make it not 40,000mm but 400,000mm

But if your spools are netto 1kg then that 1mm/s federate is wrong in your lcd display. That could be the steps settings for the puller that are wrong (or maybe your microsteps setting on the ramps). Would explain why some time ago you need the minimum puller value to be lower then what was then the minimum in the original version.

so it is close to 1 cm then. I have gotten used to it now ;-P

I don't want to be rude but for a $600 extruder producing 1kg in >10hours you could also buy two filastruders and have the same filament tolerance but twice the performance.. :(

Tried teflon yesterday and today, same story, With the Teflon the failure is later but still fails. Tried different distances, different land lengths... :(

Did notice that with the Teflon some elastic stretching and recovery is going on directly after the die, so had to place the sensor after the puller to measure the real filament width.

Yep, exactly.
Even with filastruder+filawinder you still getting better deal and less headache then building one of those.

I did a lot of research and found unhappy owners of every extruder out there. Had I known before I started this project I would have purchased an old used commercial extruder, there are lots of them around my area. But from what I read here and elsewhere I was under the impression that this extruder would produce filament I could use. I passed up a sweet little lab extruder I could have had for $2500. I have not given up here, I can now produce filament that never goes over the hole size of my die so I am half way there. Besides it takes about 30 hours to use up a kg of filament so I really have no need to make it faster than 10 hours even running three printers. I'm not sure why yours is not working, as you can see in my video mine works fine. The only time mine fails is when the blob in front of the die gets so big it curls and globs into a mass. Maybe it is the plastic, I am using Styron 347EZ ABS.

I am not limited to 10 hours, that is just the speed I am testing at, it may work better at higher speeds since it would have less time to cool between the extruder head and the die. When/if I get the bugs worked out I will try to speed it up, double the speed I am currently running may well be possible.

This was a typical reply when I emailed and asked if I would be able to return an extruder if it would not make filament within the tolerances they advertise:


Many thanks for your mail.

The standard Noztek Pro has been tested using ABS MFI-22 and produces filament with an accurate degree of tolerance, we can achieve using the 1.75mm nozzle (+ .04/ – .04), but you must take into account the ambient room temperature and the type of polymer that you are extruding.

However, the Noztek Touch is a more advanced unit, with not only motor speed control, but a more accurate thermocoulper. The big advantage her is that you have two variables to play with, temperature and motor speed. It is also powered using an Arduino processor, so that the code can be upgraded.

I hope that helps.

Kind regards

Steven Forster

As you can see, no offer to let me return it for not making good filament.

I ran another test this morning just to make sure I could reproduce the success I had yesterday, and I can.

But I also discovered why I got so much swing in filament diameter before. The swing in extruder RPM required to keep the profile die full but not over full is amazing. It will run smooth at 11.40 RPM for a while then start to dry up so I increase to 12.50 then 13.00 then 13.50 then all the way to 15 RPM! It runs great for awhile then it is getting to much and then maybe 13.40 RPM is right, then 4 minutes later 10.50 RPM is correct, and it goes like this forever.

What this tells me is that the amount of plastic being extruded at any given RPM varies greatly, it is not even close to a constant. It is no wonder that it is close to impossible for the old method to maintain a constant filament diameter. The big programming change I have asked Filip for is that the sensor now control Extruder RPM and not Spooler speed, the profile die needs a close to constant supply of plastic. One nice result is that it is an instant need for change, it does not have the delay of before when we did not get a reading until several seconds after the filament had already hardened.

CFTechno, I hate to have two people working on the same code changes but I am wondering if this is something you could do. That is change it so the sensor controls the extruder RPM rather than spooler speed. Also it would be great to have two sensors running at the same time, the second sensor would be there just to give a finished filament diameter reading, it would control nothing other than your graph software and the LCD Diameter and Min and Max readings. I am wondering if Filip would be modifying code that would not include your M700 stuff and require you to rewrite it anyway. I would really like to hear your thoughts on this, after thinking about it, if this is something you are capable of you might be the better choice. I really don't know because I do not know if this is something you could do and I don't know if Filip can modify your code. I would let Filip know so we are not wasting his time if this is something you want to tackle. I hope I am not being an ass here, I just want this to come together in the best and quickest way.

Now that I think about it, it might be nice if the second sensor could still control the spooler speed. Since the profile die will be always supplying a constant amount of plastic, the spooler control may be useful. It is an option I hate to lose I guess ;-P

enough of my dribble...

I want to work collaboratively on this with you. In order for me to do the firmware changes and feel confident that they work I will need to have the same extruder setup on my end. Some additional drawings, description, or a picture would be helpful for me to recreate the setup. Before I devote the time or introduce a big change, I need to feel sure this leads to improvement.

You wondered if I would make the changes and I agreed. Now you are asking Frans to make the changes. Please show some respect to our time. Like I said, I am happy to work with you collaboratively on this effort, but if I feel like I am being ordered around I am out.

Lets start with the PID change to Extruder RPM first and see how that works before adding the additional sensor. I've pulled in CFTechno's code with the graph changes and made the PID changes already locally. I now need a way to test them.

Hi Flipper,

I am sorry I don't mean to push, I am semi retired and have a lot of time on my hands so I get antsy. And I wanted the graph software to work, please forgive an old man.

My Lyman V5 is exactly like the manual, every part is exactly the part number listed in the manual. The profile die has zero effect on the extruder itself, so if your extruder RPM is now controlled by the sensor we should be good. I can test run it and take a video so you have the confidence to add another sensor, if that is ok.

How sure are you that it's because of the die and not because of now having a different molten plastic reservoir length (distance between de auger tip and die wall)? I can get my extruder setup such that it will also generate a predictable extrusion change without using a special die. http://i19.servimg.com/u/f19/13/37/93/48/lymanw10.png

Manual intervention doesn't work for me as the extrusion speed is to fast, that's why I'm updating my version of mackerel and the client to do the analyses and then as soon as the software has calculated a pattern it takes over the control of the extruder motor to counteract the pattern.

I am sure because the extruder is not changed, the die is feed about 1" from the extruder head. The only thing connecting them is extruded molten plastic, and I can watch this connection change size, that is what I have been watching to see the needed change in extruder RPM. I have basically copied a system used by commercial filament extruders.

Sorry, don't understand the answer. The extruder head is the tip of the auger?

The extruder head is the brass fitting on the end with the hole in it that the plastic comes out of. The auger is the screw that spins inside the tube that connects the head to the hopper.

Tried it, several times with different setups, sizes etc heated not heated etc.

Not heated : The plastic gets to cold to pull as soon as it touches the die. Increased nozzle temp. to 210C did not help (normally run at 195C)

Heated : Slightly better but the surface of the plastic still cools, as soon as there is a just a very small blob there is not enough pressure from behind to get it thru the die.

I'll keep trying...

Uploaded two pictures at http://www.thingiverse.com/make:120597

The first picture is the die taped to the fan mount, this is how I did my first tests, the second picture is my first attempt at a new sensor mount that holds the disk.

I am going to redesign the mount in a few minutes to try to over come some problems I am having with plastic attaching itself to the sensor case and not allowing enough room for a proper size blob.

Lyman / Mulier Filament Extruder V5

I have not had cooling problems, and my tension motor pulls it through the die. My die is only about 5 mm thick and about 25 mm in diameter. I do not heat my die, I would not use metal for a die. Mine is Teflon.

I will try to post a picture this afternoon.

Tried several other setups but the problem is always the same, as soon as the blob is larger then the die the plastic is already to cold to pull it thru the die. Might be that it works for 1.75mm but for 3mm for the moment I can not get it to work.

How far is your die from your extruder head? And what is your die made of?

2.5cm or around 1 inch, made from alu. Tried size 2.5, 3.0, 3.5, 4.0, 4.5 and 5.0mm, all with different entry sizes (tried none,6mm and 8mm), tried headed (195C,200C,205C,210C) and not heated, tried heated tunnel between the two (200C) Constructed a simple setup on the nozzle head with screws that allowed placing it more forward/backward is needed. It also allowed for easier taking off for threading the filament thru it.

As soon as the blob is larger then the die then depending on heated or not the filament will snap or pulled into a tin thread. The rest of the plastic will bulge up against the die.

The working procedure was : Setting the puller at a fixed speed and the extruder at the lowest speed possible. Thread the plastic thru the die and place the die. Then slowly speed up the extruder. The change in speed where the plastic still got thru the die and not was 0.1rpm.

Ya dump the Alu. get some Teflon, a metal die would need to be heated and lubed with hot water.

I just took a 4 minute video of my setup running, I am going to upload it to youtube and also upload the .stl files for my redesigned sensor case. As you will see it is just a modified version of the one supplied with the Lyman .stl's

I will supply a link as soon as I get them uploaded.

new sensor case http://www.thingiverse.com/thing:736633

video https://www.youtube.com/watch?v=B3xawERwFqM&feature=youtu.be

Lyman V5 profile die experiment

ahhhh...thank you! Now I understand what you are doing! Did you need to heat the die? How hard is it to get it thru the die on startup?

Very interesting idea!! Must definitely try that tomorrow!! Or better will try that tomorrow!

I stretch it out real thin and thread it like a needle, I have made the die sorta portable. I printed a sensor holder with a holder on the side for the die so I can slide it in from the side kinda. Filip sent me firmware to run extruder RPM and I spent the last three hours getting PID numbers that work. A word of advice, cover your sensor holder nearest the extruder with tin foil or the filament will stick to it and cause problems. I do not have it making perfect filament yet, but it never goes over the size of the hole in the die. I do not heat my die, make it thin enough so you do not have to heat it but thick enough to do the job. Mine is about 1/4" thick, the side closest the extruder is "V"ed very slightly. I need to email Filip, for some reason my "FR" changes sometimes, I think it happens when I am in manual mode.When I do it by hand it makes better filament than when I set it up to do it auto, not sure why, PID numbers maybe. Part of my problem may be that I have less room for my blob since the sensor is between the extruder nozzle and the die. It might take a few tries to get it going but once you do you will get the hang of it fast. Still a work in progress but headed in the right direction. For me the best "blob" in front of the die seems to be about 2.20 mm, any bigger and when it goes high it will cool too much and stick to the sensor holder, smaller than 2.20 mm and I get thin spots when it dips low.

For some reason I can't get pictures to download from my camera or I would post a picture. I have noticed funny crap with three printers and an extruder all hooked up at the same time, seems to mess with com ports bad.

Hope this helps.

Interesting, I will try.
Are you using groovemount or brass fitting for dye.

Humm, what is "groovemount"? My die drops into a grove kind of, more like a recessed or inverted saddle.

I am not using a groovemount, just a disk with a hole drilled in it.

I have a fix!

It works so damn good it scares me. But it will take some major changes in programming and some big but not major changes in the sensor.

I need to reprint some parts and make this workable, I will need Filip and or CFTeckno's help. But I have done this by cobbling so I know it works. I made filament that did not vary enough for the sensor to see a change.

I have some serious thinking to do, I will be back! ;-P

Looking forward to hear what you have found! What part of the software need to be changed, the PID?

Hi CFTechno, please read my above post. I am replying here because if your set up like me you get an email when someone reply's to a comment of yours and I want to make sure you see my post asap ;-P

I have gotten in contact with Filip and he has offered to change the software : )

I will still want your graph software to work and have asked Filip if he can make it happen. It may become necessary for your graph software to be modified.

In my test run I was doing everything by hand making manual changes and the results were nothing less than amazing. But after a while I got so excited I start making corrections backwards LOL

So I am guessing those that tried a profile die and failed did it differently than me ;-P

Once I have it working I will share everything with this forum, I will also upload any .stl's I make for it. I am 99% sure this is a game changer, I was able to extrude 50' of filament with little to no change in diameter. The first trial run was only about 10 feet and the sensor never left set point, on the second trial I ran about 50 feet ( so what 33 meters? ) and with a set point of 1.75 it went as low as 1.73 but never went over 1.75. And when it went low I think it was my fault, I will be printing modified parts today and Filip said the software changes would be easy and quick, I just don't know how quick ;-P And that was before I told him I still wanted to be able to use your graph software, I am not sure if he can make that happen and if he can, how much time it will add for him to make the necessary changes.

I will keep you updated as soon as I know anything,

Thanks much!

Well I can finally say I am making progress, Filip sent me a motor controller to allow more power for higher extruder RPM. His controller will not drive my motor, I have no idea why, but it gave me the courage to up my RPM. So I am now at 17 RPM, but what seems odd it that I have increased "P" of the PID to 1.0 from the .020, that is a very big change, but it has made a big difference. I have also changed "D" form the .250 to .125

It is far from perfect, but much improved from where I have been. I no longer get spikes over 1.85 mm or below 1.60 mm except on the rarest of occasion. If the pattern continues I am thinking about 20 to 24 RPM might get me where I want to be. But I want to work my way up slowly as I have been told I could burn up my $74 motor.

I will be increasing the RPM and will report back, have other things to do most of the day though.

Comments deleted.

Got my Red, Blue, and Green ABS colorants today! 2 pounds of each, now I gota get this filament size under control ;-P

Just ordered 28 oz of yellow colorant, so I will be able to make any color I want now : )

You really will have fun ;) Every color has his own settings for temperature and extruder speed in order to control width ....Been there done it.....

Hey CFTechno, I have a fix but I will need your help maybe. I am not sure if this is something you can do or if I will need Filip to help me. Filip wrote the sensor software in Mackerel_main.cpp right?

Would you be willing to try to help me?

If anyone here has played with the PID numbers for the sensor and can give me some idea of what to try I would really appreciate it ;-P

Firing it back up. I have read that Black is the hardest color to extrude so today I will try virgin natural.

Gave up on PID, impossible to get it working. Trying something completely new now. Need to rewrite both the client and mackerel to get it working.

However need others peoples input (their filament audit trails) because every one can have different extruder fluctuation behavior. So the more share their data the better....

I have 5620 mm of data, I tried different temps and rpm's, should be just what you want. I turned on auto at 5620

how do I get the text file to you?

If you tried different temperatures and rpm's you need to save each and every change to it's own logfile. What I do is I connect the client software. Then set the extruder, then wait till I'm sure that the filament in the sensor is the filament produced by those settings. Then I press "clear stats" on the extruder, go over to the client and press "start log"

I measure a minimum of 5000mm or even more depending on the behavior and the "looks" of the waves. If the waves are very regular I press "stop log" after 5000mm otherwise I measure 10000mm. After that I rename the logfile to include the settings in its filename.

Then change something on the extruder, wait till stable, press "clear stats", press "start log" etc

"wait till stable" your joking right? If I could get stable results I would be too busy dancing to record it ;-P

let's make that "settled" then ;-P What I main is that a change will influence the filament fluctuation but because of the distance between the nozzle and the sensor it takes some time before those changes reaches the sensor. You just don't want the data to be polluted with the "pre-change" filament measurements.

I normally wait twice the time it would need for the filament to reach the sensor. So if the distance between nozzle and sensor is 100mm and the puller is at 10mm/s I will wait 100/10 = 10 seconds times 2 so 20 seconds before pressing "clear stats" and "start log"

With temperature changes (upwards) I will normally wait several minutes up to an hour cause changing the core temperature of the extruder takes really long. Changing temperature downwards I try to avoid!

My extruder has never varied from 12 rpm on your graphing log. I can send you my log files if they will help but they don't show PID numbers so I am not sure how they could help.

I have to agree with the PID comment, "impossible to get it working", we do need something completely new. After two full days of changing numbers, I see no hope. I will make a change and get ten minutes of "man I think I got it", then crap, it will jump to over 2.10 mm or under 1.50 mm. The sensor should be able to work, just not the way we are doing it now.

Where in the sketch does the sensor stuff happen? I will be looking

To see and analyze the "real" behavior of your system you should not be using the "Extruder Auto Con" setting. Only that data (with a fixed puller & extruder & Temperature) will give a indication what the system is doing and when the extremes happen.

From there you keep all variables the same and change only one like for example Extruder speed. And then again check/measure/analyze the data.

This will allow you to learn which parts of your system is the main contributor to the fluctuation.

What do you mean with "where the sensor stuff happen" ? The real sensor stuff happens in the sensor and the code running in it.

I should have some time in the next few days to run you some numbers, I hope they help. I will be stepping up the RPM till she blows ;-P

I will really need to make that filament grinder after this!

Ya, I am wondering where that code can be found to see if I can understand what is happening. I am looking at the code in the main_cpp file and am trying to put all the pieces together.

I get what you saying about the "real" behavior, makes sense. I have even wondered about power, I have noticed when I turn on the spooler motor it seems like I get a big fluctuation in filament size. At first I just assumed it was the extra "tug" on the filament, but now I am wondering if it is a power drop. I think I'll try what you have suggested and also have my digital meter connected to my 12 volt supply to see if it is fluctuating.

what is your "completely new" idea?

Still don't understand what you mean with sensor code? Maybe you mean the code running on the sensor itself? You can find that here : http://www.thingiverse.com/thing:454584/#files Look for the "freescale 8 bit code" file

The spooler might cause the filament to start moving in the sensor and that again might cause fluctuations in the measurement.

Filament Width Sensor Prototype Version 3
by flipper

No I mean the code in the Lyman V5, I have found in the main_cpp file part of it but there must be more somewhere?

The sensor sends a value to the mackerel software and that is then feed into the variable "current_filwidth" That's it. What else do you feel there should be?

so this is the total control code then?...

pid_input = current_filwidth;

              pid_error_fwidth = filament_width_desired - pid_input;
              pTerm_fwidth = puller_feedrate_last - fwidthKp * pid_error_fwidth;

              if((filament_control<PULLER_PID_MAX_LIMIT && pid_error_fwidth<0) || (filament_control>PULLER_PID_MIN_LIMIT && pid_error_fwidth>0))
                  dia_iState_fwidth += pid_error_fwidth* puller_increment; //use spatial dT=puller_increment
                  dia_iState_fwidth = constrain(dia_iState_fwidth, -PULLER_PID_INTEGRATOR_WIND_LIMIT, PULLER_PID_INTEGRATOR_WIND_LIMIT);
              iTerm_fwidth = fwidthKi * dia_iState_fwidth;  

              //K1 defined in Configuration.h in the PID settings
              #define K2 (1.0-K1)
              dTerm_fwidth= (fwidthKd/puller_increment * (pid_input - dia_dState_fwidth))*K2 + (K1 * dTerm_fwidth);  //use spatial dT=puller_increment

              filament_control = constrain(pTerm_fwidth - iTerm_fwidth + dTerm_fwidth, PULLER_PID_MIN_LIMIT, PULLER_PID_MAX_LIMIT);

              dia_dState_fwidth = pid_input;


      } else {
          puller_feedrate_last=puller_feedrate;  //keep track of last puller_feedrate when running manual control
          dia_iState_fwidth=0;  //reset the integral

puller_increment=puller_feedrate X 0.6/EXTRUDER_RPM_MAX X 8; X = times

how is this math done?

if puller_feedrate = 1 and EXTRUDER_RPM_MAX = 50

1 X 0.6 / 50 X 8 so .6/400 = .0015


.6 / 50 X 8 so .012 X 8 = .096

and why does EXTRUDER_RPM_MAX even matter to puller_increment, it seems like an arbitrary number. The Lyman V5 certainly could never do 50 rpm.

I'm trying to follow the numbers through the software to better understand what is happening.

It seems like the changes this code makes is very small, but there are parts I do not understand, like...

dTerm_fwidth= (fwidthKd/puller_increment (pid_input - dia_dState_fwidth))K2 + (K1 * dTerm_fwidth);

where dTerm_fwidth is never given a value to begin, so what happens when you multiply K1 (.95) times a non declared number?

static float dTerm_fwidth; no value given, is it treated as zero? or does it return random crap?

Comments deleted.

I am interested in seeing how the PID numbers change the pull rate on the puller rollers. The change seems too slow, too drastic, or too minor, lol not sure which.

I looked at all the graphs the software produced and as soon as I saw/realized how short those spikes (or extremes) were in term of length the PID was out of the window. Have a look at a part of the one of the logfiles

First column is the length. Second is the sensor data (filament width)

260,1 2,36

265,5 2,37

269,6 2,37

275,0 2,36

279,0 2,35

284,4 2,33

288,4 2,32

293,8 2,31

299,2 2,29

303,3 2,28

308,7 2,25

312,7 2,25

318,1 2,22

322,1 2,21

327,5 2,19

332,9 2,18

337,0 2,17

342,4 2,16

346,4 2,15

351,8 2,14

355,8 2,14

361,2 2,14

366,6 2,14

370,7 2,13

So at length 269.6 it was 2.37 and at length 370.7 it was down to 2.13. So within 100mm or 10centimeters the fluctuation is already a done deal. As the nozzle distance is more then 10cm ANY input based control is always to late!

That's why I agree with what [Filip] wrote in the readme for mackerel "Filament control is with a PID loop, but there is dead time in between where the filament is extruded and where it is measured. A Smith Predictor would be a better way to control it."

I started writing some code for a smith predictor (its commented out). Smith predictors have their own problems. For one, tuning is more complicated. The PID loop works for long term disturbances with the default settings. One improvement would be to reduce the dead time by reducing the cooling portion of the filament path. Not sure if you could cool it in <2 cm or so. That would make the process more controllable. Also, with the graphing, I was trying different things to see if it introduced variation. One item that introduces variation for me is air currents on the cooling section. If I fanned it, the diameter changed. I wonder if the cooling fan airflow picks up heated air from the extruder depending on the flow of air in the room and that might introduce variation. I am making a tunnel for the fan so that the airflow is isolated from the room. The filament passes through a hole in the tunnel. I have yet to try it. Another item that introduces variation is poking into the hopper of abs pellets. The extruder output increases. Maybe there are different hopper designs. Also, if there is a repeatable variation due to the screw, it might be possible to remove it by modulating the extruder motor to cancel it out. Extruder motor could speed up and slow down at the right points in its rotation to even out the flow.

That is exactly what I found as well.

Cooling has major influence on the fluctuation however it only "lowers" the amplitude but doesn't take away the spikes. I even had a setup running with paint stripper with a temperature control as cooling. This allowed me create a gradient temperature cooling tunnel for the filament. I was running it from 50-150C in 10degrees steps.

I have tested several hopper designs including different feed lengths. Currently building a vertical version to see if the screw version on a hopper will improve fluctuation.

I have also tested a few dozen screw designs....Puller distance, puller force, temperature ranges, microstep settings, nozzle designs, extruder speeds etc...I could identify around 15 parameters and try testing each and every one of them individually .

Every test run is 6-10meters and I can run around 10 tests a day max. In fact losing testing time because the pelletizer can not keep up with me ;-) Even thinking about putting the pelletizer directly behind the puller but I like to visual analyze the test results before pelletizing them again.

I came to the same conclusion as you : The regulation should not be on the puller but on the extruder. I'll be introducing a new version of mackerel with support for new commands that will allow the lyman to be controlled by the client. Including some settings that will allow a wave like extruder speed to be setup. I'm thinking about two setting that together will control the timing of lowering the speed and one that will control the length of the increasing speed. Maybe something via acceleration settings?

Love to hear your ideas!

PS and also changed the client to improve the logfile layout, had a bug where the output was split across lines. Will release that version today.

PSS The consequence of trying to counter-act the fluctuation via extruder speed is that the objective then Is to find a setup where the fluctuation has very few frequencies with a little noise as possible (thus nice clean waves, great prediction) But the search for the lowest fluctuation in the system without any regulation might produce a completely different setup were the fluctuation has bad prediction and hence hard to regulate. If the fluctuation is low enough (<0.05mm) that's not a problem....

I had to Google Smith Predictor ;-P

So is this your new plan/idea? I have just skimmed it, kinda busy today building a greenhouse. But I will be reading up on it.

Been playing with PID numbers all day, wish I could claim progress. Just when I think I am getting somewhere I get a jump to over 1.90 mm or under 1.60 mm. It's a good thing I have hundreds of pounds of ABS to play with.

Will try again tomorrow, i need a break ;-P

I've build a shredder to pelletize the miles of filaments made. It's based on the design [fred] published on the filastruder forum (http://www.soliforum.com/topic/6997/high-speed-filament-pelletizer/)

It also helps in pelletizing colored filament and re-run those pellets in the extruder to improve color mixing.

Is there a delay built into the spooler speed change to compensate for the distance from the nipple to the sensor? If yes, can I change it?

Not that I know. There is no setting where the distance between nozzle and sensor is set so the system would not know what the delay is.

Gave up on the spooler design. Could not get it to work with the extruder speeds I'm running. Will change the 70rpm dc motor for a stepper, same price and more control!

What are you going to change in the code to run a stepper motor? Wanted to use all stepper motors in the design. Never could figure out what to change myself cause of not knowing anything about coding.

Not really sure what you are asking. The mackerel is based on marlin that has support for several axes and extruders. The Ramps board supports 5 steppers. Two are taken by the puller and the extruder so you can pick anyone that is not used like x, y or z

Adding the additional ramp to any of the axis (x,y or Z) does the mackerel firmware know automatically which one will be used for the spooler stepper? Thought it needs to be changed in the code to use the new axis instead of D8 output. I appreciate your help.

You are correct, adding a stepper stick is not the only thing to do, you need to rewrite the spooler routine in the software to be able to use a stepper for the spool.

I downloaded free PID tuning software but it wants txt file data in a certain order. It will read the stuff from your graph program but it reads the wrong stuff, it probably wants the PID numbers too I am guessing. Maybe you could take a look at it and see what you think. It may be something you can make work, maybe not.


Very nice, thank you again!

I have so much plastic I do not feel bad about throwing away 10 or 20 pounds. Styron donated 831 pounds to me for the Scout project that will only use about 150 pounds, so I will have free ABS for several years unless I go crazy ;-P But I do hate to waste anything, but the better use for me here might be for mixing colors. I am still waiting for my red, blue, and green colorants to arrive, so far I have white and black. The black is very black, even at 1%, but I have been wondering how well colors will mix in the extruder. I wish they had included Yellow so I could mix to get any color, but they did not. Might have to buy some.

I am wondering why the water cooling feature of V3 was abandoned?

People seem to be reporting better results for the V3 than the V5 as far as filament diameter is concerned.

Using water as a coolant has its own problems. Drying the filament is the first issue. Second, need to keep the temperature of the water stable. Any of the temperature will also cause diameter to fluctuate. The commercial filament machines use a water bath with a constant temperature.

So the bigger the water container the better perhaps. I will be experimenting over the next few days, I have several things to test, I am very excited to try out CFTechno's new software.

Wonder instead of a water tank system use a longer heated nipple black pipe. Currently the Lyman 5 has a 4" pipe, the Filastruder uses 8" pipe. Since it has a more melted plastic in the pipe would it be easier to control the diameter. Not sure if an uncut auger bit will be long enough for the current extruder base. Just a though.

Filastruder is 8 inch including the feeder section. The Lyman is 4inch but has a separate feeder section which is around 3inch and some extra space between the pipe and the feed section. So not that much length difference at all.

imho the melted plastic length depends on the headband position, it's width, the temperature and the heat dispersion in the pipe used.

I am currently running CFTechno's graphing software and I might be making headway, to soon to be sure.

Will report back after more time on his system.

And the water did not help, abandoned that.

Mackerel Client software for Windows. http://www.clubfitter.eu/tiki-download_file.php?fileId=78

Usage :

  • Connect your computer with the arduino running mackerel. (the adjusted version that supports the M700 command, https://github.com/CFTechno/Marlin-Mackerel)

  • Power on your Lyman.

  • Start de Mackerel client software (after you have unzipped it). Select the correct com port and press "Connect".

  • After the software has connected with the Lyman, the lyman will reset and is ready to run. As soon as It is running you can press "Start Log" on the client software. The software will show all filament data (current, average, maximum and minimum width) in a graph. It will also start a log file where all data will be save into. That data will include all filament data but also extruder speed (rpm), puller speed (mm/s), temperature (target, current), winder speed, extruded length (mm), time (remaining, elapsed)

  • When pressing "Clear statistics" the graph will also reset.


This is a really great addition. I have not tried while extruding, I was too excited to try the client and did not want to wait to heat it up. But I was able to pull some existing filament through the sensor with the puller and see a plot of a repeating pattern in the filament with a period of about 60mm. Where its coming from, I don't know. But being able to see things like this should help in all sorts of tuning exercises.

If you do a FFT on the data you will find that the auger rpm is to be found in the fluctuation. Till now however I have trouble to quantify the contribution of the auger speed to the overall filament fluctuation.

I have tried everything I can think of, I can't get it to load the "adjusted version"

What do you mean with "can't get it to load " ? You need to download the mackerel version from github, select the sketch in the arduino ide and then upload it to your arduino.

ya, downloaded it and I try to upload it to the arduino, it acts like it worked but nothing changes. I have changed the default filament size and default pull in the .h file but the changes to not make it to the arduino.

I had no problems before when you first posted the min max stuff, now it wont upload to the arduino. I tried moving to other directories just in case it was a pathing problem, still nothing,.

It is not over writing the old version.

I changed this line:

define DESIRED_FILAMENT_DIA 1.75 //define the default desired Filament diameter

not sure why copy paste is so big, sorry

but Lyman still says default is 1.68, which is what I set it to when I loaded the min max script a couple weeks ago.

Maybe a com port issue?

The arduino in the Lyman resets at end of upload so it sure looks like it is working. Baud rate or something off maybe???

OK in easy mode I get the G-Code Send line, I enter M700 and hit Send, nothing seems to happen, where would I see a response?

You might have several versions on your computer. Changing one version and uploading an other...

Connect your lymand with Pronterface or Repetier Host and send it a M700. If there is a reply you have uploaded the correct version.

OK I have Repetier Host loaded and it is connected to the arduino it says, how do i send M700?

I am at Manual control but the gcode line is not there. It does show it is heating when I started preheat, so I know it talking to it. It is even updating the temp properly.

OK in easy mode I get the G-Code Send line, I enter M700 and hit Send, nothing seems to happen, where would I see a response?

Tried it also in Repetier but seems that it doesn't like unknown commands like the m700. I tried it in Pronterface and there it works without any problems.

Re-downloaded just now and it now goes to .2 to 1.6+ so you must have updated ;-P

testing as soon as Lyman warms up : )


Thanks, sorry for all the posts, but it looks very cool. I hope I can get some better results now that I can see a graph.

That's great news! Looking forward to some printscreens from your lyman!

Comments deleted.
Comments deleted.


Thanks, will test and report!


Had meetings most of the day, got very little done here. It has gotten better but still needs improvement.

Well I have to post on my experience with reversing the sensor designed and purchased from Filip.

You can see a picture on my make page: http://www.thingiverse.com/make:120597

I mounted it on its side so the filament would not lay on the "tray" of the sensor and stick to it since the filament is still molten at that point. I have CFTechno's firmware loaded so I can see min and max numbers but the sensor reads all over the place, jumping almost faster than you can watch. It was reading from .77 mm all the way to 2.08 mm BUT the filament was 1.65 mm to 1.80 mm, the least swing I have ever made on such a long run. I ran for over and hour and the filament never went more than .1 mm from the set point of 1.70 mm.

I will print some parts with it in the morning and report. Finally a glimmer of hope, now if only I can duplicate it repeatedly.

Lyman / Mulier Filament Extruder V5

But in effect what also happened is that the cooling moved further way. If you move the sensor back to it's old place you will see that the readings are the same as you see them now.

I'm not sure I understand what your saying?

I understand that I would get finished reading if I put the sensor after the fan, since the filament is cooled to the point that it will no longer change.

I am not sure how quickly it cools with no fan, if you are saying that I could move it back and as long as I had no fan I would see the same crazy readings, I am not sure that is true since when I first started running my extruder I had no fan and did not see such wild changes in the readings.

If your saying a second sensor placed after the fan would read the same crazy numbers, I am not sure that is correct either since when I physically measure the filament with digital calipers it reads between 1.65 mm and 1.80 mm.

I understand the cooling has moved farther away, that was the idea, to measure and correct before the filament cooled. I saw the problem before as being since the filament was cooled when the sensor measured it, it was too late to fix it. It could try to correct the next few centimeters, but it could not repair what it had just measured, that filament was off size and is was too late to correct it since it is already cooled to the point its diameter can not be changed.

I am getting ready to run some this morning, if my MakerGear M2 will run it I will be very happy with the fix. Even if it will not run on my M2 but makes better parts than the 1.43 mm to 2.08 mm filament I have been making, it will be a big step forward. I will also find out if I can repeat the results continuously.

This is an interesting setup. I think much of the variation in the filament is caused by the extruder flow rate changing and not the stretching process. So measuring right after the extruder makes sense. I am not sure what target diameter setting to use there though, since its bigger than 1.70mm. The sensor readings will change if the filament is moved away from the sensor, since it measures the size of the filament's shadow, and that gets larger as the filament moves away. The sensor housing was designed so that the filament rests in the tray to keep the distance constant. Maybe putting some aluminum tape or other surface in the trough so that the filament does not stick?

I'm not seeing the PAWL stl file for the winder A07 part. Was it removed for some reason?

Hugh did not print it, he made it from store bought parts. I did print mine and can upload the stl if you would like.

uploaded it just in case: http://www.thingiverse.com/thing:716885

Lyman extruder Pawl

Sweet, thanks.

The new M code in Mackerel seems to work. The client is also finished and working. It show the filament data in a graph. I'm using that graph to help me in finding the correct PID settings. Have not found the correct settings yet but I can get the fluctuation down to +/-0.08mm without losing to much extruding speed.

A screenshot of the client running can be seen here : http://i38.servimg.com/u/f38/13/37/93/48/fila_f11.png and here http://i38.servimg.com/u/f38/13/37/93/48/fila_f10.png

Trying to read some more about PID and how to program a autotune routine for the Filament PID because that would make life so much easier ;)

Hey CFTechno,

You getting close with that graph software? I am loaded for bear and have high hopes!

I am installing a vibrator on my hopper, and installing a bigger motor controller, both Filip's idea's and the controller is his design.
My stuff has arrived for my profile die, I have your latest and greatest loaded.

I am ready to throw the kitchen sink at this puppy, I would really love the PID graphing ;-P

Thanks a ton for doin what ya do!

Cleaning the interface up, ready for beta release tomorrow :)

I tried the bigger controller but makes no difference except that the couplers and motor gearhead fails faster :(

Did not try the hopper shaker but would surprise me if that did work.

PID will not work as far as I can see because the filament fluctuation behavior (short burst of either large or small filament blobs) and the lag in the response making the response always being to late!

so your saying your PID graphing wont help?

Have you tried flipping the sensor/fan around so the sensor comes before the fan? It makes perfect sense that it is to late to affect filament size once it has passed the fan and hardened. I tried it once but the hot filament would stick to the sensor unit, I was thinking of trying it again with the sensor flipped on its side so the filament would not touch it.

OK, setting it up now to try ;-P

The filament width fluctuation graph will really help in seeing how tuning the PID parameters changes the width. However looking at those graphs and doing some analysis on the behavior makes me conclude (for now) that PID itself will not help to control the fluctuation sufficiently enough to get (or keep) below +/- 0.05mm.

In some of my tests the fluctuation is below +/- 0.05mm for several centimeters and then suddenly will jump up and down again within one second (or down and up). I fail to see how PID can solve that as the sensor will always be measuring those short burst when they are already finished and thus it will already be to late to respond.

People here said they have tried my profile die idea and it doesn't work, but what else is there left to try?

So I will try.

I have my sensor setup back words and it is heating up, I don't need it perfect but I need it better than I have seen so far.
I will report.

Well it is odd LOL

The sensor is giving reading all over the place from 0.77 mm to 2.10 mm, BUT when I measure the filament it is 2.02 mm max and 1.85 mm min, which is the tightest swing I have ever seen I think.

10 minutes later, I now have 1.70 to 1.85, unreal. The problem now is not knowing what I am making since the sensor is reading before the diameter is changed, I would need a second sensor to record the end product. : (

OK I have now moved the sensor a little farther away from the extruder tip and things might be getting better. The fist attemp had the reading part of the sensor about 1" from the extruder tip, it is now 1.75"

The numbers are looking so good now I am going to run the spooler for an hour and then measure it with my digital calipers.

I will post the results

Hey you didn't disable the sensor in your max min code did ya? ;-P LOL

I have been changing PID numbers and see no change at all in min max swings

First I put D at .500 and filament never went over SP, then I made a change to P and could not make progress, set it back and went back to D at .500 and not seeing the same result as before??? I have a headache ...

You need to clear the statistics for the max and min to be reset to zero. So every time you change one of the PID parameters you wait a few seconds and then clear statistics. After that the max and min will show the new max min for the filament under this new PID settings. I measure the distance between nozzle and sensor and then divide that distance by the feedrate to see how long I should wait before resetting the statistics.

I have just updated github with the latest mackerel version where a small bug is solved for the "clear statistics" option. It also includes the M700 commando and the minimal puller speed is set to 0.5mm/s etc...

Is this the one that will display filament diameter on a graph live?

If no, what is M700 commando?

If yes, how do I display it?


and ya I have been clearing statistics

Every 3d printer firmware supports gcode (http://reprap.org/wiki/G-code). Mackerel only supported the M303 commando (see readme) I have added the M700 code to the mackerel firmware. So as soon as you have loaded this version on your arduino your extruder will answer a M700 command.

If you would enter the M700 in Pronterface or Repetier you will see the reply from the extruder.

If you want to see the data returned in a graph you need software that sends the m700 commando to the extruder and translate that data into a graph. That is what the client software is doing. Still working on it, hope to release it soon.

Comments deleted.

So half way there?

Very cool!

Thanks again!

So are you a programmer by trade? If not, what do you do, you seem very good at this ;-P

PS funny fact : After analyzing a matrix of different extruder speeds and puller speed while recording the filament width for a limited length (5000mm) it was possible to extract the extruder rpm from that data! I tried to run a simulation without this extruder "heartbeat" in the filament width and it seems that without this "heartbeat" the fluctuation would only improve between +/- 0,002mm - 0.007mm

This was at a limited test set: extruder speed 2.0, 2.5 and 3.0rmp and with the filament width set at 1.8, 2.0 and 2.2mm for each extruder speed.

I hope to do some more tests in the following days.

I am wondering why you are using such low rpm's in your tests? I am running at 12 rpm and it takes about 8 hours to extrude a kg of filament, I would think at 2 rpm it would take all day? Is it easier to test and see results that way or something?

And I do not understand the "it was possible to extract the extruder rpm from that data! I tried to run a simulation without this extruder "heartbeat" in the filament width" Are you saying there is a direct correlation in the numbers? But if that is what you mean then "and it seems that without this "heartbeat" the fluctuation would only improve between +/- 0,002mm - 0.007mm" confuses me even more LOL sorry I am so dumb, wish I understood.

Does your new M code display "live"? Can I watch it as it is extruding or do you have to review it after a test run? I feel so helpless here, sorry to be such a pain. Any chance you could email me your early copy ;-P kinneplastics@yahoo.com

At least then I could try to play with PID numbers and feel like I am doing something, I have over 500 pounds of ABS I can waste trying LOL

Low rpm : As said before I'm running a adjusted version of the Lyman. These speeds on my version will generate around 21mm.per.second of 1.8mm. filament. Roughly 4.2hours for 1kg. At maximum speed it can do 1kg ABS in 90minutes (@195C).

Heartbeat: I'm saying that if you analyze the filament width fluctuation the speed of the extruder can be found in that fluctuation. So the extruder (or better the auger) disturbs the filament width, normally that would be bad news but it might be that the total influence is very small which would be very good news. I'm measuring\analyzing in the hope it will help in writing a filament width PID autotune routine.

M700: Yes when using the client software you can see that data live. The client will also save the recorded data. I'm also working on controlling the extruder from the client data just like you would do with a 3d printer. So less need to control it via the LCD or view progress from mobile or website.

Hope to release it soon. It's not 100% bug free yet and the client is only running on windows for now. Need to learn a new programming language to get it to run on other platforms...LOL...

I run on Windows ;-P

currently reading: http://eas.uccs.edu/wang/ECE4330F12/PID-without-a-PhD.pdf trying to learn

There is a PID autotune for the extruder, can't you copy it? Or is there some changes necessary?

Just wondering, hate seeing my extruder not making filament, no sense in making bad filament : (

No reason to copy it because as mentioned in the manual : "..The PID for the puller must be feed rate-based and not time-base like the temperature. This will allow the same PID parameters settings to work for any feed rate...."

Wow, very nice!

How does the data get to the graph? And autotune would be sweet.

When will you be ready to share? ;-P I think I would make the perfect guinea pig!

Thanks for your work!

The client software connects via de usb port and sends a M700 commando to the extruder. As soon as all small bugs are dealt with then the update of the mackerel will be released including the clients software.

You said you have made "modifications" to the V5, do you expect those of us without your changes will get the same results?

Yes, that is the objective.

I would be so happy with 0.08 fluctuation I would dance ;-P

Work hard, work fast LOL

I have just ordered the stuff to make an extruder profile die, if it works as intended it should make perfect diameter filament all the time. I worked on big commercial extruders at several of my jobs in the past and they used dies to make shapes and sizes for extrusion. The dies were large and it was easy to make them quite complex, this will be a challenge since it will be so small and have to be much simpler.

If it works it might require a rewrite of the software, but I wont know until I have it finished and test it. My concern there would be tension trying to pull the plastic out of the die too fast or not fast enough, but there are fixes used on big extruders that should work here if needed.

A profile die basically just forces to plastic into the shape and size desired and holds it there long enough for it to cool to the point that it will then hold that size and shape. Sometimes the end of the die is under water to "freeze" the polymer as it exits the die, I don't know if that will be required in this case or not, I might try a waterfall or something to avoid the hassles of working with the die submerged if water is needed.

This is just an idea I am going to try, I really have no idea if it will work ;-P

If it works I will post the info here, if it fails I'm out $60 so far.

I am surprised none of the filament extruders use a profile die, so maybe they have already tried this and it wont work, but I shall find out.

What will be the difference between this new die and the filastruder nozzle you are using?

It will have a cooling section for the filament to "set up" and the hole will be exactly 1.75 mm in size. I ordered a pack of 1.75 mm drill bits and a funnel shaped bit, it will have a screen too. I have several designs in mind and will try them all before giving up. This is how they make plastic pipe and plastic square hollow tube commercially, it is very old technology. The trick will be to do it in a cost effective way. If my idea works they would be reasonably priced to make. I have a lot of friends that are chemical engineers at Dow Chemical in Midland Michigan and have asked one of them for his input, so I think I have a reasonable chance of success.

ABS does shrink when it cools so a 1.75 mm die might make 1.73 mm filament, I could live with that.

I tried this a lot with various combinations of drilled out brass screws, PEEK and PTFE rods. It always jammed by the time the filament cooled enough to hold its shape. I got it to work very briefly once. I tapped the nozzle with 6mm threads and screwed the cold end of a Makergear hotend into it. My reasoning was that a filament extruder is basically a printer extruder in reverse. It made beautiful shiny 2mm filament, but only at 1-2"/min. My last attempts were basically a backward J-head with a 1.75mm ID. I had a bit of drilled out 6mm brass which screwed into the nozzle at one end, and a PEEK pipe at the other. The PEEK contained PTFE drilled to 1.75, and I ran the bit through the brass and PTFE together to guarantee that everything lined up, and it still didn't work. With a printer hotend, if you manage to get melted filament back up into the PTFE by retracting too much it is going to jam. The same thing happens with a filament extruder.

It is possible that it might work with the pull so long as you don't pull faster than the extruder is pushing it. You would need to heat the die first so you get something out that you can pull, then let it cool back down. The question is, when the friction hits, will the filament at the opening be solid enough yet to overcome it without breaking.

Tim here, from Filastruder.

Short version: I tried what you are suggesting, it does not work.

Long version: If you cool part of the die (I used a PEEK insert), the polymer will stick to the internal surface. Industrial extruders develop a couple orders of magnitude more melt pressure, so friction is not a problem. You will find that once you allow a region of the die to drop below the melting point of the polymer being extruder, friction will skyrocket and you will not see any output.

Not good news, I wondered if anyone had tried it. I doubt I will be trying anything different than you or ianjohnson,

But I am kinda stubborn so I will try. Besides I already ordered the stuff.

There has to be a way to economically make good filament. From what I've read on the different forums none of them seem to do any better than what we are getting here.

Have you tried extruding into water? I have worked on commercial units that extrude under water, but like you say they have a lot more pressure. Maybe a bigger motor on the extruder to raise the pressure enough to force it though a cool ended die?

The simplest, cheapest way is still extruding vertically down into a loop and back up into the winder. Gravity provides a constant pulling force, but doesn't constrain the filament the way pulling in a straight line with a roller does. If the extrusion speed changes, the filament is free to move faster or slower rather than bunching up or stretching out. The winder runs faster or slower as needed to keep the amount of filament hanging (and pull of gravity) constant. It takes up more space and isn't as desktop friendly as direct pull. However direct pull requires accurate measurement as close to the nozzle as possible combined with PID tuning to match changes in extrusion speed which is hard and complex, or very consistent extrusion which might be achievable with a proper extrusion screw which is expensive (Filabot has one for $600).

I've tried the linear CCD, and also came up with a low-tech version of a laser micrometer - https://www.dropbox.com/s/pvncx6etotbvyet/2014-08-26%2013.44.01.mp4?dl=0 The laser shines through a pinhole into a phototransistor. As the laser/sensor pair sweeps past the filament, the sensor will be shaded. The filament diameter is measured by how long the sensor is in shadow. It is a little sensitive to sideways movement of the filament, but keeping the beam as small as possible helps. Vertical movement of the filament doesn't affect the measurement, and since the filament doesn't need to be constrained it can be measured right at the nozzle. I got it to the point where it could give me consistent measurement, but never tackled the PID. Getting the measurement to work was an effort, PID is a whole other thing, plus it requires adding two steppers and drivers. It was a lot of extra cost and complexity to achieve what I was already getting with the simplicity of the dropped loop.

How tight of variance can you get with this hanging method? What would be the minimum and maximum diameters over a 1 kg spool of filament?

Thank you,

So with your nozzle I should be able to come close to these numbers by hanging my filament rather than winding it up?

I can't guarantee that, since there could be something in the rest of your extruder that is causing the diameter issues. I'm just reporting other user experiences in case they help you get to the bottom of your issues... and to give you hope that its possible to extrude quality filament1

.05 variance is typical. Actual diameters depend on temperature and nozzle size, as well as keeping ambient conditions constant.

Looking forward to the results ;) The shrink might counter the swell that normally happens if the extrusion is fast enough or the die length (land?) is not long enough. There......all my 2cents worth of knowledge given away in 1 sentence

Video making Black filament on my Lyman V5:


This is Styron ABS Magnum 347EZ, with 4% Black Colorant, I have already printed parts with it and it prints as nice as any filament I have ever purchased.

Max 2.02 and min 1.25 could cause some problems in the hotend and in the print with the layers?

4% is higher then I have seen before, normally I see 1% or 2%.

Even at 1.5% the extruded filament is very very black, next patch I will try at 1%. You say here min 1.25, is there a reason for a min? Is the final look not the only concern?

Thanks again

Your filament width was between max(imum) 2.02mm and min(imum) 1.25mm according to the video. That's a big fluctuation imho

Oh, I thought you were saying 2.02 and 1.25 as % limits for colorant, sorry.

And yes it is causing problems with layers, the Prusa i3 extruder head does not seem to have any problems with the size fluctuation though, but my MakerGear M2 will not run it. That is why I am trying to make a die.

The supplier told me 4% is what they use, this was donated for a Boy Scouts of America Varsity Camp robotics class by a large plastics extruder I buy sheet from for my thermoforming business. The 347EZ was donated by Styron.

The first part printed perfectly but since then I am having a de-laminating of the layers, so this is most likely caused by too much colorant? Cool, I will cut it back to 1.5% and see how the filament looks and works.

"some problems in the hot end" so it might plug the nozzle with color particles? I am having a problem there too maybe, not sure what is all happening yet, I am on a steep learning curve.

You seem to be the most knowledgeable here so I hope you don't mind me asking you for some desperately needed help.
Did you notice I am using you min max code in the video ;-P

I am thinking my biggest problem is filament diameter. I do have Filip's sensor but my filament when set at 1.68 mm varies all the way from 1.43 mm to 2.08 mm. I "think" ( not sure ), at 1.43 mm or something slightly more my extruder feed loses a tight grip on it causing it to stop, at 2.08 or somewhere less it is too big and will not fit though feed holes and jams. Maybe some of the problems I am having have to do with too much colorant, I do not know, but I will re-extrude today at 1.5% colorant and try to make some prints, but...

Can you PLEASE give me some pointers on getting less swing in filament size? I have played with PID numbers and sometimes see no change at all and other times it makes it much worst, I have never seen improvement. I get the impression that Filip and Hugh have mostly if not only, made 3 mm filament. I am wondering if the written code for the sensor is designed around 3 mm filament. Have you or do you make much 1.75 mm filament? If so, how much variance in filament diameter are you getting?

Sorry to be so long and ask so much but you seem to know more than most, if not all.

Thank you very much,

Yes saw that you were using the adjusted mackerel software ;-)

I do not know much but try to read and learn as much and as fast as I can. I think I have read all postings on the filastruder forum in order to learn more. This is what I think :

If you add to much colorant it can change the property of the filament (becomes brittle) and can cause trouble both in extruding (width fluctuation) and printing (layer bonding) it depends on the material used in the masterbatch (your color) and how much you add.

Filament width fluctuation will cause issues in printing : It might be to thick to pass the hotend causing plugs and will cause inconsistent plastic flow creating layer bonding issues and perimeter issues.

I still do not know why the width fluctuates and what the best method is to control it. But isolating the extruder AND the nozzle will improve the fluctuation immediately. Cooling the filament will influence fluctuation, extruding temperature will influence fluctuation and extruder speed will influence fluctuation. No idea which is the major contributor however :( Puller speed seems to have minor influence unless extruding real thick filament (>4mm)

I'm testing both 3mm and 1.75 and the smallest fluctuation I can get is around 0.1mm BUT not for a full run. That's why I'm writing that new version of the mackerel firmware to include a complete audit trail. That will allow me to play around with each and every parameter and then see the consequences it has on the filament fluctuation..

According to the maker of filastruder the best control for width is speed. The slower the extrusion the better the fluctuation.........

Wow, you are correct, taking the colorant from 4% to 1.5% fixed the brittle filament problem! I was not sure why it was so brittle, I thought maybe it was just the new 347EZ was more brittle which worried me, it is great to see that fixed, thanks a ton!

What do you mean "isolating the extruder AND the nozzle will improve the fluctuation immediately", what do you mean "isolate"?
How do I isolate them? Do you mean to put insulation between them so the extruder is not heating the nozzle? If that is what you mean, what would work that wont melt?

I have the fan shield as close as I can get it to the nozzle to get the sensor as close as possible.

Thanks for you speedy response! I thought I got the impression you are in France, is that right? If no, where are you located, I am just curious.

Yes, living in France ;-)

With isolate I mean that you make sure that both the barrel and nozzle are completely covered in the welding blanket you are using maybe even extending the isolation a bit over/past the nozzle. Also make sure that the fan doesn't create any draught near the nozzle, PID tuning the temperature is also important.

Ah, ok, will do


Is the light on your sensor blinking or solid? Filip told me it should be solid but I think I am getting better numbers with it blinking. I am still running tests to confirm. I did rap the nozzle with welders cloth and am running a test now.

What RPM are you running you extruder at? You said it makes a difference, so I would like to try your numbers on my V5 and see how they work.

I would be very happy with a .1 swing, set it at 1.68 and 1.58 to 1.78 mm would work just fine I think.

I wonder how well some of the other extruders and extruder kits work, I have read a lot of their forums and see a lot of unhappy people. But I don't know if only the unhappy ones are posting, are there happy customers too? I see one advertises .05 tolerance, and all they use is a nozzle, no fan, no sensor, just extruding onto the floor. I wonder how well they really work.

Solid, But before I start extruding I will make sure that the sensor measures 0 before extruding, if not then I will recalibrate until it does. Sometime I need to clean the sensor before it will reset itself to 0.

I can give you my numbers but I'm running a extruder that is completely changed from the original Lyman. This because I found no information why some design decisions were made so I'm altering everything and recording the consequences for each and every change. I also read what the maker of filastruder said about the possible design errors and I'm trying to address those in this altered design.

Because you are using the filastruder nozzle with the smaller hole the backpressure (even if you enlarged it) in your extruder is higher then in mine making you motor stall earlier when running with my numbers. I'm pushing mine to extrude at 12grm/minute at the moment and hoping to improve that and the fluctuation. But it takes a lot of testing, time and plastic.

In my experience it is impossible to extrude low tolerance filament without width sensor, how do you know that what you're extruding is good? Making filament is easy, reducing swing is not :) I'm also trying to figure out all the parameters. I have a question though, when using a stepper motor for the extruder, can you be sure that you have constant rpm?

There is always the possibility of steps skipping and thus causing inconsistent rpm either at the extruder side or the puller side. Using a encoder would allow measuring and solving but would also increase the price..

Question is how much steps can be missed before it shows in the filament width.

The RPM question I guess I can't answer, I'm not sure if the display only displays the set point or actual.

Even with the width sensor it is hard to extrude low tolerance filament, I wish I could get it much lower.

My issue with filament size is that my MakerGear M2 has a very small opening inside the extruder, if the filament gets just a little too big it will jam, I have even had commercially made filament I purchased lock up in it. The two Prusa i3's I have do not seem to care, I thought I was having problems with filament size in them but when I cut back the colorant to 1.5% from the 4% as CFTechno suggested those problems went away.

And to make the problem worse the two Prusa i3's do not belong to me, I am teaching a robotics class at a Varsity Boy Scout camp this summer and they purchased the two printers so I could get the forty robots printed on time. I have no idea what will happen to them after camp, they may sell them to the highest bidder.

You can see the robot I am printing here: https://www.youtube.com/watch?v=AVnZjbAslXs

I am having problems making the filament width sensor board. Were can i find the gerber file so I can order the PCB. I see there is an 8 bit microcontroller, Where do I find the program to put on it. I have read the manual and can only find the schematic on pages 25, 37 and the bill of materials of page 17.


The instructions will point you to the link for the sensor and there you will find the link to version 3 : http://www.thingiverse.com/thing:454584

Filament Width Sensor Prototype Version 3
by flipper

Ok, I found them both. The new PCBs are on order! The will be an exiting experiment working with HDPE plastic. For my final's project I am building a special heated build chamber 3d printer to use HDPE plastic, and this recycler to melt down plastic shopping bags for the filament. If I am lucky, it might get me a scholarship.

Can someone update the Heat Band part number from Zorotools? G472516, doesn't seem to be listed anymore? Or post the specs, cause actually I need the 220Volt variety.

afaik you can order any bandheater you like as long as it will fit the size of your coupler

The Styron ABS Prime 555 will not print, it is too thick, but I am learning. It has a flow rate of 2.5 and to print we want something in the 10 to 12 range. I should have some 347EZ ABS by Styron soon, EZ stands for easy flow and it has a flow rate of 11.5

No doubt that is part of the reason it extrudes so slow through the Lyman V5, taking 10 hours to extrude a 1 kg spool.

But that says a lot about the Lyman V5, being able to extrude such stiff plastic with out any problem! It should be a breeze to extrude plastic with a flow rate of 11.5 compared to the 2.5 flow rate I have been extruding.

FYI though, I did upgrad to the High Current DRV8825 Stepper Motor Driver Carrier from Pololu for the M1 extruder motor..

OK here it is! The much anticipated ( by me any way ;-P ) video!


Get some popcorn and a drink and enjoy!

Hey Filip told me about your modification to add min and max filament sizes to the display, I downloaded it and I love it!

Very nice job, it works sweet!

Thank you!

Thanks for the feedback !

I'm working on a few other changes like the full audit trail for the run either to the serial port or to the sd card. But I'm trying to stay as close as possible to the original so not just patching it into the code so it takes a bit more time.

I will check also your remarks about the length and include the changes that you made to allow for a lower puller speed.

Found the length issue: Well not really an issue ;-) The puller steps per mm in the firmware are set as default to 2653.6 So when using the nema 17 1:14 geared stepper motor as defined in the BOM then the PSteps in the firmware are not set correctly by default! The firmware have the Psteps/mm set for a 99.507:1 geared stepper motor. So you have to change the PSteps/mm to 375.3.

An audit trial that is displayed graphically would the sweet, so you could watch the swings to see how PID changes really affect stuff.

Sounds good, yes?


For displaying the trial I'm using a TSL1402R sensor with Arduino Mega+display, see http://www.thingiverse.com/thing:668377

Filament Width Sensor - TSL1402R + Arduino Mega (Work-in-progress)
by zasf

maybe include some more info ? Like the number of measurements per second? I would also like to hear what that setup with the filament rollers does that's visible behind it.....

The sensor is read 4 times per second. At the moment I don't do any smoothing, low pass filtering, average, nothing. I believe it is more sensible this way. If the trail is shaggy you can spot when the filament has bubbles or even little imperfections. Rollers are used as puller in a pure "Lyman style" :) there are three of them to make sure that the filament diameter is not pulled by the winder which would alter the diameter. Let me know if you have any other questions

thank you for that info and 4fps is not bad at all! Do you think you could integrate it into the mackerel software and still have enough performance from the arduino mega to do the other work the firmware is doing?

I saw those rollers and also some extra block. Those block force the filament into an other direction?

The puller changing the filament width would mean that the filament is still hot/warm enough to change it's diameter? If I understand things correct then that means you are running it at high extruder speed and rather hot?

Can I also ask how you made those rollers? I bought the polyurethane rollers but getting them to Europe was ridiculously expensive. So if there is any good alternative I would really like to hear it :)

I think it would be easier to use a graphical lcd with Mackerel and add a menu entry to show the trail, that would be a nice feature someone with coding skills could work on. When I started my build Mackerel was not out yet, otherwise I would have sticked to mainstream :)

The rollers are made in plastic locally with a lathe and we put some bicycle inner tube rubber. Puller rollers are the same as runner rollers but with different dimensions. Now that I think at it, these could be also 3d printed, it's just a cylinder with a cut on the side. I also tried to order them but couldn't find any suppliers in Europe. Maybe it could be an idea for Lyman v6, another step towards self-sourcing the extruder.

Three rollers are used for pulling so that the filament makes a "S". In my setup, width sensor is in the middle between puller and winder. I wanted to make sure that the winder is not pulling the filament and therefore alter the diameter. Only the puller is in charge of that. The "S" setup for the puller could be another great idea for v6, it is simpler than current v5 puller and stronger.

Again thank you :)

Might be that using a graphical lcd would use up to much cpu cycles and therefor slowing down the sensor reading. Also the time interval visible on it would be rather short. Today I will be testing some client software that reads the data from the extruder (I added an special M command to the firmware) and show the graph for the full length of the run. If I'm done testing I will post some images and also release that version of the mackerel firmware.

How are you driving those 3 pullers? Each has his own motor or are you using belts or gears? Maybe make a new thingiverse entry? I'm sure that more people from Europe are interested as those polyurethane pullers are very hard to get and expensive

Yes, thinking about that by transferring the data via de serial port as well

One other note:

My 1 kg spools fill up at 40,000 not 400,000. So the Lyman V5 150000 default length would run all over the floor for 3 mm. That number should be 15000, I think someone put one too many zeros on there. ;-P I put 40000 mm of 1.75 mm on a spool, that leaves about 3/4" of spool remaining which is about the same as when I was buying filament.

It was pretty cool to go to bed last night with the V5 running and wake up this morning to a finished full spool!

I plan to upload a video this afternoon!

3/32" which is 240 mm,

3/32 -> 25,4mm/32*3= 2.38125mm

So the Lyman V5 150000 default length would run all over the floor for 3 mm. That number should be 15000

150,000mm = 15,000cm = 150meters. 1kg 3mm ABS~ 130meters, 1kg 1.75 ABS ~ 400meters (or 40,000cm = 400,000mm)

According to the manual the length measurement is in mm so that 150,000 is correct but maybe there is a conversion error in the source code somewhere.

It was pretty cool to go to bed last night with the V5 running

Hope you have smoke/fire detectors in that room where the extruder is running.

I do indeed have a smoke alarm, I currently still have the V5 in my computer area in my living room.

I did also assume the V5 is reporting cm not mm, but if someone left it running expecting to come back to a full spool they would get a very big surprise, ten times the filament they were expecting. That would be a big job to spool up by hand or throw away.

Sorry to post so much but I always hope it will help someone.

I enlarged the hole in my nozzle from 3/32" to 1/8" and I went from taking 33 hours to extrude 1 kg of 1.75 mm filament to taking only 11 hours.

Also I plugged my first screen in less than 2 kg of filament, so a screen is a good thing to have. All that crap will plug the tip on your 3D printer faster than you can imagine. My extruder motor started jerking and I was wondering what was happening, now I know it was loading it down trying to push plastic through a plugged screen.

This is definitely a learning experience. I am having a great time!

Found someone who was making their own nozzle with a metal screen. All you need is a brass plug, 5/8" hose washer, and a snap ring. Seen several version of the 5/8" hose washer, compared the metal screen get the tighter woven screen. I purchased the Danco washer in the end. Made the nozzle, never had any problem with particles in the filament.

During the initial purge to get all the oils out and particles make sure you leave the screen out. Afterward use MEK that dissolves the abs not acetone cause it makes the abs into a puddy mix. Believe the total cost was about 5 to 6 dollars.


That is the same 1/2" pipe plug I perchased, I used one of the extra screens that came with the filastruder nozzle I bought. I would however recommend putting the washer in first, otherwise your used screen area will only be the size of your nozzle hole and would plug faster. By putting the washer in first your used screen area is the size of the hole in the center of the washer, much more surface area, much longer between required cleanings. I have never heard of MEK before, I will check it out, thank you!

Thanks for all your information, I got problem with plug printer too.
Where dit you get your nozzle with screen.

I ordered one from Filastuder and ordered extra screens. I also bought a brass 1/2 pipe plug and drilled it out, the extra screens fit it perfectly. I just ran a full 1kg spool last night, it took 13 hours and 5 minutes. When it started it was pulling at 1.10 mm/sec, when it finished it was at 0.69 mm/sec. So I will be changing the nozzle out after every run until it cleans itself up. I soak the nozzle in Acetone to dissolve the ABS so I can clean the screen.

You can buy the nozzle here: http://www.filastruder.com/products/nozzle-die-with-integrated-melt-filter

Hey I think I got it!

I changed "#define PULLER_PID_MIN_LIMIT 0.2 //min output limit of filament dia control in mm/sec"


"#define PULLER_PID_MIN_LIMIT (int) 0.2 //min output limit of filament dia control in mm/sec"

so I did an explicit conversion, I found "extern int feedmultiply;" in the Mackerel.h file and thought maybe it was the variable that passes the tension correction and it is an int not capable of storing 0.2 so I took a chance.

And it fixed it!

Now I need to play with the P and I numbers to see how much swing I can get rid of. This is getting fun!

It now passes back a correction of 00.34 to mine and my average is 1.74 mm, not bad, not bad at all.

edit: I have talked to Filip and he says what happened is when I put in (int) 0.2 it changed 0.2 to 0, so I could have just put in...

"#define PULLER_PID_MIN_LIMIT 0.0 //min output limit of filament dia control in mm/sec"

if I understand his email correctly. I did not try this since it is working, but I no doubt will at some point.


Has anyone else made 1.75 filament on their Lyman V5? I am having a problem with the filament sensor. At 12 rpm on the extruder screw I need a tension puller speed of about 0.47 but in auto mode it seems as though 1.0 is the minimum. I did change the "#define PULLER_PID_MIN_LIMIT 0.2 //min output limit of filament dia control in mm/sec" from "#define PULLER_PID_MIN_LIMIT 1.0 //min output limit of filament dia control in mm/sec" but it still seems that 1.0 is the minimum. I set it manually at 0.40 and get it close to the 1.75 mm, but when I put it in auto mode it drops to about 1.16 mm in less than a minute and never comes back up.

I tried to speed up the extruder screw above the 12 rpm default, but when I do it overloads and faults because of high current and I even have the 8825 high current driver from Pololu. I am using the exact motor for M2 suggested in the manual, I even bought it from Kysan.


Looking at the code the 1mm/sec puller feedrate is controlled by the PULLER_FEEDRATE_MIN 1.0 in configuration.h I have changed it to 0.5 and will test it.

indeed changing it from 1 to 0.5 will change the lower limit of the puller to .25

I finally got mine up and running. I am just getting started so I have a lot to learn but I think it is going quite well so far. I have it set to make 1.75 mm and after two hours, 4350 mm of filament has been extruded. The diameter is swinging between 1.60 to 1.80 mostly, with an average of 1.69. I made the mistake of trying to print with some of the first filament I extruded and learned the hard way about dirt and metal chips in a newly made extruder. The current spool is after several hours of running and with a screen in the nozzle, it looks much better, at least I can't see any specks in it like I found in the first stuff after it plugged my printer nozzle. I could not find a #37 drill bit locally so I used a 3/32" which is 240 mm, the 1.75 mm holed nozzle would not produce 1.75 mm filament just as Hugh said. I had to edit the .h file to allow me to set the tension roller speed to less than 1, I currently have it set at .60 and am getting pretty good results for just starting out. My filament cooling fan still has not arrived and I am running without one. The spool winder is working flawlessly, I don't see any reason at this point to add the rubber band upgrade. I have taken some video and will upload it to youtube soon.

I have not messed with filament PID settings but most likely will in the near future. I have several colorants come Monday or Tuesday and can't wait to make some colored filament.

Where can I find the BOM (other than the stl's) for V5?

Under "Thing Files" the LYMAN MULIER FILAMENT EXTRUDER.pdf pages 15-22

Thank you! it was hidden between all the sexy stl's.

Is this filament extruder able to extrude abs with nema stepper motor? ?because I had made an filament extruder which was unable to produce abs filament because every time the motor get dead or wears out quickly

I just uploaded a new STL file of a new Level Wind Guide with clips on each side. Instead of weight I notched the top of the pawl and attached a rubber band over it to the clips. Also uploaded is picture showing this guide.

Another change I made is the Auger. I purchased an Auger from Filabot as their auger has a little larger diameter. I had to machine the hex end toward the screw about an inch and cut off the excess to fit the extruder body. Then I cut the auger end so the total length was about 10.75". Next I ground that end to a cone shape at about 45 degree angle so the point was centered. This length fit my extruder so the point end of the auger was 1/4" from the nozzle. I am using a plain brass hollow plug drilled with a 4.25mm hole.

This is the extruder settings with this auger.
M1 - 17rpm
M2 - 5.96rpm
Temp - 210 C
FR - 11.93mm/sec
AV - 2.61mm (2.60mm is for my slicing)
Variance +/- 0.1 with occasional +/- 0.15mm (Filip's printer sensor adjusts good for the variance when printing)
Tela - 3:29 for 150,000mm

I am waiting for a quote on my Feed Screw design so it is not yet tested. Etienne in France is also getting a quote.
I would be interested in knowing how many folks would be interested in this Feed Screw Auger, if it is an improvement. If you are interested send me an email: hlyman@comcast.net

So you purchased the $630 Auger? That would more then double the cost of the V5.

What improvement did you see? Was it worth the money?

I am wondering what you are using to insulate the heat chamber, it looks nice and must not be melting, what is it?

No, I purchased the $11.00 auger.
The insulation is a welding heat blanket I got on eBay.

Thanks, I just ordered an 18" x 18" x 1/4" weld heating blanket.

I hope this thing works when I'm done LOL

I will post youtube video's of mine extruding filament as soon as I get it going, still waiting on those last three parts, now 4 with the blanket ;-P

I am surprised there are not several success videos on youtube

Saw that a new auger design has been added, interesting!

Has that new tapered auger been tested? Any results to make public already? Or an indication of the cost for such a custom made auger?

See above post.

I see in the manual how to change the defaults but no suggestions for 1.75 mm extrusion.

Should extruder RPM be changed from 12?

What about filament length till shut down? Assuming a 1 kg spool.

I find on the internet the numbers of 130,000 mm for 1 kg of 3 mm ABS and 400,000 mm for 1.75 mm ABS, anyone know how accurate these numbers are? Does the 150,000 default overflow a spool or fill it to the max?

Should windup speed be adjusted from the default of 6?

Does anyone have any experience extruding 1.75 mm and have some numbers?

Also the default for 3 mm seems to be 2.8, why not 3? Should the default for 1.75 be something less than 1.75?

I will most likely never extrude 3 mm filament.



For your desired filament diameter, make the nozzle hole 1.5 time bigger. So for 1.75 use 2.64mm hole (#37 drill bit).

In weighing a spool of 150,000mm 2.60mm filament I find it to be close to 2 lbs. My 8" spool fills to where there is 3/4" left with 150,000mm, so I could get another 50,000mm but then it is hard to handle in the printer. Who cares just extrude another spool.

I use 2.60mm filament as my printers like it. Filip uses 2.50mm filament. Also, I find that my printers will work with filament from 2.2mm to 3.10mm, (I made my own hot ends) so if my filament does have a bad blip it doesn't matter as long as it is being adjusted with Filip's sensor on the printer.

Adjust the M1 & M2 speed till you get the desired diameter.

Most things will become more clear if you have your unit up and running.

Extruder speed will control the amount of plastic it will extrude, like (example) 4gram/min (also depending on temperature). If you set it lower you get less gram/min, higher more gram/min.

The puller will pull that (again example) 4gr/min into a filament with a specific width. The higher the speed of the puller the lower the width and vice-versa. The correct combination of settings depends on your extruder, motor, plastic, temperature, cooling etc etc. Even the color you use will influence your settings!

If you want to exactly calculate the meters you need for 1kg then take a look at the spec sheet for your plastic where you will find the gravity value. That value you can then use to calculate the exact length for 1 kg

Extruding filament at a width also means that you have to include the fluctuation. If that fluctuation is (again example) 0.2mm you are extruding between 2.6 and 3.0 when set at 2.8. Would you set the width at 3.0 your filament would measure between 2.8 and 3.2. That 3.2 might then be blocking your hotend as most hotends have a channel width of 3.2mm.

So does the solid state switch get hot? Is it mounted onto the power supply cover? Will it melt the case if mounted to it?



Used the sensor to measure some PLA I bought, thought it would give a good indication what the target should be. The sensor measured the width to be between 3.03 and 2.94 Measuring by hand gave roughly the same values. So something around +/-0.05 would be a nice objective;-)

I just uploaded a pdf file of the drawing for the tapered auger screw. I am still waiting for a quotation to get a prototype for testing. The auger is sized to fit my current extruders (3) with an aluminum hot end 2-3/8" long. The difference between this screw and Filabot's is that the taper section is only in the transition section and not to the end of the screw. This was taken from a drawing sent to me by Etienne Berthhaud of France. Once tested and it shows a better tolerance filament I will have some made and make them available if there is enough interest. I think Etienne may do the same.

I designed the extruder, but I am limited on the electronics. You need to contact Filip Mulier, he is the expert. I have filament extruded with some spools as far off as 0.30 +/- diameter. I am printing on my two printers everyday making parts (100 at a time) for my padding press (paddingpressmagic.com). I have Filip's diameter sensors on each printer. It changes the extruding speed of the hot end to compensate for the diameter changes. It requires Filip's marlin firmware upgrade. I have smart graphic LCD's on my printers that shows the diameter of the filament and the extruding % as It changes. Further I have designed a new auger screw bit that I will publish when I get one made and tested. At this moment I checked the printer LCD and the diameter showed 2.19mm which is off by more than 0.40mm of my extruded diameter of 2.60mm, printing at 100mm/s, but my parts are printing fine. These blips are just momentary. Also my TAZ3 prints better than my DIY printer. Each has the same hot end extruder that I DIY made. As I am now pushing 86 you may not hear much more from me as time goes on. Have fun.

I too would like to thank you or you great work here!

I am wondering what you used to insulate the heated section, it looks nice and must not be melting.

I added pictures of mine today to the "I made one".

Thanks again!

Your are welcome, it is nice to see folks making their own extruders.
I used a welding pad blanket wrap around the hot end for insulation. Obtained on eBay.

Just wanted to say thank you! for all the hard work that you and Filip have put into this v5! :-)

I tried to change Mackerel to include the display of max and min filament_width. Besides adding the two extra variables and resetting them to 0 at the "clear stats" command I include the following code in mackerel_main


I would like to show those two values one after each other at the same place as average.. I saw that that the display of average happens at the "ultralcd_implementation_hitachi..." but have no idea how to control this switching between those three values. Someone else has an idea? Flip?

I include a new counter alt_cnt and then changed/include the following in the ultralcd... (around line 491 )

  if (alt_cnt<5){                               // FVD
     lcd_printPGM(PSTR(" Av"));
  if (alt_cnt>=5 && alt_cnt <10){               
    lcd_printPGM(PSTR(" Mx"));
  if (alt_cnt>=10 && alt_cnt<15){
    lcd_printPGM(PSTR(" Mn"));
  alt_cnt = alt_cnt + 1;
  if (alt_cnt==15)
     alt_cnt=0;   // FVD End

feels quit and dirty but works ;-) It will show Average, Max and Min filamant width, alternating between them every +/- 5 seconds

This is a nice idea. Can you make a full Mackerel code available. If I try to add this to the coding I no doubt will mess it up.

Took more time then expected to learn and understand git and github but if all is allright there is now a fork with the changes included : https://github.com/CFTechno/Marlin-Mackerel

I'm working on a version where a full audit trail for the complete extruder run is available either saved on the sd-card or send out via the serial port.

No problem. Will create a fork from Filips mackerel github and include my changes in it.Will try to make it available today.

wat is the minimum gear ratio needed on the nema17? would 1:5 be enough?

According to the manual you can adjust the settings when using an other gear ratio

I have finished printing all the parts and am waiting on a few last ordered parts to arrive.

Does anyone here have any experience coloring natural ABS with one of these? I am concerned that such a small melt chamber may not always have enough colorant pellets in it to properly color the plastic.

If it does not work well I am wondering if chopping or grinding the colorant to a much smaller size would help or would the finer partials cause problems?


Net yet ;-) But will be ordering some serious batches of colors next week :-)

Visit the filastruder forum, it's full of information. They tried al sorts of things already, even color in powder form but stopped using that. Also read that chopping the filament after the first run and re-extrude it will improve color mixing (if needed)

Very cool, thank you

Read them all and all their links, will defiantly try this.

I had 718 pounds of ABS Prime 555 donated to my project, hope it extrudes and prints sweet.

If the MFR of that "ABS Prime 555" is high enough you should have no trouble extruding and printing.

The LG ABS (43 g/10min) I'm testing is very nice both in extruding and printing. Only trouble for now is that I can't get the filament width fluctuation for the moment below +/- 0.1mm

Styron told me it is exactly the same MFR as the stuff FABB Filaments uses, and there filament is the best I have ever used so I expect the 555 to print nice.

Is +/_ 0.1 bad? seems like 1.65 mm to 1.85 mm would be close enough, no?

Are you doing this on a Lyman and if so do you have the width sensor from Flipper?

Yes I use the sensor from Flip and it allows to see the smallest and very short bursts of width fluctuation. If you would measure it by hand then the chance of measuring exactly at that position where the "blob" is is really small.

Filament width fluctuation is what makes the difference between good and bad filament. (other then bubbles and moisture) , filament fluctuation will make the print flow inconsistent and will show up immediately on the perimeter of your prints but will also (badly) influence the interlayer bonding.

I was talking 0.1 for my 2.8 filament but let's say that you get 0.1 for 1.75 so between 1.65 and 1.85. The surface of a circle PI x R^2:

1.65mm = 2.138mm2
1.75mm = 2.405mm2
1.85mm = 2.688mm2

That is around 11-12% volume change!


I see posts from people making filament all the time that have no sensor and they claim to make great parts, I have close to $600 in this, you are scaring me ;-P

For an example : http://www.soliforum.com/topic/9604/general-filament-extruder-problem-homemade-not-filastruder/ and look at the difference in printed parts in the photo he posted

The photo shows a printed part when the filament fluctuates +/- 0.3mm (2.2 - 2.8 mm or 22% volume change)

But this was simply him not understanding the difference between extruding PLA compared to ABS.
So his filament size problem was an easy fix. But you did point out something I was not aware of, I have seen this problem with filament I have purchased, to a much lesser extent, and had no idea what was causing it. That same filament would jam in my feed tube because it would get so big.

If these people can make good filament without a sensor, on an extruder they have to hang vertically from the wall, it does not seem like it should be all that difficult to do.

I see on Hlyman's latest post that the new software changes the speed of the screw if I am understanding him correctly, this is how big commercial extruders work. I have worked with big sheet extruders for many years and am very familiar with their operation. Adjusting screw speed is the way to go, much preferred over the use of the draw rollers, I am not sure what the new
"Filip's marlin firmware upgrade" does with the draw roller, it would be interesting to know.

He also states he is designing a new screw, the drill bit is quite different from what a commercial extruder uses, but I can not imagine that they would be cheap to make. I have considered using a wire feed welder to fill in some of the center of the screw near the extruder head to look more similar to what the big boys use.

Something like this: http://grabcad.com/library/plastic-extrusion-screw

The problem is the screw needs to be very smooth, in a big extruding house we polish them to shine like a mirror. It would take a very long time to clean up my welds to that point, probably not worth the time. The reason for the filling in is that you need to keep the pressure higher in the melting head than in the feed part of the tube. Otherwise you will get the pulsing he talked about causing his filament to vary wildly in diameter. ( He may have been almost boiling his plastic at those temperatures with PLA )

This can be overcome somewhat with a more constant feed of the screw at the hopper end, when co-extruding ( a process that puts a thin layer of a different plastic in or over the sheet ) the feed hopper is sometimes vibrated to keep the feed more constant. A second, much smaller extruder is usually used so it needs the extra vibration since a small change in material feed will result in a big pressure swing in the extruding head.

So if I do have problems one of the first things I will try is to add a soft flexible connecting hose between my hopper and the extruder and mount a vibrator to it. I purchased a small one for another project that ended up not being used so I have one on hand, it is simply a small motor with an off centered weight.

Anyway sorry to ramble on so much


  • Yes there was an issue with the temperature & extruding speed differences between PLA and ABS in the thread that I posted, so indeed an easy fix if you know the cause!. But it is a very clear example how important it is to keep the filament width fluctuation to a minimum, that's why I posted the link.

  • Lyman is talking about using the same sensor [flip's] that is used on the extruder and using it on the printer. That way the printer software (marlin) will adjust the plastic flow from the nozzle to compensate for the filament fluctuation. IMHO a patch to solve issues with the filament

  • The Mackerel software for the extruder in auto mode will control de puller speed in order to control filament width. Till now I could not get it to improve on the results that I have when manually setting extruder speed and puller speed. The reason is the pattern in which the filament width is fluctuating. The "blobs" are very short bursts, if the puller would have been in auto mode it would always cause worse results because of this. I will record them for 3 minutes and plot them, post the plot so you can see how short the duration of the blobs is.

  • There is already a special auger available. It's used on the filabot and is "only" usd 630 http://www.filabot.com/collections/components/products/filabot-original-ex2-feed-screw

  • With regards to people printing great parts: How great is great really? It depends on your standards. This is what I have printed on my Prusa I3 with my own made filament and using my own definition of great it's not great :-) http://i38.servimg.com/u/f38/13/37/93/48/test_g10.jpg

However some prints with less detail (plates to hold bearings for these gears) in 100mm/s and 0,1 layer height looked very nice! But testing the interlayer binding showed there were some issues.

Thanks for the Filabot link, at 5/8" it should work in a Lyman. $650 is more than I will spend unless the drill bit wont do the trick.

But in the future I would love to upgrade to it.

Has anyone ever put one of them a Hlyman?

Lyman got one I read somewhere in his posts. Never saw a review afaik

Hi Frans,

You talked about a "blob", I am wondering what you use as your extruder nipple? Are you using the brass 1/2” NPT plug with a serrated 1/4” serrated hose tip or something else?

It seems to me that a brass pipe plug with a hole of the proper size would help to eliminate blobs and filament size fluctuation, have you tried this? It seems to me the hose tip would naturally cause blobs, and more so when making 1.75 mm filament since the hole is so much larger than the desired output.

I read on one forum that a 1.32 mm hole makes the best 1.75 mm filament, I do wonder how they drill such a hole.

Thanks for you comments, I have learned a lot from you,

I'm using a similar thing as I could not find the one in the BOM.

I also read that the other extruders without the puller use a nozzle with a specific size in order to get the correct width, however as there is no other controlling mechanism they (the users) can only place their extruder higher or lower from the ground and use gravity as a "puller" to control the final width. I saw some pictures where the extruder was places on a table and on some extra boxes to get it high enough. There should also be no obstruction when it falls as that will influence the gravity "pull" and therefore the final width and fluctuation.

I will try both once I am up and running. I hope the last parts arrive this week.

It just seems to me the hole without the tube should work better, the over sized hole just seems like it would cause problems too.
The hole in my hose fitting measures 4.15 mm.

I think the tube would naturally "blob" up, as it will not fill evenly unless the draw pressure and the extruding speed are perfectly matched. I would think you would always have to over extrude rather than under extrude to keep the process going which would cause blobs. Just my thinking, not sure I'm right lol

Has anyone tried: #2133 DRV8825 Stepper Motor Driver Carrier, High Current from Pololu? Anyone know if they will work for M1?
They say 1.5 amp per coil without cooling and 2.2 with. I plan to add heat sinks, even if their the ones that came with the A4988's

I got an email from someone who mentioned that they used these and were successful with them.

thanks for the reply Flipper

hlyman I am ready to cut my screw/drill bit, I am wondering about distance from nozzle. I see you posted a while back that you were experimenting with distances and would post your results but i do not see an update. Do you have a sweet spot as far as distance from nozzle I would appreciate hearing you.

Thank you very much,

I unscrewed the brass nozzle from the coupler on mine to measure. The distance from the end of the screw to the end of the coupler was exactly 1 inch. Hugh put this together for me so its the master's handiwork. Hope that helps.

i have made a gearbox to enable the use of 2 nema 17 motors rather than the more expensive and under powered (it only gives about 80 kg of torque same as my gearbox) nema 23 i havnt put this into operation yet nor calculated stepper ratios etc (anyone good at that stuff want to message me please or skype me via facebook:dhudd3 yes that is my username lolz) instructions will be posted latter (most likely tonight ) but its pretty straight forward

nema 17 motor torque combiner and raiser
by dhudd3

I don't understand your "80 kg of torque" is that 80kg.cm (8nm) ?

The NEMA 17 (kaysan 1040220, 1A) in the BOM has 9kg.cm or 0,9nm according to the datasheet
The NEMA 23 (PHIDGETS 3333_0, 2.8A) in the BOM has 150kg.cm or 15NM according to the Phidgts website.

Main problem is that the ramps step stick only drive at a maximum of 2A. So there is some serious loss of torque imho.

FYI: In the filastruder info there is a reference to the beta motor and it's claimed to be 28nm.

the nema 23 is well uder powered as you stated so by using two nema 17s instead you can achieve the same result as most cheap steppers are actually about 1.8A you only receive half of the torque from the motor (150/2) plus 5 as its not quite half to be nice gives you about 80kg of torque same as my gearbox which provides 81kg
4.5kg (standard nema 17 motor) 2(as we now have 2 motors not one)) 9 (the final torque increase value(each set of gears increases by 3)) == 81kg cm

with more time i could probably bring the ration even higher (12 or 15 instead of 9) and sacrifice more speed as it isnt realy needed

my aim is to build the entire project from standard 3d printer parts (all engines nema 17 standard motors normal bearings (608zz) and the usual assortment of parts you would use on one nuts bolts washers threaded rod etc) we can easily run 2 engines for each the auger and the 9kg winder engine and save money (i don't no about you but i have a few of them ) a replace the winder with a single nema 17 for a total of 5 motors im hoping with some stl mods i can do this

hope that helps clarify anything this doesnt replace the nema 17 motor at 9kgs only the under powered nema 23

Yes, I understand, however there will be some losses because of the gears as those are not 100% efficient.

Do you think that the printed gears will survive the forces? I was surprised by those forces and the 3d printed coupler was not capable of handling them so it was replaced by a brass one.

the gears them selves are quite strong i have tried to break them by forcing areas to rotate while i hold one end no skipping or tearing of the gears also the gears may not be 100% accurate but they do turn extremely easily therefore you shouldn't have to many troubles when it comes to trying to achieve the rated output

rather than use brass did you try editing your print settings to allow more outside edges((vertiacal shells in slic3r)this will make it stronger as well as slicing it at a smaller size (0.2 or 0.1 rather than 0.3) if it keeps breaking just add a mm to the side or so shouldn't after that

im currently running the connectors i use on this model on my 3d printer z mount no issues `

the gears them selves are quite strong i have tried to break them by forcing areas to rotate while i hold one end no skipping or tearing of the gears also the gears may not be 100% accurate but they do turn extremely easily therefore you shouldn't have to many troubles when it comes to trying to achieve the rated output

rather than use brass did you try editing your print settings to allow more outside edges((vertiacal shells in slic3r)this will make it stronger as well as slicing it at a smaller size (0.2 or 0.1 rather than 0.3) if it keeps breaking just add a mm to the side or so shouldn't after that

im currently running the connectors i use on this model on my 3d printer z mount no issues `

Continued testing. I might be doing something very wrong but when the puller is at a fixed speed the filament width will fluctuate (2.8mm goal) +/- 0.6mm In auto tune mode it's even worse, sometimes as high +/- 0.8

I got same issue, width fluctuate too, try 1.75mm and got between 1.6 to 1.9, did you drill holes in the Barrel as specify page 4, I just did one.

Yes, Drilled the hole but that had no influences on the fluctuations

When using the filament PID, it has to be tuned and takes some trial and error. If its out of tune, it will be worse than not using it at all. See page 57 in the manual for some tips. I've found that the extruder can sometimes surge up and down, and it might be due to how full the hopper is, possibly air pockets in melted plastic in the screw barrel, etc. I've found 'priming' it, by pushing filament in through the hot nozzle while it is not running sometimes helps to ensure enough plastic is melted around the screw.

Can you (or someone else) provide me with the ID of the Brass plug nozzle 1/2" 1/4 hose or is the 1/4" already the Inside diameter? I could not buy via the mentioned source in the BOM so found something similar on ebay with a 4.5mm inside diameter. Maybe that's the issue?

I measured my brass nozzle, made for 1/4 inch hose and the ID is 4.5mm. I make 3mm filament, so a smaller brass nozzle might work better for 1.75mm filament.

Thank you for the info! I'm also trying to make 3mm filament but till this moment the fluctuation is +/- 0,3mm and +/- 0,1 depending on speeds (puller and extruder) and temperature combinations. Trying to improve insulation and will then do some more testing today.

I disassembled the unit. Between the nozzle and auger there was a nice blob of plastic with almost no air bubbles. The auger's first 3 cm was nicely covered in plastic en the next 1 cm was a mix of pellets half molten / half non molten.

I understand that the filament PID need tuning but what worries me is that with a fixed auger speed and fixed puller speed the fluctuation was that large. I tried different auger speeds (6.5-17) at different temperatures (195-220) and different puller speeds (0.70-1.50). Also compared sensor data with manually measured data,they were the same. Increased cooling. With more cooling the data seems more stable with no fluctuation for several centimeters at all (!) and then suddenly either a very small or very large blob will appear.

Extruder results during these test was between 3.1grm/min - 5.4grm/min.

The unit is reassembled and the space between nozzle and auger is made smaller to see what happens when running the same tests.

After reassembly did some more testing. When lowering the temp to 180C and the auger speed to 9rpm, puller to 0,5rpm. removing the cooling the result is between 2.9 and 3.2mm (2.9grm/min) or +/- 0,15 which is better but still not great.

With cooling the same settings will give me 2.8-3.0 and 2.6grm/min?? I really do not understand why that is.

It seems that any combination of auger speed and temperature resulting in more then 3grm/min will generate more fluctuation. Which leads me to believe (very carefully) that there is a pressure consistency problem. What do you think?

I'm in the testing phase and already fried 8 motor drivers (and one arduino).....ordered some DRV8825 stepper drivers, those are designed for high torque stepper Motors.

Extruding temperature seems very important to get high speed extrusion. Got as high as 30inch/min at 215C and no result at 175C

What is the collected experiences from using different augers?

I bought a 5/8" Dewalt Auger from Amazon, when I measure it only the cutting edge gives a hole of 15.9mm.

The rest of the auger is only 13.5mm in diameter, is this usable or should I return it and get another brand?

Hey everyone, I have been building this v5 and it has been alot of fun. Unfortunately I have run into a wall with the firmware. I checked my ramps board and it works with sprinter. Mackerel seems to be working on the board... but the LCD will only flicker (the back light, no letters). I have now tried two LCDs and am resorting to asking others. Please if anyone has run into this issue, please let me know what was done to fix it.

P.S. I have been going through the software... but I am not a software guy, I very well could have over looked something.

I had a similar problem. I believe its a wiring issue. I ended up cutting off the notch on the 10 pin wire ribbon and inserting it backwards on the board (only on the ramps, still 1 to 1, 2 to 2).

At this point I would be surprised that it is a software issue, since the code has not changed and has worked for others using the specified RAMPS boards. When you tried sprinter did the LCD work in that case? Also do you have a RAMPS1.4 and a SMART 2004 LCD Controller? I assume you have checked that the LCD cables are hooked up correctly? (1 to 1 , 2 to 2) Flickering on the backlight might indicate a cable mixup.

Thanks for the response flipper (and may I add the board you sent me looks great). unfortunately sprinter does not support an LCD feature. I can find a compatible firmware to put on the board with an LCD display. I am using a ramps 1.4 and I have tried both a SMART 2004 and a SMART 12864 (had is laying around and just hoped it would work) both gave me the same flickering screen. I have checked the wires again and again (which is believe is the definition of insanity) but I have no problems checking them again when I have the extruder in front of me later today.

Did you use the illuminated switches? Are you sure those are connected correctly? Do you have the same problem when you keep the heaterband and SSR disconnected?

Thanks for the response. I do have the rocker switches and they are operating properly. I also disconnected the SSR and tested it without the ssr. Unfortunately I am ending up with the same issue. I do not have my filament sensor hooked up right now, but it should still be responsive, right? The LCD is definitely getting power, it is just not displaying any text (also it is not the contrast, in a move of desperation I checked that was well).

Firmware willl work without the filament sensor connected. It does not need the filament sensor to extrude.

I must admit with the low RPM motor is having no problem turning so the firmware does seem to be working. This is why my predicament is so vexing.

Problem is solved. very embarrassing, I will never admit to what it was. Thanks for all the hard work that was put into designing this project!

Excellent news!

hi everyone just wondering what the interest would be if i posted a 1/9 reduction gearbox for 2 nema 17 motors essentially supplying the required torque and speed to turn it instead of using the nema 23 meaning you can use a standard printer kit

the stl is finished and i printed it myself seem to be working for the required job but is untested as soon as its tested i will post a number for anyone who would like to use it

I am interested.

Hi hlyman,

Do you extrude much 1.75mm filament with this? If so what are the tolerances? I see you say .1 for 3mm filament. Will this be as good or better than most filament sold online? I have access to very low cost ABS, less than $1 per pound and am excited at the idea of making my own filament. I have ordered the sensor and will be building one over the next few weeks, thank you for making this open source.

Hi Lyman,

Although i looked at the manual, I could understand the exact position and duty of the bearing. Why did you use thrust bearing and loacation of the bearing is just end of the auger bit or fixed to the Hopper Motor Mount part ? I couldn,t find it at the manual.

Thank you in advance.
Any one is welcomed to answer.

from my understanding its duty is to counter forces created by pushing plastic forward so there is not a constant push against the connector to the drive motor. I assume there is a ledge on the drill bit and it pushes against the hopper wall.

I can understand the bearing as a reaction force or kick-back protection. But i cant understand where it fits. Is it tighty fitted into the hole in motor mount part or elswhere ? I would be glad to learn it if anyone has any experience on that

it goes on the screw shaft and just rests/pushes against the hopper wall, it does not go into the hole.

Is it an idea to change the fixed "EXTRUDE_MINTEMP" in de firmware to a setting that defines the minimum below range that the temperature must have before the extruder can be started. So if it has been set to 15 then the extruder motor can not be started unless the temperature is at a minimum of 15 degrees of the set extruder temperature.

Example : I set the extruder temperature to 150C en this "minimum below range" is set to 15 Then the extruder motor can only start when the temperature has reach a minimum of 135C (if PREVENT_DANGEROUS_EXTRUDE is set)

EXTRUDE_MINTEMP is code that came with Marlin, to avoid running a 3D printer 'extruder' when the barrel was cold. So its more of a safety feature protecting the extruder motor, couplings, and gearbox from wear and failure. The risk with changing it from an absolute temp to a delta value is that it gets the user involved and the user could inadvertently enter a wrong temp (like 100C) and then the extruder would start, even though the temp is not high enough to melt plastic and the extruder could fail. Are you running plastic other than ABS? If so then maybe it should be based on the plastic type (and melting point) rather than user temp setting.

Using others plastics with lower extruding temperatures was indeed what was in my mind when I wrote this.

is Step-syn - 103H7123-1344 is ok for Extruder gear motor?


I did an internet search and that model did not show up. Is it a standard stepper? What is the gear ratio? The torque and gear ratio specs you are looking for should be close to this motor: http://www.phidgets.com/products.php?category=23&product_id=3333_0

Comments deleted.

First off looks great and seems to work well for you. I did notice that in the video you had a hand full of pellets that was not feeding in the hopper. I'd recommend putting the extruder in a 40 degree angle (extra angle should help push more into the auger) and angle hopper to accommodate the new extruder angle. Other than that looks like a well oiled machine. Well done!

Hi Mr Lyman,

What a great job about your filament extruder !

I come from France, and unfortunately, it’s not possible to order online the urethane drive rollers 3/8 ‘’ ID 35A on http://www.mcmaster.com/, (shipment to France impossible)

After a long time searching, the same product doesn’t exist on eBay or somewhere else,

so my questions are:

Do you know an equivalent product of your urethane drive rollers ? (with the shipment to France possible)

The hardness of your urethane drive rollers (shore 35A) may be changed by a harder one (shore 92 A for example) ?

best regards,

Someone has an European source? Because I ordered a few (8) and including the shipment cost, import costs and other taxes two rollers did cost me euro 61,98! or usd 79!...

Yes, could not find it. Website really broken, scroll doesn't work not in IE nor Firefox
Best link I found was http://www.misumi-europe.com/shop-uk/eu/ItemDetail/10300421250.html
Searched with "polyurethane roller setscrew" Could not find one with the correct hardness.

Thank you very much Mr Lyman

Hi Mr Luman
Thank you so much!

I have made myself one.
It is working fine, extrusion speed is about more than 60" in one minute (may even increase)
Only 1 small problem: The filament is not shiny ( polished), it's just blurry, no such filament is commercial.
But I have found a solution, it is not by Extruder machine, which is in the plastic (technical).
One again, Thank you so much

Nice, post a photo.
If the the filament extruding is not shinny or polished it is because burrs collect at the exit of the nozzle. I clean the nozzle each session. I get both, but it makes no difference to my 3D printers, they don't know the difference.

Hi Mr Luman,

Thanks for all the excellent info.
I'm about to start building myself one.

I see mention of a BOM in the instruction manual. Where can I find this BOM?

Best wishes

Read the manual.

Thank you Mr Lyman for the quick reply.

I am definitely reading the manual, but I'm not as clever as you.
I need some extra help please.

I'm looking at the "Heat band" and I found a description and picture of it on page 4 of the manual. It is also mentioned on page 7 and that is all I could find of it in the manual.

My problem is that I am not familiar with heat bands and I need to find something like that. All I can see in the picture is something with 2 wires and it is covered with heat resistant tape going all around the pipe hiding the heat band.

I would typically like to know the operating voltage, ( I assumed it is 120V and I would like to get one to work with our operating voltage of 220V) What is the heat band made of what is the width and length and thickness. What is the wattage it can handle and the temperature range it can work in, etc, etc.

That is why I asked for the BOM, because typically in a BOM there is a manufacturer and a part number. With that I can get all the specs and get all my questions answered. Then I can look for a local supplier and if one is not found do a international search for a supplier.

There is no BOM included in the files on thingiverse and no mention in the manual of where I can download one.

But I will read the manual again and see if I can figure it out as you suggested.

Best wishes

When you find the BOM's in the manual, let me know. Try pages 16 thru 22.

Thank you for your reply Mr Lyman.

I have found the BOM on the pages you suggested.

I had a problem with my PDF program not showing the tables below the pictures. It was just a blank area below the pictures.
I figured that out when I had a look at the PDF file on my android phone.

I have to get another PDF viewer. (I have the default one that came with Windows 8)
Thank you for your help.

Best wishes

I am making small batches of the filament sensor boards for sale.



Thank you Mr. Lyman!

Hi Mr. Lyman,

Excellent work, excellent design!
All is good, but without the Firmware this will never go... I understood that this is not final version and you may continue working, but even this version of Firmware will be interesting for all. Moreover, I think you will receive a big number of ideas and mods...

If this project is a real Open Source, could you please upload the Firmware too?

Thank you in advance.


See page 41 in the manual.

How far does the auger need to extend into the extruder pipe assembly? Does it need to extend through the melt chamber (the heater wrapped coupler) and right up to the plug or do you cut it so that it stops just short of the melt chamber? I would assume that extending into the melt chamber would cause a lot more heat to be transmitted back along the auger which might result in partially melted material gumming up the works.

For those of you in Canada who are looking to source parts, Lee Valley Tools has an auger that looks very similar to the Bosch one for about $22.

My first extruder which won the DFC had the auger about 40mm back from the nozzle. I have tried different dimensions up to 6mm from the nozzle. I am still working with it . My current dimension is 12mm back from the nozzle and I will be trying 25, 35 and 40mm back from the nozzle. When finished I will post my findings. The auger has not gummed up the works on my extruder at any dimension.

Filabot has a nice auger for about $17.00. I am using it, however I have to machine the end to fit my motor coupler.

Do you have an update for this?


Thanks Hugh, Belated congratulations on the DFC win.

I am looking forward to seeing your results. Out of curiosity, are you printing different length auger motor couplers in order to set the position of the auger?

Thanks for the Filabot information as the shipping to Canada is quite reasonable. The BOM mentioned ebay as the source for the auger and I have been seeing an insane trend there of $30 shipping costs for a $20 auger.

Not being my area of expertise, I thought maybe I should read up a bit on polymer extruders and discovered what you already knew, that that the screw/auger generally extends into the melt zone.


Hey!Lyman,could you tell me the material's name you put in the Extruder V5.I want to buy some in the internet~thx!

I use Enpol 8433 natural ABS. I bought 500 lbs as a sample order. Cost was $2.87 per lb delivered, phone, 770-441-5033. I am sure you can get pellets on eBay. They should be fine grade.

What is better in this version from version 2 of your extruder?
I mean, what justify the 600$ cost, respect to version 1 or 2 that was around 200-250$ cost?
I see the integrated spooler, what else?

Read the manual.

Hi! I have a question about diameter sensor. Which one did you use?

The sensor used is shown in the Manual, .created by Filip Mulier. See pages 17 and 33.

One with an open throat.

Simply fantastic !!! Thank you for providing all of the project and will make logical.

I'm in the process of building the extruder system. For the winder section is it possible to change the 12v dc motor to a stepper motor? By using an additional stepper driver on the ramps. Thank you.

You could use a stepper, but you would need to change the firmware to drive a third stepper motor. Currently the code to drive X,Y, and Z is commented out to improve performance. I don't think it would be a performance issue to make the change. Not sure if a stepper motor is the best choice though, since it is basically a constant speed motor. The best design would use a constant torque motor. One way to do that is use a dc motor with a constant current source (not constant voltage). That way the rubber drive will not wear out.

I understand what you mean about the torque vs speed. Would like to see how the stepper motor will work in the design. For the code what would need to be changed to use any axis. Sorry I know nothing about the code learning as I go. Enjoy trying different ideas sometimes it works, sometimes it doesn't. I appreciate your help. Thank you.

Your comment has been forwarded to Filip Mulier for a reply. I know that using a stepper motor on the spool winder would require a revised firmware coding.

In you recent video you show that you made a better variation of the filament spooling guide.
will you upload the stls?
i just printed those in your original file, and i ve seen that the round filament guide and the pawl is not so stable, thats why the lead weight is applied. However, is also goes forth and back, but the square version semms to eliminate this wobble...
One more question...
about the pawl: what is the thickness of the bottom (arc) part that slides on the worm gear? mine is 1,7 mm but i still have sliding problems with the gear..


The stl files are uploaded in a zip file.
You need to do more dressing of the parts, the thickness may be to thick or to thin for your worm gear.

May you please upload the extruder base without outriggers?

Hi?Lyman,would you please upload the STL file named CF-429 which in the sheet A03,I can not find in the ZIP pack.
Thank you ~

It is STL 8 in the ZIP pack. See sheet P01 page 23

Could you Send me the STL files named "A05 PART 4 STL 18 PED CASE" without the OUTRIGGER please?The OUTRIGGER trubles me,cause it can not fit in my pinters 200x200x150mm plantform.So I need the STL version without OUTRIGGER,my email address: chenran5130678@163.com
thank you so much~

The stl file "A05 PART STL 18 PED CASE NO OUTS" has been uploaded.

Thanks for your quick reply~lol!Best wish to you~

Why don't we move this to github.com so others can improve upon it? It seems like a much better forum for fixing bugs/contributing/managing versions. I think it'd be best to have one source of information instead of having to search through comments to find updated information.

Thank you again for your contribution, this machine is amazing

Is there anyone that has built this in europe? And if so can you help me build one myself letting me know where did you buy the parts and what parts you substited for the 240v ones?

Thank you so much, please, I would really like to build one like this, but its being a nightmare to say the least
Please help me

You could try a Filastruder, the power supply is 110-240v, automatically switching.



And let me know if you have success.

Comments deleted.

Thanks for sharing this great design. I am currently building one and I'm not seeing STL-4. Maybe I'm overlooking it, but if not, would you mind sharing it? Thanks again!

Thanks for noting this omission. I have added Part 19 STL 4 to the downloads.

Mr Lyman, I wanted to contact you because my employer would like to buy this machine. Is that possibble at all? Thank you!

Contact Liz Havlin, she is selling fully built kits based on version 4 and 5.

I do not nor will make extruders for sale, to old.
Check with Jebba (Jeff Moe).


See the new sensor design for the extruder and a printer, by Filip Mulier:


Filament Width Sensor Prototype Version 3
by flipper

Hi and thank you so much for your contribution

I started ordering a few things to bild this, but because Im in portugal, the costs are like 10x higher to import the pieces.
I have a lot of problems, if you could help me with some of them would be great

1st- Here the power is 220V, what do I need to change to make the machine, and maybe where can I get it
2nd- Could you maybe do a BOM for those of us who live in europe from european distributors?
3rd- I found a way to get the motor from phidgets, but will only receive it in 4weeks since the portuguese distributor of phidgets will only order then, and of course it will be much more expensive
4rd- I dont expect to pay less then 1300$ for building this because of living where I live, to give you an example, by now I only bought the sensor electronics, the board costed 15$, the components costed around 50$ with shipping included... instead of the 11.94$ you state on the BOM
5th- Could you or anyone help me out? Maybe someone could buy all of the materials for me, and ship them to me, and if it is someone within europe that would be even better, that way I wouldnt have to pay for customs and wait a month to receive the items

I will wait for your response before keep buying all of the materials

Thank you so much

Sorry, I am not selling or making any parts for any extruder. Check all the comments as there may be someone who is.

Hmm, I didnt ask for buying parts from you or making them for me but ok

Hello, how can I get in contact with the inventors? Does anyone have an email address? Thank you!

Contact only available through comments here.

Does anyone have an extra width sensor board they'd like to sell?

is a direct drive nema 23 motor not sufficient to drive the hopper? even if it is run at full current capacity using a bigger driver?

You question is forwarded to Filip Mulier for the best answer. However, you can always try it as the firmware motor speed is adjustable. Other than that I do not know.

I am not sure if you can get a nema 23 motor with enough torque. A guess for the torque needed is 50 to 90 kg-cm, or 700 to 1250 oz-in. The geared motor provides up to 150 kg-cm. You might be able to use a nema34 sized motor in that torque range. However the firmware (based on Marlin) provides a very short direction and step pulse, so other drivers may not work without pulse stretching circuits.

Im having a hard time finding exacly which pins to plug the width sensor in to. I see on page 46 of the manual showing the row of pins but not exact. would anyone be able to send me a pic or something showing exactly the pins +5v,v out and gnd go to


andre.mercado (at) gmail

Try page 42 which is a picture of the ramps, green to gnd, red to 5v, yellow to send. The 8 pins are group aux 2.

i see the AUX-2 block of pins has two rows. lets pretend that the top row is A and bottom row is B.
A1 | A2 | A3 | A4
B1 | B2 | B3 | B4

which pins +5v , v-out and gnd correspond to what?

i think its +5v = B1 , v-out = B2 and gnd = A1 is this right?

thank you for the help

First Aux 2 has 10 pins.
If you can read the diagram on page 42 then you will see that:
Green arrow points to grd which is A1.
Red arrow points to +5v which is B1
Yellow arrow points to sensor out B2 (v-out)

Thank you sir !

Very nice building. This is a important contribution to make all the plastic pollution a little bit lesser. Lets make these devices and recycle as many as plastic as we can. Thank you for this open source gift!!

Very nice building. This is a important contribution to make all the plastic pollution a little bit lesser. Lets make these devices and recycle as many as plastic as we can. Thank you for this open source gift!!

Gaah.. I have already almost finished an almost lookalike solution with a "self reversing screw" as it correctly is called (the wormscrew).
Now i can trash all my work, and focus on another solution i simultaneously have worked on to make simpler winder.
One differnce with this and my version is, that my screw have 3mm pitch to get correct winding with same speed on screw and spool.
And i don't use a tube for the guide to run over..

Good work anyways..

Good job. See the following video on my new level wind mechanism.


Publish a picture of your design.

Would replacing the A4988 driver with a DRV8825 increase the speed of the motor and therefore increasing extrusion speed?

"The DRV8825 can deliver more current than the A4988 without any additional cooling (based on our full-step tests: 1.5 A per coil for the DRV8825 vs 1.2 A per coil for the A4988 Black Edition and 1 A per coil for the original A4988 carrier)."

Once again...
i knee, before the teacher
thank you!

So I built one of these, and put motor drivers in E0/E1 as directed. Only the driver in E1 causes servo to run. I have tried swapping ramps boards and verified that both drivers are good. This implies that the Arduino Mega is bad. Before i replace, is there a simple way to change code in order to use a different port for the driver (such as X,Y or Z)?

Some basic things to check are - Do you have a jumper (or a switch) on the RAMPS pins normally used for the x end-stop (see manual)? Without this, the E0 motor will not run. Also, the motor only runs when the extruder is above 170C. If those check out I think the easiest way to make it use X, Y, or Z is to change the pins.h file. This file defines what mega 2560 i/o pins are used for each function. You should be able to change the pin numbers used to drive E0 to pin numbers corresponding to X. For example - see the section in the pins.h file for motherboard 34:

define E0_STEP_PIN 26

define E0_DIR_PIN 28

define E0_ENABLE_PIN 24

Note these numbers (so you have them to revert back) and change these numbers to the corresponding ones for X:

define X_STEP_PIN 54

#define X_DIR_PIN          55
#define X_ENABLE_PIN       38

And after a bit of hacking, it is working great. Have it extruding fairly well, but still fine tuning feedback loop. Next step...see about generating custom material filaments. I have hopes of printing out ion exchange resins.

Make sure your Mega 2560 is R3 and the Ramps is 1.4 version and they are not China copies. Try turning up the pot on the E0 driver. The M1 motor needs 2amps. Filip has had no problem with his ramps, mega and drivers, whereas I have had to change drivers, ramps and mega having your same problem using China copies. As for changing the Marlin coding that is Filip's department, so check with him.

Have anyone try mixing different types of plastic like PET to make thing printed with our own filament can be more flexible.

For people from Europe that are trying to source parts for their extruders: I am also building a similar extruder and I have faced the McMaster problem. I recommend AmazonSupply, where you can find most of the parts (I've bought pipe, fittings and bearings).

I found an typo error in the BOM for the Ball Bearing in the Level Winder. Item 6 on page 21 of the manual R1. The Bearing is MR608-2RS not the listed MR608-2R5. I have check it with Hugh and he asked me to post it

I added a picture of a sample print, printed with V5 filament. For those of you that would like a sample I would be happy to send one to you for small donation of $15 sent to my PayPal account "hlyman@comcast.net". This sample will give you an idea of the quality you can expect extruding your own filament. The sample is printed with a 0.50 nozzle with a 0.40 layer height.

I found an error in BOM for the heat band. The number from Zoro Tools is G0472516. This is a 120V heat band, but it also is available in 240V.

I am looking to try and cut the cost down a little and the main cost is the stepper motors, I was wondering what kind of torque I should be looking for to, while maybe slower, still keep things working? That is unless someone knows where to get some geared stepper motors for cheap.

The only way I think we could accomplish this would be to all pool in and buy in bulk.

Hugh, I have uploaded a fairly accurate model of the Bosch auger you used. I hope you find it useful.

Bosch 5/8" Auger

That's nice. I am going to try the auger bit that is sold by FilaBot.

What's your experience for using your own extruded filament
1.Does it feel the same as the commercial ones?
2.Is the diameter of the filament even?
3.Is it more bristle or more flexible than the commercial ones?
4.Have you try mixing different COLOR pellets and what's the result?

  1. I use only my own extruded filament and have now since my first version was published.
  2. Yes it is smooth as commercial grade. However if the nozzle builds up dried melted pellets at the nozzle exit the filament becomes a little rough, but prints the same. The diameter does vary more than commercial filament. To get it to commercial grade requires a resizing die or one of these:
  3. It depends on what pellets you use. I use Enpol 8433 Natural ABS pellets, which melts at a little higher temp than some others. It is used by commercial extruders.
  4. I use color pellets to add color, 2%. It works fine, but it takes time for the barrel to clean out the color when you change colors. I made separate Sub Assembly Extruders for each color I use and only have to plug them in.

The Manual has been revised and uploaded as Lyman Mulier Filament Extruder V5 R1.

In the description of Part #15 you state "This is a Nema 23 stepper gear motor, 1:15 gear ratio, 2.8A, 150 kg-cm
torque. It will do 116 rpm...". However in the parts list the motor is listed as a 77:1 gear ratio which the vendor describes as having a max speed of 25 rpm.
Could you clarify?
I would like to experiment with increasing the extrusion rate, so if the 15:1 motor will work I would prefer to have the extra rpm available for future experimentation.

As always the errors are beginning to show up. The correct motor # is 3333_0. The 3334_0 motor is one that we tried and changed to the 3333_0 motor and I got them mixed up. Thanks. The fastest I am running with the motor drivers on the ramps is 18rpm which is extruding around 10 - 11 mm/sec. I think Filip is running up to 25rpm on his machine.

Keep up the great work! Bravo!!

Thank's to both of you for another great design!

Hello dear Mr. Hlyman, what about arduino firmware for this HW? Where it is possible to find?Moreover, what manual you are recomending to read? Where it is locating?

Can the filament sensors sense clear material. I have a need to print more transparent items than what the ABS can offer. So I am considering to make PET or even acrylics(may be too bristle though) filament. Also what's the max temperature you have run your machine at reliability? BTW, it's very detail and refined work.

I think the sensor would have trouble with clear material, since it needs a shadow edge to register the width.

I suspect that it will still cast a shadow. Just because the filament is transparent doesn't mean that the light travels straight through it. The light gets refracted and reflected. I have been thinking about using a laser light source to help sharpen the edge definition. I would be interested in hearing opinions on this idea.

I tried extruding PLA transparent filament yesterday. The sensor did not work correctly for it. As I have not experimented with PLA enough to advise how to extrude it. It is a work in progress. Printing with the PLA I extruded I got one good part and one bad part. There was too much variance if the diameter of the filament. It will take time for me to determine the speed and temperature settings to correct this problem. Also remember that the extruder will work the same as V4.1 without the sensor.

A masterpiece.. It will a long one for me since i have to either buy in steps from the us... or source them in europe
but have allready ordered the parts from mouser... the freescale chip is on backorder until october...

I found the freescale chip in stock at Digikey
http://www.digikey.com.au/product-detail/en/S9S08SG8E2CTG/S9S08SG8E2CTG-ND/2186853http://www.digikey.com.au/prod... for Australian dollars or
http://www.digikey.com/product-detail/en/S9S08SG8E2CTG/S9S08SG8E2CTG-ND/2186853http://www.digikey.com/product... for US Dollars
is there anybody willing to buy from McMaster then ship to me in Australia I'm after urethane rollers

Paypal me and I'll ship it to you

Not to mention the fact that McMaster Carr doesn't export to just anyone...
"Hi Björn,
to the complexity of U.S. export regulations, McMaster-Carr accepts
international orders only from our established customers. This decision
also applies to orders shipping within the United States, because it is
based on the final destination of the items. We cannot accept this
order or future orders.
Your order has been cancelled."
That sucks... But wait, are there any "established customers" around here?

Got the same bull crap:
Hi Bo,
Due to the complexity of U.S. export regulations, McMaster-Carr accepts international orders only from our established customers. This decision also applies to orders shipping within the United States, because it is based on the final destination of the items. We will not provide a quotation or accept your orders.

i mailed them this before trying to do an order:
Hi, i'm located in europe and wonder if you can be convinced to accept and order IF:

  1. You dont have to deal with any export forms at all
  2. I arrange the shipping with UPS, including any forms to fill out
  3. All you have to do is box the order and wait for the driver to collect

yep that sucks BIG time.... it sucks as much as we dont get an mail when a comment is replied to
btw... the reason they give are pure BS... there are no regulations on the parts in the printer... they can be exported out the us without any problem. And they can just ship with UPS, in that way they dont have to deal with export forms themselfes, UPS will handle it and they do so while its in transit. and you do not have to pay customs up front, they send you a bill about 7-14 days later.
I gues they have enough orders that they dont give a "beeeep".... This is another BIG reason why i have cried for a kit, of course with a kit i cant do it in steps but have to buy the whole machine at once...
With UPS i could even arrange the shipment from here... all the seller would have to do is to box stuff up and wait for the driver to collect... they would not have to lift a finger.

Independent of why they choose not to export I'll rather give my money to someone who is happy to get the sale so I'm looking for European sources / redesign what's necessary instead.
I'll have to figure out another way to measure the thickness of the filament anyway because my current supply of PLA pellets will make transparent filament unless I add a colourant.

fedex was by my door the other day with all the electronic bits
printer is running... now we just need to source the parts on this side of the atlantic.. ie in the EU would be best.
to some extent i would also be happy to do kits, but in form of batches of say 5 or 10.. people signup and wait. when there is 5 or 10 they pay and i start to make the kits.. something like that...
but we need to find parts.... and i tried but failed at half of it

yeah, if you find all the sources for the extruder let us know...
i found the source for the electronic bits that is used to messure filament...

I have found a way of getting McMaster Carr to ship to me in Australia. I just check the McMaster Carr web and I have Shipping Tracking Information yippy. The way I'm doing it is to use an on follower so it give me an US Address. Here a coupe I found http://www.myus.comwww.myus.com, http://www.reship.comwww.reship.com, http://www.shipito.comwww.shipito.com, http://www.comgateway.comwww.comgateway.com just to name a few "Do Your Research". So now I have all the bits to build it now just waiting on my 3d printer to turn up now.

what was the price for the bits before shipping on?

When I check on the tracking information UPS saids "The shipper has
requested a delivery change for this package. / The package will be
returned to the sender. So when I contacted McMaster Carr I get the response
to the complexity of U.S. export regulations, McMaster-Carr accepts
international orders only from our established customers. This decision
also applies to orders shipping within the United States, because it is
based on the final destination of the items. We cannot accept this order or future orders.
We have canceled your order and will credit your credit card.
See that bit about orders "shipping within the United States"
I guess some people in the USA Can't order from Master Carr if they think their items will be on shipped out of the USA.
Back to square one

This looks great wish I had the money to make one :( looking at the BOM could you sub this for the kaysan one http://www.phidgets.com/products.php?category=23&product_id=3329_0http://www.phidgets.com/produc... while it is $9 more it would save me $20 in shipping and looking to save every penny possible right now. Again great work and thank you.

This gear motor will work fine. The Psteps in the firmware are set for a 99:1 gear ratio as default. As the gear ratio for the Kysan motor is 14:1 ratio the Psteps need be changed 375.3 which can be done with the LCD in the control menu, see page 50 in the manual for how to change Psteps/mm.

Ok thanks, I guess I should ask a broader question then and ask what would you recommend for some minimum torque ratings for the stepper motors? I know going cheap isn't great but that is all I can afford and I can slowly upgrade from there wit