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


Delta-Pi Reprap 3D Printer

by MikeP-NZ Feb 24, 2013
Download All Files

Thing Apps Enabled

Please Login to Comment

Is this design already on online in OnShape? Is there a .step exported version for easier import in FreeCAD? I would prefer if open source projects used software that can also be used by the general open source community ?

how much does it cost to build

i wish someone can make an file to print with this.. http://www.thingiverse.com/thing:208458 many similar design for a kossel base using the t-slot design, just modify the 19mm(3/4) square tube to fit in an angle for this delta printer.. it would eliminate the wooden base design.. free hand drilling worries me of not getting the incorrect triangulating angles..
Found this one..http://www.thingiverse.com/thing:34717 But the square tube is for 15x15mm

Kossel frame 2020
Arn Kossel Mod
by heyarn

what is the size of the max print size/bed?

I just wanted to drop by and comment on this printer because I built one about a year ago. I have made a few minor modifications to the hot end hub and rods but otherwise it is to spec.

It took me a good 6 months to calibrate it and learn enough about printing to get solid quality out. Now that i'm all dialed in though, this has been an excellent machine. I print something almost every day just for fun, and I do a little business on 3D hubs from time to time. My hot end is an E3d V6, and the highest resolution I've gone to is .05mm!

I really appreciate the work you put into designing this, and the support that I got when building it. Reddit was definitely helpful, but I had some important questions answered here as well.

Overall I love this thing. It isn't as pretty as a name brand printer, but it's something I made with my own hands and it was very rewarding.

Hi.. Can you help on what max build size this has?

I’ve dismantled this printer. I’d recommend building something more mainstream, but the build volume was about 8x8x14 inches

Hi there Mike,
I really like the options you decided upon when designing the Delta-Pi. I will be building one in February now, with a few small changes: first, I am going to try to use a kossel effector, and that will require a couple of changes to the carriages to account for the smaller spacing between the arms. Second, in my case I am quite happy with the print area of the MK2B heatbed, so that's what I intend to use. And third, I may change a few dimensions by a few millimeters such as the height of the columns and the distance between them (and consequently the length of the arms), but nothing very extreme. And yes, I saw the pictures with the Traxxas 5347 ball joints, these I ordered already as well as carbon rods.
So: thanks for sharing your nice design!

Great design. With a flaw. 30 BB623zz are quite too much.I don't have Solidworks, but you can take a look at my angular rolled that can save 4 bearing per axis and save other 3 for the idler pulley using a printed idler with only one BB.
ANYWAY... I'm printing one just now.

I don't think the use of two small 623ZZ ball bearings per idler pulley is a flaw at all, it's just a design choice. I am also quite comfortable with using eight small 623ZZ ball bearings per carriage, perhaps you are forgetting that these small ball bearings can cost as little as $0.12 each.
"ANYWAY" I took a look at your printed corner roller which uses a single slightly larger 624ZZ bearing and it would require a complete redesign of the carriage. I am not even sure one can build a simple loading structure for your roller and there is also the issue of printed rollers not quite having the same precision and wear resistance as steel ball bearings. All things considered I think your corner roller will remain in the domain of "unproven concepts" for quite some time, whereas the Delta-Pi carriage design is already proven to work.
Posting a comment and calling a proven design "flawed" as you did is not only incorrect but rather impolite too.

Happy to hear that u can find bb at that price. Here in Italy medium price si higher than 1.5€ each. Didn't understand why u become so angry. Is open source as u remark: u have to accept that someone could not agree with some design choices. It's a fact. Said that, I think that this printer is a really good choice vs others designs using t-slots aluminium extrusion bars.
About my rollers... I'm using them in a cnc milling machine, printed in taulmann nylon, so they are proven to work.
As a not polite person I invite You to be a little more polite, answering to your followers and to criticism in a more soft manner. It pays more than u can expect.

The great thing about open source designs is you are free to explore any modifications you like. That's how great designs evolve. People try different things, some work, some don't.

I'm sorry. I obvious use the word "flaw" in a non criticism sense. I think this is a great project so far, and as I said, the only one I put some time on, as printing it as it is, looking at the project and (as a designer) taking note of what can be improved and taking some time to figure out how to optimize costs, reduce pieces and make this beautiful printer a little lighter in weight.
Feel free to contact me here or via mail: cristianogariboldo@gmail.com

Hi there,

Quick question: what's the rough build size of this machine?


All of the parts are trickling in from around the world. I will be building this soon!

Check the latest Flickr images before you start.

Welp, it's too late for me to go ball and joint right now. Anything I should be concerned about?

If you are going with the original design then the make sure you don't drill out the holes for the pivot bolts in the arms. You need to screw the bolts through to keep the linkages tight and free from play. But don't do the bolts up tight because the linkages need to be pivoting freely without too much friction. They will last quite a while but too soon you'll find that there is play appearing in the joints and it becomes very noisy when that happens.
The move to ball joints is very easy. The carriages and hub are the same.

I was wondering if you could assist me in getting my RAMPS board firmware set up. I've tried to use your configuration.h file, but I can never get it to compile?

Your best bet is to run my config.h file through Repetier online editor. Take some time to read and understand what the selections mean. There are many possible combinations of hardware and settings so it is rare for one person's config.h to work for another person without making some changes. If you are still stuck then try asking specific questions on one of the forums. Questions like "Can you help me?" generally get ignored.

Okay I got everything in order. This weekend is build time!! Question, how fast can this thing print?

Maximum print speed is more a function of the hot-end design than the rest of the printer.
I usually print at less than 80mm/s. Not to say it can't go faster, I just prefer quality output over speed.
Rapid moves of 300mm/s are OK.
Faster than that and you run into the limitations of the pulse frequency that can be output by an Arduino Mega.

thank you very much MikeP-NZ

please can you send me updated photos of delta pi you have armed

I don't have time tonight but I will take some new pictures over the weekend and post them to the Flickr site.

I've added new photos of the Traxxas ball joint arms. The printed arms work fine for a while but they are tricky to get right and they wear out too quickly. I am also using FSR touch sensing to level the bed and have a new filament drive.

good morning, the carriage may have a 20mm x and y axis movement exes?????

There are 2 versions of the carriage files. One for 3/4" square tube (19mm) and one for 20mm square tube.
Depending on your country you will probably find it easier to get metric or inch size tubing.

how difficult is it to build a Delta-Pi with a larger build surface by spacing the sliders further apart? The rods have to be longer, but apart from that, it seems that there are just a few values to change in the firmware config.
Do i overlook something here?

That's pretty much it. Make the arms longer and space the towers further apart. For every mm you push the towers out from the center you should add 2mm to the arms length and if you want to keep the same build height add 1.7mm(sqrtof 3) to the tower heights. Don't just scale all the distances.

Good morning, I'm from Argentina and decided to build a delta pi ... and I printed all the parts and the assembly is almost finished, but I would like to use a firmware already tested.
leave my email so I can send photos of the impressions achieved.
from already thank you very much

Great to hear that. I am way overdue to update a few things here.
One thing you may want to do is to use the Traxxas 5347 ball joints instead of the printed linkages. The printed linkages are fine for a while but I found they start to wear and the printer becomes noisy.
To convert you need the following:
1pkt (12 pieces) Traxxas 5347 Ball joints.
12x M4x25 grub screws
M3 threaded Rod 12pcs x 80mm long
M3 nuts and washers.

Comments deleted.

Hi yblock, I'm pretty sure the STLs are correct for a NEMA17 motor.

I was trying to stuff the motors inside of the mount. haha. Your pictures have been very helpful

Hey guys,
can someone please upload the firmware.
my printer is completely've only big problems with the FW

I have added my config.h file from the repetier firmware version 0.91
This doesn't do automated bed levelling but all adjustments are now in EEPROM which makes calibration so much easier.

Great thanks, it was a great help :)

someone has already tested the Merlin-hotend?

Hey Mike (I may as well direct these straight at you)

So ive been trying to calibrate the bed height and I’m running into some issues.

Ive setup some Scripts in Repetier Host to move the hub in front of each tower so I can dial in the endstop screws.

The scrips I have look like this:

Script 1

; Z Tower


G0 X0 Y-80 Z0 F8000

Script 2

; X Tower


G0 X60 Y40 Z0 F8000

Script 3

; Y Tower


G0 X-60 Y40 Z15 F8000

Script 4



G0 X0 Y0 Z0 F8000

Do these look right to you? They
seem a little low considering the printable radius should be around 125mm.

And the main problem Im having is that if I calibrate the printer to the correct height in the center the outside 3 points are about 15mm below the build platform and vice versa if I pull back the z max height in the EEPROM settings by 15mm the outside point can be calibrated properly but the center is above the platform.

I read that this has to do with the DELTA_RADIUS in the firmware but changing this just seems to mess up the movement.

As always, thanks a lot for all your help, I feel like im having every problem possible but on the plus side im learning heaps about how not to build a printer. :)

you are in the right track. your delta_radius may not be equal to 125mm due to build errors. I use 126.75 which gives me a little bit better movement.

you may also visited this link a few times but here it is: http://minow.blogspot.comminow.blogspot.com

it is a boring and long process but if you follow it to the smallest detail you will succeed.

also you can check the joints. try to move the hub when your motors are holding torque. if the hub is able to move and hold position without motors moving then you have bad joints. this was my issue.

yeah i have been looking at that website and ive been trying to raise the Delta Radius to lower the printhead. i was doing it in intervals of 10mm but that didnt seem to lower the center at all it just seemed to throw off the coordinates of other points , would changing it by 1mm or 2mm really fix an offset of 15mm?

The delta radius value has no effect at the center. The effect is to raise or lower the outer points.
You set the center by adjusting Z_MAX_LENGTH first then adjust DELTA_RADIUS so that the height at the edge is correct.
Don't forget that depending on which firmware and which version you use one or both of these values from configuration.h may be overridden by a value from EEPROM

If you have 15mm height difference then there is something wrong with the values you have entered.

Yeah, i've been making sure to change the Z max length in the eeprom and not the firmware, partly because the firmware doesn't seem to want to let me turn the eeprom off. Ok well i'll do some measuring and tinkering and see what i come up with.

each time you change delta radius, you need to change max z height (or leave it the same and re-do the endstop calibration. that blog suggests changing endstop values)

I started with an interval of 120 to 130 and narrowed it down until I found the value that works. My largest offset was around 6mm. lastly I found a bed-leveling test stl which covers the whole bed and made my final test.

an easy way out could be to use a z probe with johann's marlin. It works but repetier is much better at delta movements so I don't recommend.

hmm, well i sort of got the 3 outside points and the middle level but this required upping the printer radius from 175 to 250 which offset the delta radius to 199 which is causing all sorts of movement problems and isnt even staying level constantly around the board.
So im thinking the problem is elsewhere, im just not sure where is would be. I dont think i did many major changes to the firmware when i got it off Mike all those months ago but the 15mm drop off seems a bit too consistent for a hardware issue

Your issue sounds very similar to my old issue. Do a G1 X0 Y0 Z1 then G1 X90 Y0 Z1 and again G1 X0 Y0 Z1. If the hub does not return to origin using the same path it used before, then you have a problem with your joints. You can adjust screws (I am not very good at this, look for Mike's suggestions in older comments) or do as I did and buy Traxxas joints (or some rod ends or magnetic joints).

You can try 2 other things:
-See if the hub can move without carriages moving. At a specific coordinate it should only hold a single position. If the hub can stabilize at several points you have a joint problem. This produces surprisingly consistent errors.
-See if the hub is parallel to the bed at all times.

can someone give me the the measurement for the printable radius to put into the repetier host. ive got the aluminum build platform off of the website but the printer obviously doesnt go right to the edge so im not sure where to measure to. Thanks

It should be 125mm. All that setting does is set the size of the grid on the screen.

Can you please release the source of the new hub? I want to edit it to make a mount for e3d hotend. Currently I have 20mm long spacers and a mounting block which makes the heater block almost in line with the hub (I have glass fiber-silicone tubing to save the hub from deforming. It doesn't even get warm). This way I save some z-height but more importantly I use the fan mount that exists on the hub with a much quieter fan. I have been printing all plastics parts because the current ones I use are very weak. I will replace them all with stronger parts so I can print a new hub just for the tidier look.

By the way,
My printer's total printing time has reached 3 days. I am getting amazing prints. Lowest layer height I tried was 50um. I am almost certain that it can print at 25um (maybe 10 too) but I don't want to spend the time to try it.

Hey guys

I was hoping someone would be able to help me with a problem i'm having trying to set up the programing on my delta pi.

So i had not worked on it in a while and when i booted everything up its having a strange issue when I tell it to return to home.

I made a video to better illustrate http://youtu.be/fT0F6iWNuUAhttp://youtu.be/fT0F6iWNuUA

When i tell it to return home, the X and Z towers seem to move as if they have already click in there limit switches while to Y tower moves up for a longer distance but will still stop before hand if the distance in too long.

I've had this problem for a while now and but it did move alright before this and i havnt changed any of the hardware, im running Repetier firmware and Repetier host.

Also im finding that the Ramps board is running very hot, i've put a rather large computer fan on it but the heat sinks still get hot to touch after a few minutes of moving the carriages up and down.

if anyone can offer advice or some firmware alternatives it would be greatly appreciated.

Hi Stuart,
A couple of things spring to mind. Make sure you separate the end stop wires from the stepper wires. Twist the end stop wires together along their length. Twisted pairs are a lot less susceptible to picking up noise.
Make sure you have pull-ups enabled for the end stops. and wire the end stops normally closed.
These will all help to reduce the chance of false triggering of the end stops.
If you're RAMPS is getting hot try reducing the current to the stepper motors. If you haven't used a voltmeter to set the Vref of the drivers to a known value then this is a good place to start. Remember the current rating on a stepper mortor is the maximum allowable current. You don't have to set it that high.50 or 70% of that is often plenty of torque and makes the machine quieter too.

Mike Paauwe

It's 125mm. All that setting does is set the size of the circular grid on the screen.

Well that was a lot simpler than I expected, I had the endstop wires and the motor wires wrapped together. After unwrapping them they are working fine again so thanks heaps for that.

As for the drivers, the trimpots are all set to about 1.25v the motors are 1.8A. I do remember having to turn them up to this voltage because when i first plugged them in, the motor didn't have enough strength to push the limit switches.

do you have a way or a trick for setting up links? they are either too loose and the hub moves a lot or they are too tight and the hub does not move parallel to the heatbed.

I am considering buying traxxas joints.

This is a vital piece of information and I am sorry if it comes too late. I thought I had written it in the instructions but it appears not.
The key is "don't drill the holes in the clevis or the pivot arms any larger than 2.5mm. The holes need to be a tight fit on the M3 bolts. You should be screwing the bolts through the plastic. This tight fit is critical to the printer working accurately.

Well, It is too late for me. Instead of reprinting them, I just ordered traxxas rod ends and new carbon fiber tubes. I am guessing that I will be able to mount them to my current setup (same hub and same carriages). Otherwise I will be drawing new parts (which will take a long time due to my abilities)

You should be able to make the Traxas joints work. The key thing is that the balls must be spaced apart the same distance on the carriage and the hub. Push them put wider if you need to get clearance from the hub.

For anyone considering traxxas joints:

They are working really well. There is no need to modify the design. You can just mount the joints to the outer side of the link assembly holes. However I added 3 of those small M3 washers to both the hub and the carriages to obtain a little more clearance from the hub.

I had to take out the anti-backlash springs because they were too strong. They were causing a lot of noise in the joints. I will get some rubber bands or at least looser springs. There is no measurable backlash now but I just want to be sure there won't be any in the future.

I guess if you assemble the original design correctly you don't need traxxas joints.

How tight should the carriages hold on to the towers? Currently, mine slide off around 3cm down and stop when i release them. Nothing is connected to the carriages yet. Feels like they make too much noise.

The method I used for setting the tension was to press my thumb against one of the outer bearings and roll the carriage. If I could easily stop the rotation of the bearing and cause it to slide on the tube then it is too loose.
You don't want excessive preload on the bearings. without belts or arms attached it should have very little rolling resistance.
It should certainly fall under its own weight when vertical. I haven't checked but I would expect it to move under its own weight with just a shallow incline. Don't forget that there needs to be a washer between the bearings and the carriage. and the washer size given in the BOM is important. Larger washer sizes may not work properly.

Thanks for answering.
I follow your drawings as close as possible. The problem is my parts were printed with 3 different printers and 5 different plastics so it is hard to get the dimensions right. It seems like I need to increase scale a little bit because the carriage cannot slide more than 5cm under its own weight.

That doesn't sound right. My carriages had a bit of play on the tube before you tighten the preload screws
No part of the carriage plastic should touch the square tube. Only the bearings should touch the square tube.
The bolt holding each of the 8 bearings in the carriage should be tight. Before you fit the carriage on to the tube check that each bearing is completely free to spin without rubbing on anything.

I set the scale 98% and printed one. All 8 bearings touch the tube all the time. It won't fall under its weight but it is very easy to get it moving, certainly not a problem for the steppers. Then I tried 99%. It slides way easier almost like free fall but only 6 bearings can touch the tube at a time. It feels very loose. I like the first one way better. It seems like a need something that is in the middle.

You should be printing at 100%. The carriage should be loose on the tube until you do up the preload bolt. The preload bolt flexes the side of the carriage through the live hinge to control the bearing pressure. If you print at 98% then the spacing and offset of the arm mounts will be wrong too.

I decreased the minimum number of perimeters and set the scale to 100%. Now everything is perfect.

By the way do you have a good idea to mount e3d hotend? I was thinking this thing but with some spacers to make use of the fan mount in the hub:

Only thing is the heater block comes too close to the hub.
I really don't want to print a new hub as it takes too long to print one.

8 months have passed since I decided to build this model and I still couldn't print all the plastic parts (motor mounts and carriages left). All other parts have been lying on the ground for the last 6 months.

I reckon you still use your printer. How is it performing? Can you provide some more print videos?

Also out of curiosity, how fast can it print?

Development is stalled a bit on the prototype due to lack of any free time. I hope to have a big push over the Christmas break.
I've got bed probing working on the protoype. But haven't had time to post updates. There are a few others who have successfully built and are printing with theirs.

Is there anyone in the US that would print the parts for me and ship to oklahoma?

A bit late, but I could print them for you. I have Red, Yellow, Orange, Blue, Black, White, Pink, Translucent Green, Translucent Brown, Natural PLA, and Clear.

The translucent brown prints best. I'm currently printing these pieces for a friend in the Translucent Green / Brown, so I'll post a picture of how it turns out in the makes section soon.

yay new things, i really need to get around to finishing this, holidays are coming up soon so hopefully i'll make some more progress.

Would anyone be willing to print off the printable parts for a price?
if so, how much would you ask?
Parts would need to be sent to Sydney Australia.

Would print my own, but dont want to waste anymore money trying to get my solidoodle to work.

I could. :) Will message you.

I've printed just about everything bar the central hub. I can't seem to get this to slice and print. Is there an error in the published file, or could someone send me a different copy to try.


I also got problem printing the hub. I found the hotend just moved around and printed nothing but air. When I opened the file with some STL tools I couldn't see the object as if the object was floating somewhere I couldn't see. Anyway, I was able to fix it using Autodesk software and the print came out nicely. I'm not sure you have the same problem I have but just let me know if you want to try my file.

Mike, decided to live dangerously, and build one. Getting the parts from the UK. Only one change I can see, use electronic end stops rather than micro switches, Can you point me in the direction of the changes for the firmware, and which version you are using. Arduino mega with Ramps 1.4 will be in control, with a 2004 display and SD card reader, PSU to be decided, and e3d-online hot end. tube will be 20mm, and I may make it a bit taller than the "standard", to allow some wiggle space. Also looking at a possible ait guide to go round the hot end, to help with local cooling.

Can you put something up about how you calibrate the unit, and the best piece/s to use for test printing. Are you using a glass or aluminium bed, and is it heated? What levelling have you put in, and possibly most important, how's it performing?

Longer term, I'm sort of hoping it may be possible to have a multi filament hot end to allow colour printing, but that will only happen when the basic unit is working.

Anything I need to watch out for?


(Based near Dublin in Ireland, and 3 D printing is still pretty new here)

Hi Steve, I added a couple more photos and a link to the Flickr photo set where there are even more photos.
I have added bracing to the back corners and turned the machine 180° round so there is a single tower at the front and 2 towers at the back. The firmware I am using is Repetier but I am probably going to switch to Marlin and implement the bed leveling probe. Bed leveling is a major step forward for this design and I would recommend you incorporate it from the start.

Mike, Many thanks for the rapid response and update, Strange you should mention the bed levelling probe, been looking at the notes on that earlier and deciding that it would be a good thing to do. Also note your comments about repetier and Marlin, perhaps I might ask you to put up a suitable config file at some stage as a starting point for people like me. Will post on progress once it starts happening.

This may be a dumb question, but how did you mount your build platform? I Can't tell what it rests on or how its leveled.

I used countersunk head bolts that are sunk into the top surface of the build plate in 4 places. There is a spring between the build plate and the base board and a nylock nut underneath the base board. I adjust the build plate so that all 4 corners are exactly the same height.

your design looks great, now i just need to find someone to print out the parts and i can start working on my own

if it's any help, Hugh at http://www.3distributed.com/http://www.3distributed.com/ is going to be printing a set for me for 20 mm tube in the next few days, and I'm going to be starting construction here in Ireland within the next few weeks, once all the bits are here. I will be using a hot end from http://e3d-online.come3d-online.com, and the only change I can see at the moment will be to use electronic end of movement rather than micro switches.

I spent quite some time looking at different deltas, and the concept of this one looks so nice and simple in terms of the mechanical side, I am reasonably happy that getting it to print accurately should be simpler than on some of the other printers I've seen.

Hopefully, I won't be wrong.


In my opinion is the best design I have seen for a 3d printer.

I have to say congratulations on a superb design. I am most impressed by the linear bearings. I have been trying to create bearings like the ones you designed but your design kicks mine in the ass. I love how the bearings tighten, very elegant. I will now be adopting your design in my work. Thanks

Thanks for the kind words

Since you released this on february and had some time to tinker;

What is the largest (good working setting) print dimensions you can obtain with the design? :)

Since releasing this in February I've had very little time to work on it further.
The recent development of automatic bed leveling routines built into the firmware should make this design super easy to assemble and calibrate.
I've added extra panels to the back corners to stiffen the structure but other than that I am still using these same STL files on the prototype.
Check the latest flickr photos to see what I mean.

Where would I go to obtain the firmware you talk about? :)

Does this have any advantages over belt driven systems?

This is a belt driven system. If you mean does it have any advantages over linear rod and recirculating ball guides then I believe it costs less and performs better. The 8mm rod used in the Rostock design can vary widely in quality. The cheap rod is usually undersize and rarely straight over the required length.

Hello, i've been slowly assembling the th parts of the printer according to your production drawings and Bom and i have run into a bit of a snag.

I'm assembling the top and bottom boards the drawings state that you need 8 Socket Head Cap Screw M4 x 30mm per pillar (4 in the bottom tube clamp and 4 in the top) but in your bill of materials it says that only 12 are required for the project. also the Bill of Materials says that 16 Countersunk Phillips Machine Screw M4 x 40mm are needed but i can't figure out where.

Are these meant to go through the bottom board?

On another note i am about to buy stepper motors and was wondering what extruder you used, in it here on thingiverse? I just would like to know so i can buy the motor required for that at the same time.

Use 4 countersunk at the top of each tower so there is no head protruding from the top of the top plate. Use the 4 remaining countersunk to attach the hot bed. Use SHCS for the bottom of the towers.

I used http://www.thingiverse.com/thing:18379http://www.thingiverse.com/thi... for my extruder.

Greg's Wade reloaded - Guidler, Tilt Screws, Fishbone Gears

Would it be possible for you to upload the diagrams for the top board and bottom board as .stl files. The lesser cutter that i use doesn't recognize .dxf files, if not, that is cool too. Do you know any way convert .dxf files into .stl files.

I know you can open .dxf files in solidworks and then save them as .stl, but that requires you to have solidworks. You can also open .dxf files in Adobe illustrator and save it as a .ai vector, most laser cutters i know can use this type of file.


can you tell the dimensions of the top and bottoms boards (length and width). I downloaded the specs, but they don't give me a sense of scale on the part regard the the boards. Also where exactly should I place the hole on the boards.

I've uploaded a PDF drawing with all the dimensions included. The shape is changed and a few more holes added but everything else is still the same.

hi. i am planning to make one for myself. i want to supply everything locally. can you take a look at this link:

this is the closest thing i could find to arms. i am not sure though because it translates as carbon fiber pipe.

also is there an easy way to make a jig? there is a laser cut thing but i assume it will be expensive for a single piece.

My Jig was a scrap of plate with 2 holes drilled in it. The important thing is that all arms are the SAME, not that they are exactly 250mm. The holes were the right size for the bolt head and 250mm apart. Then I applied glu and assembled each linkage, set it in the jig and then placed it to one side for the glue to cure. That tube looks just like mine.

thanks for replying.

you recommend polished stainless steel. sellers tend to not take me serious as i am asking for very little amounts so if i fail to find a good polished tube, can i survive with another finish (like brushed etc.)? there are a lot of sellers but if they have generic steel and they make the finish according to orders they take, they may not do it for me for a single piece.

by the way, i am going to start building as soon as i get my prusa i3 and lyman filament extruder calibrated.

Polished in this case doesn't mean a mirror finish. 180 grit or better (higher number) is perfect. Most folk would say this is brushed or sanded rather than polished. But polished is the word used by the industry.

thanks for answering my questions. today i made my first commitment and bough steel tubes. when i asked for polished finish the man thought i wanted mirror finish. i ended up buying satinated finish (should be at least 180 grit). also i had to settle with htd 3m belts as gt2 was impossible to find locally.

You would be better off if you hunt down the GT2 or S2M profile belt. There's plenty of online retailers selling it and on ebay.
The belt connection on the carriage won't work as intended with the 3M belt and you'll need a 12 or 14 tooth pulley for the 3M belt.

i bought 16 teeth pulleys (why did you say 12 or 14? i thought more is better.) as for the carriage, my prusa i3 which was designed for t2.5 works very well although it was not intended. i hope that 3m belts will work. otherwise i will need to get creative with them.

My design calls for 20 tooth pulleys with 2mm pitch. which would be 13.3 teeth on a 3mm pitch so the nearest pulley would be a 12 or a 14 tooth pulley. More teeth will give you a higher speed at the cost of lower resolution. You may also have problems with the idler pulley because it has a small diameter you may have noise or vibration issues as the teeth roll over it. It MAY work just fine but I know the GT2 belt will work better.


After messing up belt orders I had to order again and now I switched to GT2.

In a comment you mentioned that you used this extruder:

However, at the wiki you wrote that the extruder was slow at retracting so prints were stringy. Do you suggest using another extruder? Or did you switch to this one recently?

Greg's Wade reloaded - Guidler, Tilt Screws, Fishbone Gears

I am still having issues with the extruder and PLA filament jamming so this part of the design is for you to experiment with and make your own selection. For what it's worth I no longer think that retraction speed was to blame for the stringy prints.The printer platform is working well. The extruder issues are an unwelcome nuisance.


After messing up belt orders I had to order again and now I switched to GT2.

In a comment you mentioned that you used this extruder:

However, at the wiki you wrote that the extruder was slow at retracting so prints were stringy. Do you suggest using another extruder? Or did you switch to this one recently?

Greg's Wade reloaded - Guidler, Tilt Screws, Fishbone Gears

What was wrong with the old Carriage file that required a fix? Should i print my parts again?

There was nothing wrong with the 19mm carriage. I just deleted the thumbnail image and there's no way to re-create it except delete the file and upload it again. The 20mm carriage was missing the support material for the bridge.

how long is the carbon fibre tube that makes up the arms? I'm looking around your drawings and I cant find a set dimension. i want to figure out how much i need to order from your site.

6 pieces at 206mm.

does that include the 20ish mm that goes into the clevis on each end?

Yes that is the right length to give a finished length of 250mm for the arm. You need to "set" the length precisely when you glue the arms. I recommend taking the time to make a jig.

ok thanks, I do have access to a laser cutter so I might try and mock something up that way to get the best accuracy.

If you don't mind me asking, how long do you think it will be until some sort of kit is available? i don't mean to sound pushy its just this will be the first rep-rap i build and i don't want to start without a better idea of what i'm doing. I am very excited about this project.

There won't be a kit for at least a couple of months yet.

Whats the status of your kits? I was going to buy linear rod for a Rostock from your site, but now that I've seen your design, I'll go with this.

I can't wait until you release a kit for this. I really want to try and give it a go.

To allow any size square tube, you could change the carriage design to look at little more like a Kossel. His beams are currently 45 degrees rotated compared to yours but that would not be hard to redesign.


This design only works because the beams are rotated 45°. The small bearings run on opposite corners of the tube.

Looking for square tube stock in the US seems to be a little difficult as everything is in standard, is 3/4" tube an acceptable replacement? That comes out to 19.05mm so I would imagine its close enough if not the same.

19mm and 3/4" are essentially the same in this case. I will need to create a modified set of parts for use with 20mm tube though.

Are you planning on switching the design to 20mm tubes?

I intend to provide files for both 19mm/3/4" and for 20mm tube once prototype is printing.

how is the noise level of the bearings on square tube? nice work.

There's no doubt it's noisier than my Prusa-Mendel. But I think it's the top and bottom board and the desk top it sits on that radiate the noise rather than the tube.

Very nice job. What is the straightness spec on the square tubing?

I don't actually know what the spec is but it's straighter than the last set of 8mm linear shafts I bought. They were almost 1mm over 760mm

Looks great! Are you planning on making this a kit when you get it all sorted?

Yes a non-printed parts kit is definitely in my plans. Maybe printed parts too

Your universal joint design is about the best I've seen. Great job!!

I like the use of square tube and "generic" bearings.