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


Curta Calculator Type I scaled at 3:1

by mwu Dec 6, 2016
Download All Files

Thing Apps Enabled

Please Login to Comment

Hi there! I figured out the gear alignment dilemma (my own stupidity) however upon attempts to slice the step drum lower//upper halves in cura, support is not generated on the teeth. Do you have any tips for getting the program to recognize these overhangs? (I have set it to generate everywhere and 0 degree angle, so it's not that.)

So how sturdy is this thing?
I'd love to print it as it's probably the coolest thing I've ever seen. But I wouldn't really like to spend my time repairing it.
Anyway, a million thanks for making this @mwu!

I unfortunately don't have a lot of information on longevity. I took an early build to a Mini Maker Faire and I had to disassemble and reassemble it a few times due to parts getting stuck. I later figured out that some of those parts were designed incorrectly. Some of those were my fault and some appeared to be either flaws with the engineering drawings or some of the drawings may be from different iterations than other parts.

When I made Adam Savage's Curta, I spent more time on ensuring that all the parts ran as smoothly as possible. I ran quite a few calculations on his, and I believe it would hold up quite well over time.

The biggest issues I've had:

  • The transmission shafts can bow during computation -- particularly subtraction. This can cause the gears to slip and result precision to be lost. For the most part, this just means incorrect results. Over time, it may strain the shafts enough to snap one depending on whether they are printed vertically or horizontally. The more friction in the system, the worse this is, so spend as much time getting smooth action as you can. In my more recent builds, I am printing with carbon fiber filament to strengthen the transmission shafts.
  • Strain on the main shaft can cause it to snap. This is a tough one because it is combined with the step drum and is a long print. I have been considering redesigning the main shaft to be separate from the main drum so it can be printed horizontally for strength. This would also reduce the amount of support material necessary for printing the step drum.

Hi there! I am going into my senior year of highschool and as my project I am building your curta calculator model. I have approximately half of the components printed, and the components from BoltDepot have been ordered. Recommendations for the assembly process would be amazing as I am having a hard time comprehending the gear alignments, even after closely reading the build manual and watching the build video. Thank you very much for the opportunity to own such a beautiful machine.


What about the gear alignments are you having trouble with?

I was wondering what the file name for the support columns was. I couldn't find it.


Thank you so much!

I am a student planning to print this using an SLA printing process. Would you recommend this? Maybe the parts will be too fragile? I wanted to try this out using an SLA process for higher quality prints for a more realistic look.


I do not own a 3d printer at present, i have never ordered a print yet either my question is what would this project cost (ballpark) to have printed? Trying to get an idea, so i know if its in the cards for me or not.

Hi there!
I have been gobsmacked by this creation since I got to see it on the Adam Savage review.
I will be trying to print this file on the next 3 months (if university allows me to do so...), so I wanted to know any special consideration or any good advice that you could give to me.
Thank you for this awesome design!!!

There are a few remixes of parts (not all tagged as remixes, but search for Curta) that can help make it easier to build or improve its reliability once built.

I highly recommend using a slicer that generates easy to remove support material. I used Simplify3D, but other slicers have improved their support structure generation significantly since I initially completed the project. Do a few tests and see what works. Simplify3D does have student pricing, but other slicers are free (choose the cheapest that works for you).

I think I've taken care of most of the printing related details in the BoM and build manual, but if you have any specific questions, I am happy to answer.

Finishing and painting really add a lot of polish to the project, but are also the more difficult parts. I used paint stencils cut using a Cricut. Any of the home cutting machines should work. After sanding and filling to smooth the prints, paint the surface black and then apply the stencil. To prevent the white lettering paint from bleeding under the stencils, apply a light coat of black after applying the stencil, then do your coats of the white for lettering, then remove the stencils after it's dry to the touch. If you do lighter coats of the white, the lettering won't stand away from the base paint as much. I used flat paints and then applied a clear coat. The nice thing about a clear coat is that it'll reduce how much you can feel the lettering standing up over the black paint.

The 'Double Transmission Gear' also needs support. It is not stated in the BOM though.

Also, a lot of file names don't match with the part names in the Build Manual. Is it possible to update the names in the Build Manual?
Thank you for creating this awesome project!

Thanks for sharing this great project.
I have a question about the stencils:
When i open a file in Silhouette studio, the numbers look somehow wrong.
Do i need a font for this to show it right?


I had similar problems with the cricut software. There are no fonts needed, but SVG support can vary between software. I am unsure whether the problem lies in the software I exported from or in cricut and silhouette's software. I ended up exporting a high res PNG and using cricut's tools that work with raster images and then scaling appropriately afterwards. I am unfamiliar with the silhouette software. Hopefully it has similar functionality.

Thank you for the reply. Do you happen to have those HiRes PNG files. I will try with the Silhouette software and/or with the cricut software. I tend to buy a stencil cutter machine myself. It may come usefull for future projects also.

I've uploaded the svg files and a zip file containing the high res images (thingiverse wanted to just display any images I uploaded individually in the gallery rather than list them as project files). I've also shared the actual Cricut projects I used (unfortunately, they share via Pinterest)

Thank you very much !. The images are super.

Thank for making this accessible for us, this must be the most advanced project to print on a 3D printer.
I started Friday evening and Sunday I had made this ( I know I do not use the manual, I am visual mostly and this works best for me)

I plan to finish my Curta in 2-3 weeks, I have 2 printers Anet A6 and a Monoprice select mini.

That's impressive for the amount of time put in. I don't have my printer nearly as fine tuned and I have to do a lot of finishing of the parts to get stuff done. I do think that the carriage (that upper portion) is the easier to assemble as far as amount of work to get things fitting and moving well. So heads up that there will be more work on the lower portion.

While for a mechanically minded person most of the work can be done without reference, I suggest you do reference the manual or the original engineering drawings for the order of parts. The order doesn't always have a clear reason without really digging into the mechanics. In fact, one of the biggest (maybe the number one) reasons for a Curta being returned to the factory (when Curtas were being produced) was because people had disassembled one and couldn't reassemble it to be functional due to the number of similar parts that have to be in the right order.

2-3 weeks is an aggressive target, do you also plan to finish and paint it? If so, does that fit in that timeline or is the 2-3 weeks just for initial assembly? I once calculated how long it would take me just to print all the parts if I could run back to back prints... The estimated time was around a full week of non-stop printing.

The prints look great! I look forward to seeing the finished product!

Thanks Marcus

I have not taken decision of painting but think I will, of course I look in your building manual, but i do not follow it from page one to last, i do it part that belong together I print and assemble right away. Yes i will post it when finish, I have always been fascinated by mechanical "old" things. Her is another one an Automaton https://www.youtube.com/watch?v=C7oSFNKIlaM

Hello mwu, do you know approximately how large the project is in square mm or square inches?

The sum of the volume of the printed parts or the volume of the plastic used? I haven't done either calculation. The plastic used would depend on print settings -- I list the basics of the settings I used in the BoM, but the amount and type of support will alter that. Volume of all the parts you can determine by going through the CAD parts in OnShape. If you open a part studio, select the part in the bottom of the list on the left, then select the scales icon on the bottom right, it will show you volume of that part. You'd have to go through and do that for each individual part and multiply by the number of each part there are in the finished Curta.

Hi mwu

What an amazing design. I originally intended to buy a Curta, but then I stumbled over your design here and it's much more fun to make a piece of math history instead of just buying it.

I just started 3D printing and got myself an Ender 3.

Last year (if I read correctly) you said 0.4mm nozzle and PLA is enough - does it still uphold or did you change your mind since then?

Greetings from Switzerland

Thanks! I worked on this project as a way to get a Curta (I couldn't justify spending that much on something that would be a mostly unused gadget on my desk). I am very fortunate that Adam Savage found the project and was willing to trade a Curta for a 3D printed Curta, so now I have both and an incredible video of his unboxing of it -- just that video would have been worth all the work :).

Yes, the parts I had the most issue printing were the various collars for the gears (named 10218 through 10222 and 10230) that ride on the transmission shafts. I redesigned them slightly to widen the parts of the collar that didn't work with a 0.4mm nozzle and ran it through an stl repair tool to fix where two bodies didn't join fully that caused additional problems with slicing.

PLA is great for all parts except that the transmission shafts can bow some under operation. For the most part it isn't enough to cause a problem except for subtraction. With subtraction the cascading carries causes enough friction to make the transmission shafts bow enough to sometimes slip a gear and lose accuracy in the result. I have been printing my transmission shafts in PLA with carbon fiber embedded in it which increases rigidity (hopefully enough to prevent this problem). Also, this remix of some of the parts solves that problem on the last three transmission shafts (the most problematic ones) with a brace as well as provides a carry levers that are much easier to print and a few other things.

Curta Calculator Mods
by mcmaven

Thank you for your fast reply! I checked mcmaven's redesign and I'm a bit worried to build it together myself as he has no build manual.

I'm reading the first pages of the manual now and trying to match the parts to the corresponding stl files. I guess I start printing everything first and putting it together at the end. Your video and the pictures are great help!

Comments deleted.

Awesome those carry levers are a pain sometimes.

With the selector knob at the top of the shaft, zero should be dead center. I used a cricut to cut paint stencils and aligned them by eye to be center with the knob. Them you can move the knob one place downwards at a time and see it count upwards. Each time, the number should sign with the center of the knob. If not, I picked up and replaced the stencil.

Im wondering if anyone has a picture of what the carry lever springs look like, also do these need to be baked like the other springs.
Thanks in advance

I'm out of town so I can't take new photos, but I've shared an album on google photos that I think has enough to see what they look like. They do not need to be baked.

What us the dimensioned diameter of the ring with the results cam? My printed ring is 62.46 mm.

There is a notch in the Tens Bell. I wondering if that notch is supposed to push the carry lever from one side of the notch to the other.

Yes, that "ramped" portion of the tens bell raises the carry lever back up, resetting it after a carry operation. There are two such ramps on the tens bell. One for the carry levers of the results portion of the Curta and one fore the carry levers of the turns counter.

The springs on the carry levers are to ensure that the lever snaps to the correct location so that variances in force and friction don't cause problems in the operation. In other words, the springs help ensure consistency in operation.

The ramped portion for the turns counter is too short to push the carry lever. The carry lever just scraps along the top of ramp. However, the ramped portion for the results portion is high enough to push the carry lever.

The diameter of the turns counter ring is 0.5 mm smaller than the diameter of the results ring. I assume they should both be the same diameter.

I have printed two Tens Bells. Each one has the same problem.

My slicer is S3D.

I've gone back and taken a look at the source files. The turns and results rings look equivalent in diameter and the height of the ramp. However, it does seem that the ramp is not as high as it should be on both. I will work on updating these, but since you are only having a problem with the results ring, it seems there is an alignment problem somewhere.

I am looking forward to your revised Tens Bell. In the meantime, I will push forward working with what I have.

I am currently printing your revised Tens Bell, however,

I have superimposed the two Tens Bells on S3D. It looks like the diameter of the turns counter and result rings are identical. The ramp is a little higher but that does not help. My carry levers just skim along the top of the rings. The rings do not push to levers.

How the CURTA Works, is an excellent video that details how the Curta functions and is a great way to understand the mechanisms. It's actually how I was originally introduced to the Curta Calculator.

It calls the ramped portion of the tens bell a camming surface. That pushes the carry lever from its lower carry position upwards into its normal position. What part of that doesn't function properly with your build?

Comments deleted.

I am working my way through the carry levers. I have done three of them and now understand to how make them them and verify them.

A bit of background on why I am doing this project.

I am a docent at the Computer History Museum in Silicon Valley. The Curta is always one of the first artifact that I talk about on my tour. It is is an interesting technical and important historical object. Also the story of it's inventor, Curt Herzstark, is heartwarming.

The Curta in the museum's gallery is behind a glass case as are most of the the museum's artifacts. People can see it but not touch it. I bought a Curta Type II from eBay a couple of years ago. I enjoyed it for a couple of months but then decided that I wanted to share it. I wanted the museum guests to get a hands-on Curta experience. It is now one the artifacts that our interns have available for guests to manipulate on certain weekends.

I got interested in 3D printing about 18 months ago. I knew that I had to make a Curta after seeing your project on Thingiverse. I took me many months to finally decide to do it. I am now enjoying the process. I don't know how it will end up but the journey is rewarding all on it's own.

Thank you for bringing this into my life.

People asked me why I released my Curta work for free. This is why. I wanted to inspire others and to share an amazing journey I went on.

  • I didn't know much about CAD, 3D Printing, tolerances, finishing, or painting when I started and I learned a lot of fun skills along the way.
  • I also learned a ton about the calculator, how it works, and all of the little ways that the mechanics ensures the correct answer (as well as many ways in which it can go wrong -- particularly for a 3D printed model).
  • Finally, I learned a bunch of interesting things about math. For instance, the sum of sequential odd numbers provides sequential squares. That tidbit of knowledge is how one does square roots on the Curta calculator -- by subtracting sequential odd numbers.

I am having a tough time with the carry lever springs.

Your suggested length of wire, 57mm, makes the middle segment much too wide. I have been trying different wire lengths but still have not found a length that works with the tool. Do you have a better size estimate?

You video and images make it look like the spring fits below the notch to make it a push spring. The original Curta has the spring above the notch as a pull spring. Which type are you using?

I did have to play around with wire length some. If the middle segment is too wide, then the legs are too short or the wire is too long. When you insert the wire for the first two bends, the length of the wire from the inserted end to the edge of the tool should be 23.5mm. If that is correct on your tool, then you would need to alter wire length as you have been trying to do. If not, it could be a bit of plastic blocking the wire from being inserted fully.

I'm not sure what you mean by push / pull spring. The legs of the spring are intended to "squeeze" the carry lever on either end of the angled notch on it. This is supposed to prevent the carry lever from being stuck between its up and down position. I have put together a shared album of pictures I took this morning to show how the carry lever spring sits with the carry lever and the bracket it sits in. Please note the comments I placed on each photo.

The depth of the hole in my tool is 23.41 mm which should be close enough. I have found the 53.5, for whatever reason, works for me.

I was confused about the function of the spring. I had thought that it was to either push or pull the lever back into a position after it had been moved. Now I see that it snaps to one side or the other of the notch.

Again, thank you for your efforts in helping me.

I have printed the separated step drum. I am not able to pull the step drum up into the subtract position. The anti-reversal plate is too thick. If I reduced the thickness of the anti-reversal plate by 0.86 mm then I can pull it up.

The consequence of the thinner plate makes the step drum sit lower in the add position. Sitting lower makes the transmission gears not line up with the gears of the step drum.

It looks like the lower main axle was not inserted far enough into the step drum. Of course with the separated step drum the lower main axle is printed into the part so there is no way to check this.

The lower main axle was created by cutting the main axle and then combining the top portion of the main axle with the step drum.

The separated step drum was combined with the whole main axle and then divided to where it is easier to print the two sides.

Final alignment of the axle and step drum is equivalent between the two. Just in case I made a bone-headed mistake, I just double checked the CAD files to verify that.

If it aligns properly for addition before sanding the plate down, I would suggest that fitting the plate into the groove for subtraction is the problem. I do have to sand the plate some (maybe up to ~0.4mm for a final thickness of ~2.6mm) to fit it to the collar on the step drum, but at almost a full mm, I think there is another issue. Maybe you have some amount of plastic drooping on the collar on the main shaft? Getting dimensional accuracy on top of supports can be tough. If you have calipers, check the size of the two gaps in the collar on the lower portion of the main shaft. They should both measure to 3mm which should allow for that plate to sit in those grooves.

The gaps between the collars are ~3 mm. The surfaces are clean and smooth. The 3 mm plate fits between them just enough for a sliding fit.

The distance between the top of the top collar and the entry point of the shaft into the step drum is 15.5 mm. Cab you verify that this is the correct distance? The correct distance compared to the original, non-separated step drum?

It should be 15.15. The screenshot I attached shows the selected surfaces in CAD (top of the collar, and where the shaft meets the bottom of the step drum). The distance is displayed at the bottom right.

Since this is different from your print by a significant enough amount, I re-exported that model to stl, downloaded the stl from here, and then compared the two in my slicer. The position of the main shaft and collar in the current model here on thingiverse was identical to the one I exported from CAD.

There seems to be a difference between your print and the stl file. I don't remember if I updated that file after uploading it. Assuming no change to the file, it would be problem with the printed dimensions in the z axis. I suggest you download the current version of that file and compare it in your slicer to ensure that the collar starts and ends on the same layer. If that isn't the case, print the new file. If they are the same, then something is wrong with your print. I would double check z-axis steps per mm if the problem isn't an old copy of the model.

Thank you for your efforts. The file that I printed is identical the current file on thingiverse.

I am currently doing a long print. When it is done I will check the accuracy of the Z axis. If it is wrong then this may mean that everything that I printed for the Curta is badly dimensioned. :-(

I checked my Z-axis. It is dead on. I re-printed the part. It is within spec. Moving on.

That is strange, but I am glad the problem is fixed.

I ran into unexplained printer issues too. The counter body's geometry did not line up with the tops of the transmission shafts. It would match up at one point and it would be off a small amount more each hole going around the counter body until at 180 degrees it was very noticeably wrong. I checked everything I could think of and reprinted many times. I checked with multiple people where the problem might be including Simplify3D who provided excellent support for their product. Finally, I flipped the part over and printed it upside down from the original orientation. That fixed the problem and I have no idea why. As far as I can tell the problem was not in the model, slicer, or final gcode. I think it was a problem with the precision of an Arduino or a bug in the firmware or something along those lines.

Dear friends, I do not have a 3D printer and I would like to know the possibility to achieve or purchase the material printed in order to build myself the Curse calculator Type I.
Thanks for your indications and comments.

I would be interested in printing these parts for you.

Thanks for your help, there are a lot of parts for a Curta Type 1 calculator. I wait your indications.

Has anyone considered multicolor prints to embed the letters and numbers into the print? I know I do not have the skills to convert the prints to multicolor and add the text, but I am sure there are folks here that have the skills.

I considered doing it, but it's tough to put lettering on a curved surface as paths in order to cut the model. I used OnShape which doesn't (or didn't when I made the models at least) have that capability.

Do you have the separate STLs for the step drum and the bottom of the main axle?

It is a lot simple to re-print just the bottom of main axle when it breaks rather than the combined part.

Also, separating the upper step drum from the upper main axle would be helpful for replacing broken upper main axles.

I sure do. They are:
https://www.thingiverse.com/download:3856967 <-- Main_Axle_and_Step_Drum.stl
https://www.thingiverse.com/download:3040914 <-- mainshaft-_main_shaft_bottom.stl

The 1.8x28x40 spring is not available. Has not been available for many months. What is a good substitute?

Thank you.

I am having good results printing the parts using a two extruder printer. I use one extruder to print water soluble, PVA, supports. I can print the supports right against the part without damaging the part. I have also found that using 0.3 extruder for the PLA prints the parts with better dimensional accuracy.

Comments deleted.

Where did the hardware list for the build go, i cannot see it anymore?

You would be looking for the BOM (Bill of Materials). I have two copies -- a PDF version and a spreadsheet version.

In the PDF version, the list starts out with 3D printed files and then goes on to list the hardware on page 4.
The spreadsheet version has multiple sheets. The first sheet shows 3D printed parts and the second shows non-printed parts (hardware).

Both copies can be found under "Thing Files" near the top of the list. If you hover over the file name in the list, it will show the alt text which will also show the file extension (xlsx for the spreadsheet).

Curta Calculator Type I scaled at 3:1
by mwu

What a muppet, I never saw the extra sheets in the excel file. Thanks for the fast reply.

Hi sorry to bother you but this project is truly amazing. I have a smaller 3d printer but is it able to take this back to 1:1 scale ?

It can't be printed at 1:1 on an FDM printer, and I doubt that any current printer could make a functional 1:1 copy. I had to do 3:1 to be able to print it with a standard nozzle. A few people mentioned attempting a 2:1 scale, but I haven't heard from them. I think 2:1 would require an SLA printer.

You are an amazing brain of mechanics.
I will try to print one for myself... if possible. If not I will buy one from you.

Thanks, but I am not building these to sell them because of how much work is involved in finishing the parts so they assemble and move smoothly enough to function. I am also not selling the parts as kits because parts can easily break while finishing them and it could be any of the parts -- many of which are long prints.

I have some ideas for making production of parts faster, but I have a lot of other projects on my hands right now.

What is the size in millimeters of the biggest piece of this 3:1 model of a Curta?
Please Answer

Sorry it took so long to get back to you (family emergency). If using the original main shaft and step drum (seems to be stronger structurally), the tallest part is 213.75mm. If using the divided step drum, the tallest parts are the transmission shafts which are 170mm. Just under that is the lower housing which is 168mm in height. For width and depth (x and y dimensions), the largest parts are the lower housing and the upper housing. They are 159mm in diameter.

So the 3D printer must have at least a 159mm x 159mm x 170mm build volume. I recommend one where the z axis can reach up to 214mm for the original main shaft and step drum.

So, a well tuned Mini Kossel 3D printer would work just fine? The build volume on those is 170mm diameter by 300mm tall

Yes, that should work.


Thanks a lot for this great project! I am trying a build with my Anet A8. So far it is going pretty well. I used blender to make some rocket gantry style supports for the shafts. Otherwise the extruder nozzle brushed them and knocked them over. I also modified the spring clips, making them 3 times heavier gauge; but reducing the rounded protrusion so it doesn't interfere with other things. They seem to work at least once. Each time you remove them they are likely not to hold again. Of course, they are now not to scale.

I have a question. On my build, the lockouts held by these clips on the first results and first turns counter shafts do not interact with the nearby rings on the tens bell like I would expect. Instead they are just above the ring. This seems odd. Are they supposed to be like this? The other lockouts with carry levers do interact with the rings.

I think I answered my own question. I imported several of the stl files in blender and positioned them as they would be in the completed Curta. The lockouts interact with the tens bell rings at least to a degree. My build must be suffering from accumulated small errors in printing. I think that I will adjust the lockouts a little and reprint them because it is the easiest way to fix the problem.

Sorry it took me a little while to respond. There are a few things you should check out.

  • The under side of the main casting interacts with the top of the tens bell body.
  • The bottom of the crown gear tops to the transmission shafts interacts with the top of the main casting.
  • The clips interact with the tops / bottoms of the lockouts.

All of those interaction points need to be smooth, so if there are imperfections in the surface of the print, they will need to be sanded / filed to fit properly.

I don't think I needed to alter the lockout stl to get them to match up with the tens bell, but that may work.

I am not sure how far you are in your build, but you may also want to check that aligning a gear to the single tooth section of the step drum and then shifting the step drum to subtraction correctly aligns it to the 9-tooth section and vice-versa.

If the step drum doesn't align properly, the step drum alignment is based on the bearing plate.

  • The step drum must seat properly in the bearing plate
  • The locking plate must align to the collar on the bottom of the main shaft correctly in both positions. The locking plate itself can be off a little bit if the surface of the bearing plate it sits on is not smooth.
  • The bearing plate must be the correct distance from the main casting (the support columns interface properly).

I hope this helps

Thanks for your reply. I am at the point of putting in some of the setting shaft assemblies. At this point I am really starting to see how the Curta operates which caused me to notice the problem. I will look my build over carefully with your comments in mind before taking any more steps.

There have been a fair amount of artifacts due the support needed for these complex parts. The first layer above the support tends to be loose and the support was difficult to remove from the narrow grooves in the main shaft and other parts. I am using Slicer and belatedly realized that some alteration of the support parameters might help with these issues. I had to do quite a bit of filing of the locking plate and the matching grooves on the main shaft to make them fit. At first the support columns did not seat well. The bottom of the main casting where the tens bell hits had loose artifact from the support; so I filed this and now the support columns fit well. Similarly, the bottom of the crowns of the transmission shafts had loose artifact and I trimmed that off. The crowns now fit beneath their ring very nicely. The weird thing is that it seems that most of this filing should have helped with my problem already.

Another thing that I did should have helped. I noticed that the bolts holding the spring in the tens bell poked out of the top of the tens bell a little and would hit the main casting; so I made a little plastic washer to put under the heads of these bolts.

Anyway thanks again for your comments and for all your work in making this project.

Your experience is not out of line with what I do. I fit parts, test them, sand / file them, refit them, etc. while I build one. I've found that similar to other manufacturing methods, the basic process only gets you so far with 3D printing. Further work is necessary to take a good foundation to a finished product. Unfortunately for the Curta, 3D printing creates fairly imperfect parts compared to molds. The parts also have to align with each other just right so the tolerances are pretty tight for 3D printing even at 3x. This means being anal about the fit and as you said, having an understanding of how it functions, can be the difference between building one and building a functional one.

Thanks for bringing up your solution with the tens bell spring. I also had an issue with the bolts being a little bit long. I think I actually filed the ends of the bolts down, but adding a washer is definitely an easier way to go. I may actually add a little bit of thickness under where the bolts rest to eliminate this problem -- kind of building the washer into the part.


I just stumbled upon this and think its an amazing print! I have a few questions about it though. Will I need supports on any of the parts or could I position them at a certain angle on the print bed so that they dont need it? Also, how long did the overall printing take?


Yes, supports are required. If you check out the BOM, it lists what supports are required for each part. The Curta is a fairly complicated mechanical device so I stuck to the original design as much as I could so it isn't exactly 3D printer friendly, but it is printable. Arrangement of the parts on the print bed can definitely reduce the need for supports, but not entirely. It's definitely needed to arrange some parts in specific ways to get the strength necessary to function.

I once added up the print times for all parts to just over a week, but that number makes some assumptions. Print speed is highly dependent on the printer. It also assumes that the printer is running 24/7 with no stops and no failures which isn't really possible with a normal printer (gotta take prints off the bed, reload filament, etc). Finally, I used print estimates from Simplify3D which don't include firmware settings like acceleration / deceleration, max velocity, etc. In reality, one would need to print parts and come back to the printer later to remove them, queue up new parts or a reprint if it failed. That will likely introduce many hours of downtime for a printer. I think with a well oiled machine, all of the parts could be printed in 2-3 weeks. Unless you have a monitored printer farm or automated print removal (or both), it would be hard to pul it down below that.

Thank you so much for a detailed and immediate response! I have one more question. I know that the printable Curta is around 3 times the scale of a real one. Having said that, what is the smallest that it could possibly print and still work the same as the 3x version?

For an FDM printer, you probably can't go below 3x. An SLA printer might be able to do ~2x, but it would require reworking the tolerances for all the parts in CAD. 1:1 might only be possible with the original materials (or better -- I'm sure there are some modern materials with the right properties that could do the job).

I think I am going to attempt this to compliment my real Curta...
You think PETG would be an ok material? However after reading your manual filling some parts seems to be a requirement so I might have to order some ABS.
Thanks for all your hard work!

I look forward to another Curta being made out there!

PETG is a great material for everything except for the transmission shafts. Those tend to flex a little bit under a carry from one result digit to another. In fact, I had problems with that even with PLA (more rigid) and after eliminating as much friction as I could. I have been updating my Curta to use PETG infused with carbon fiber for the transmission shafts. It requires a hardened nozzle, but it is a decent amount more rigid than PLA.

Do you mean filling or filing? There is definitely some work to ensure that parts perfectly fit one another and some to make the exterior smooth. I don't have any experience with acetone vapor smoothing, but you'll need to maintain mechanical tolerances. I think manually smoothing and filling with Bondo would produce a better result (but without any experience to back that up).

Ok great. I might have to play around with tolerances since PETG has a different shrinkage if you did yours out of a different material. When I get to the transmission shafts and end up breaking them maybe I will try to use some of the machines at work to turn and mill them (however I know thats going to be a while away).

Yes I meant filling with body filler. PETG has the negative property of not liking paint or body filler sticking to it. I have recently been using this stuff called XTC-3D at work to fill in some ABS. It works really well, self levels, and is really easy to sand. I just need to try it out on some PETG. It is a 2-part epoxy basically so it should work.

Thanks for the tips!

I also completely printed mine with PETG. I had to reprint the transmission shafts with PLA to be more rigid. Because of the shape with the notch PETG would be permanently bowed. Also download the files from the ONshape and print the shafts horizontal and the crown after for best results. I have been using a 0.3mm nozzle for all the fine parts such as the crowns and gears (10218 etc).

I saw your calculator in Adam Savages web site, and got hooked. I currently have over 150 hours of print time into it.

To make the shafts, pins, and other parts without supports I've used Meshmixer to split the STLs. For shafts I split them lengthwise, print both halves, then glue them together. The results have been very good. The shafts are strong, because the fuses go lengthwise, and the extra layers on the bottom side of each half forms a strong central beam, which makes them stiffer, and enables them to handle more torque. Using dichloromethane Acrylic solvent on PLA, you don't have much setting time, but the joints are really strong.
I was able to print the transmission shafts this way, with the crown gears on them, and the Reversing Lever I split into 3 parts and was able to print without supports.

For the Bearing Plate, I imported the STL into my CAD system, measured its dimensions, reconstructed a model of it, and split it into 3 parts. I added two 1/16 inch holes through it for alignment. The holes allow me to use some 1/16 drills as alignment pins when I glue the pieces together.

The Step Drum Lower is giving me trouble. I considered reverse engineering the Step Drum Lower, but would take a immense amount of time to recreate an accurate model of it. In your assembly video, you have a Step Drum Lower print that you push the Main Shaft into, but on this site you don't have an STL for that version of the Step Drum Lower part. I was able to print the version that you offer here, where the Main Shaft is included in the STL (with supports), but the shaft broke off it while I was picking out the support material.

I used the technique above to create a strong Main Shaft (which you did post an STL for), but I don't have anything to insert it into. Could you post a Step Drum Lower STL that offered a hole for your Main Shaft STL? Like the one in your video.

I attached a photo of the parts that I've printed so far.

Thanks for taking the time to make this design available. I can't imagine how long it took you to reverse engineer the Curta.

Regards, Steve

You should be able to create a copy of my onshape document to make changes to the step drum. You can find it at https://cad.onshape.com/documents/57e3f5c529d2af11276b1529/w/eafe4a700fe87f2b71bc3d31/e/f1e5e18535312417d6894ed1

I found it. Thanks! I split the Step Drum Lower, segments, and shaft into 13 parts that can all be printer with no supports. If you are interested, I can send you a zip of the files.

Are there some parts that need to be printed multiple times? And in wich way I can draw the numbers? (I've got the artistic skills of a walrus)

Yes, many parts need to be printed multiple times. There is a spreadsheet in the files which lists all of the parts and how many times they need to be printed.

I unfortunately don't have an easy way to paint the numbers right now. I was working on a method using a dremel and a pantograph, but it didn't get good results. It might work with more expensive equipment, but that would defeat the purpose.

The easiest way to get the numbers on the dials is to print off the numbers, cut them out, and glue them to the dials.

Coming from Surveying, one of my instructors many years ago had one of these and I have wanted one ever since.
Consider this “Challenge accepted!”

Super greatful for all your hard work

if you could spell grateful

It is pretty challenging and it is a long build, but I look forward to seeing another make posted!

Thanks for the Warning ;-)
I spent a couple days trying to print the transmission shafts in various positions with different support. As you know because they are so thin and tall there can be issues with adhesion to the bed and getting them to print straight and strong. I was using PLA for support with PETG for the parts. As the shafts are so thin I was getting some PLA infusing in the layers on the support creating brittle layers.
Thank you very much for supplying the CAD files!
I am able to print the shafts horizontal and the ends upright and glue them together later.
Haven't yet tried this as I am printing the lower body (started that to give myself time to find a solution to the shafts). Using PETG at 30mm/s. I am currently 2 full days in on an 80 hour print (at 95mm).
Wish me luck!

I need to package up the files I created for printing the transmission shafts horizontally. I printed them without the crown gears at the top, but the crown gears do need to be oriented correctly for them to work right (the crown gear needs to be aligned in a certain way to the gears that slide up and down the transmission shafts). To accommodate that, I added a feature to the crown gear that fits the groove on the transmission shaft to ensure they always align properly. I don't want to confuse others, so I need to update the documentation at the same time I upload those files. If you want access to them now, you can export the STLs from the OnShape workspace which has the changes.

Btw, when printing the transmission shafts horizontally, I add as much supports as the slicer will generate and add a raft. I've also been printing these in carbon fiber infused PETG as it's the stiffest filament I've found.

Thanks for the reply! I have been using PETG for all the parts so far. Only downfall is I only print that at 30mm/s which is why the lower body took 40 minutes shy of 3 full days. The lower finished with no problems, looks gorgeous.
I printed the 10207's horizontal, only used a skirt, no supports or raft, turned out great.

Only question, are the new STL's you mentioned different from the CAD files currently in the OnShape? as that is what I printed. They seemed to be 5mm shorter than the STL's on Thingiverse if I lined them up to the same spot inside the crown. I had printed the other transmission shafts (10208, 10209, and 10216) vertically, they are weak and slightly crooked but I can use them to align the shafts that I print laying horizontally before I glue the crowns on. Again, are those 3 any different from the CAD files on OnShape as far as the direction of the Notch to align the shaft?

Only about 200 more parts to print ;-)

OnShape has the most up-to-date versions. One thing I do need to do is add a small marking to the shafts to show how far to push the crown gears on. 16mm of the shaft should be covered by the crown gear -- so far I have just marked the 16mm by hand.

I am at this stage now and is going well. Again this has been quite the test for the printer and slicer settings.
Your STL files for parts 10218,19,20,21,22,30 all reflect how the parts were assembled in OnShape. On the bottom half S3D was trying to print 2 independent walls as per the design of the part and was creating a weak wall due to a gap that could never be filled. I exported the files into Fusion 360 (just what I am used to) and fused all the components into one part for each of those parts and re-created STL's. S3D was then able to treat them as solid objects and print them completely solid. I am using a 0.3mm nozzle for these objects as my 0.4 was always leaving gaps between the layers of the prongs (tangs).

China has lost all my bearings and I have to still wait another 2 months for replacements. As you warned me, this is a long build ;-) over 400 hours of printing time (slower because of PETG).

Thanks again!

Good to hear that you're making progress. I need to get back to mine and repair it (needs a new step drum which broke almost two years ago when it fell off my desk), finish, and paint it so I can take it to some Maker faires. I've been having some trouble with my printer lately. I am using Klipper on a Replicape and it is refusing to print after doing an upgrade. I think it might be due to my webcam. I'm going to try unplugging the camera and try it again tonight.

That is awesome! Glad we could help pressure you along ;-).
Like a sadist, I am contemplating printing another with wood PLA and then staining it. Not authentic but I think it would be pretty cool.

wow, a stained wood one would be awesome.

Came here because Adam Savage :D


to see this masterpiece in action :D

Thanks! The wait from when I sent the Curta to him to when the video was posted was agonizing, but totally worth it. That video is amazing and I would be lying if I said I hadn't watched and rewatched it a few times :D

Oh i believe you :D

It would be soooo awesome to meet him or work with him once _

I met him on his Brain Candy Live tour in Durham, NC. His Curta was mechanically complete, but not fine tuned or painted. It was very fun to see his reaction in person and to talk with him about the device. He is very much the same personality you see on TV and on Tested. The enthusiasm is infectious -- he's very fun to be around.

Does the printer need to be extremely accurate in order for this to work?

No, in fact some of my prints were pretty bad. However, building this isn't ever going to be as simple as print and assemble -- to get the correct fit (minimizing play while not having too much friction), the prints need to be sanded and filed to fit (manual fitting). Doing that also has the side effect of reducing the reliance on print quality. However, the better quality you can get, the less work you'll need to do later to manually fit the internal parts and to finish the external parts.

Hello there. I have a couple questions. I really want to print this for my calculus teacher as a gift but my school's printer only has a build size of 203x152x152mm. What parts of the build actually need the required 160x160x170mm? Is there any way I can change those parts so I can use the only 3D printer available to me?

It's pretty typical of the X and Y dimensions of the print volume to be square. I would think the build volume is 152x152x203 (listed as X x Y x Z). The housing (the knurled grip parts) are 159mm in diameter. If you have the ability to modify the part and chop off the knurling it might fit in 152x152 on the X and Y dimensions. The Z-axis size should be fine.

Hi (again)
What orientation did you print the reversing lever at?

It doesn't get exposed to great stress, so I printed it upright. It will print mostly support-less, but it does help to put some supports in the cut-outs near the bottom. If you have problems keeping it from falling over, use a few loops of a brim to extend it's footprint on the bed.

Great! Thank-you.
Im still working on this. I have just finished assembling the selector shafts and knobs and need to put in the ball bearing and springs....

Always wanted a CURTA but with prices starting at $1000.oo + its been difficult to justify.
Now a model that I can make on my 3D printer....now i'm finally going to get one and have the pleasure of assembly and learning how it works.
Keen to do some extra upgrades with machined parts.
Appreciate the BOM, construction notes and videos You have obviously spent a lot of time to produce an accurate working model.


Thanks, I'm very interested in seeing some machined parts if you do that!

Thanks for this great design! I'm about 80% through printing the parts and have started to file/sand/fit the first few parts together.

Before I buy the hardware, do you think it's essential to use metric or can I use imperial nuts and a tap/die kit?

You could probably get away with using an imperial hardware and tap / die kit. I haven't tried it, though. Even after doing the math for conversion, you may need to play around a bit to find the right sizes (whether the next size up or next size down is the best fit).

Hi, im having some issues slicing the small transmission gears (10219_-410002.stl, 10230-_410008.stl, 10221.stl, 10218.stl, 10222.stl), it doesn't slices the bottom drums (pic attached). Do you have any idea what might be wrong? Also, did you use supports for this pices? Thnks a lot!

If you drop you extrusion width down some it should produce something printable. Yes, I did use supports.

So I got myself into the 3D printing buff too. I am looking into getting a 3D printer and I was wondering what size build plate is required to be able to construct this? I have access to a larger 3D printer for a few of the larger parts at my college but I need my own. Opinions are greatly appreciated! Thanks! Magnificent work!

The printer would need a 0.4mm or smaller nozzle (pretty standard), and a minimum of a 160x160x170mm build volume. There is an option to print the step drum in one piece (the bottom portion of the shaft is split). That requires a 220mm build height and uses a lot more support material.

Would it be possible to upload the Bill of Materials as an excel file?

It is already there. It is the third file down in the file list.

Posted an update on my work towards an easier and cheaper way to letter and number the Curta: https://www.thingiverse.com/groups/curta/forums/general/topic:25464

I've started printing parts, but I'm having some troubles figuring out which parts are which. The build manual is super useful, but you use different names than you use in the BOM spreadsheet and file names in several cases. Would it be possible to convince you to reconcile these when you have time?

I'll find some time for updating that. In the meantime, if there anything you're stumped on, let me know and I can point you in the right direction.


I haven't tried it. 45mm would put more tension on the carriage. Maybe too much.

Is it possible to print it in a smaller scale, for example x0.75 or even x0.5?

There would be a lot involved -- all of the tolerances would need to remain while the dimensions scale. Also, the bolt sizes would have to scale to the nearest equivalent instead of an exact scale amount. Some of the parts are as small as a normal 0.4mm nozzle will go, so a smaller nozzle or a different type of printing would be necessary.

So, possible but you would need to do a bunch of CAD work and have an unusual printer setup.

I just broke the small support pillar (look at attached pic). What purpose does this small piece serve? Would it be okay to glue something there or should I reprint?

You could probably get away without having it there. If it is still there, just cracked, that is fine. The first one I built had that happen. It mainly just adds strength so that the bottom section doesn't flex any. PLA is pretty rigid to begin with so it is probably fine.

YES! I was so hoping you would say that lol. It was so small that it blended in with the support material. Glad to hear that its not super critical.

how long should the support be

110.25mm -- If they weren't right, the selector shafts wouldn't align well -- they do in your last picture.

how long should the supports be

hey my number 9 it lining up what do i need to do pls help if i lift in by 2mm it works

I can't see enough to be sure, but it looks like the problem is too much space between the main body and the bearing plate. The bearing plate must have printed flat (no warp). The support columns must be completely inserted into both the main body and the bearing plate. If not, the distance between the bottom of the main body where the tops of the selector shafts rest and the bearing plate which determines position of the step drum can be incorrect.

fixed printed new bearing plate .

thanks you

Awesome! Good to see everything matching up nicely.

what do u use to open dxf its no numbers i use trueview

I used a program called qCAD to create the files. I paid for the professional version, but there is a free version. The download is a trial, but if you remove the professional plugins, it will become the community edition. You should be able to print to scale with it.

If you don't want to deal with that, you can use eDrawings Viewer. They list their professional version first. Scroll past that to the free eDrawings Viewer.

what is the area size for the number on top and the 8 number on the side .
how big is the paper to print the numbers ur way is to much for for the number cant u use a label maker just need the size ur files dont come out the right size if u just print them.

I did print these off on paper from these files before I got the vinyl stencils for painting them. I had no issue with scale, so it should be correct. Many printers (or software drivers) do some automatic scaling when they print. Check that the print scale is 100% rather than fit to page, and that any automatic scaling is turned off.

I will post back here once I am at my computer and can measure the size it should be so you can double check your results.

How much filament would this print in 100% size need?

I used about three rolls including failed prints and reprints I required

I'm about done printing all the parts. To make it as accurate as possible, can i know the paint code you used for the black?

I used a Krylon flat black paint. The place I bought it is right across the street from work. I'll get specifics today.

Thanks so much for going the extra mile for me...means a lot. What you have done here is so awesome! can't wait to have mine done

First, I'll tell you what I used / did exactly, then I'll describe where specifics won't matter. Here's the list of my steps for painting, based heavily on Bill Doran and Joel Telling's video on finishing a Harry Potter wand.

  1. Sand from 80 to 220 grit sandpaper
  2. Use Bondo spot putty to fill big gaps
  3. Prime with Duplicolor's filler primer
  4. Sand some more with 180 to 220
  5. Repeat 2 & 4 until smooth -- no layer lines or other defects showing
  6. Spray a couple coats of Krylon Industrial Acryli-Quik Ultra Flat Black -- go with somewhat light passes until it's fully coated
  7. Wet sand with 1500 grit foam pad
  8. Repeat 5 & 6 until satisfied with a nice clean and smooth finish
  9. Spray with a clear coat. I used a matte clear coat between the knurled grips and a gloss clear coat from the knurled grips to the parts at the top and bottom of the Curta (spray them separately, though -- not assembled).

That's a lot of very specific tools and paints I used, but they don't really have to be those. I actually switched black paints part way through because I had trouble with getting consistent finish on the first black I chose.

As long as it's a nice deep black, it will do fine. I chose flat so that I could get both the matte and gloss finish with one consistent color, but if you want to cheat a little bit, you could buy satin and gloss blacks to eliminate a few steps.

The filler primer isn't completely necessary -- a few of the later parts I didn't use it, I just used the bondo filler on the entire part and then made sure to sand it nice and smooth with multiple grits. Up to you which you think is easier.

The clear coats are nice because they help protect the paints. In fact, I slipped and scratched the paint on Adam Savage's with my fingernails while fitting a couple parts together. I lightly wet sanded with 1800 grit and then lightly sprayed the area with more clear coat and I couldn't tell I had damaged it.

Any wet sanding foams or papers should work, but I liked the way the foam pads formed to the shape of what I was sanding and the foam helped provide even pressure.

The lettering was done in flat white paint -- pretty much doesn't matter which one, just go for the brightest white you see (not something slightly off like ivory). The silver turns counter section was done by painting a base layer of grey followed by an aluminum metallic paint. I tried a number of metallic paints, but they all seemed to be somewhat translucent. Hence the base grey coat. The red ring on the crank for subtraction was done with a fire engine red. Sorry I don't have specifics readily available for those colors. If you really want specifics I can look those up when I get home.

Don't make it nice < niais < nescius := not-skilled, ass.
little bit: pick one

the number 0 -9 you said ur where looking for a different way . so u dont have the cad file for the base cap

I've built the tool (a pantograph) I need for the lettering and numbering, but I will probably need to iterate on it. I printed out some test pieces to paint last night. I should be able to test out the pantograph this weekend to see if it will be accurate enough. The CAD work I've been doing for the pantograph is public on OnShape if you want to take a look. Details on assembly, parts required, guides to use with it and methods for using it will be released once I iron all that out.

There is an stl for the base cap, but not a dxf for the lettering on it.

YOU have the CAD files for the base cap i like to put letter in the base pls
what happen with the number i almost done printing need the number pls

What number are you looking for? The bottom plate design I have is customized for Adam Savage -- I don't have a generic one.

I had a mistake in my BOM. The size of the upper carriage spring was listed as 1.8x28x25. That was incorrect. I have uploaded a fixed copy to the proper size -- 1.8x28x40. I didn't realize it until I ordered more at the wrong size. I hope this post prevents some of you from ordering incorrectly as the only source I know is in China and shipping takes a while.

I have a printer with a small build plate (area). I know that part of the coolness in this project is its size, but how small can I make the overall build before it's too small to print?

The scale is already about as small as it can be done on an FDM printer. It might be able to be scaled down a little bit if you use a small nozzle (such as the experimental high resolution 0.15 nozzle from e3d). I haven't tried it so I don't know if that would work. It would be very hard to get some of the parts to stick at that size and it would also involve scaling the parts down while keeping the tolerances the same (not scaled down).

Scaling it down means going back to the source CAD files and doing the work there rather than just scaling down these files in your slicer.

I'm not sure what kind of printer you have, but there have been a few posts from folks here who may be attempting a 2:1 scale Curta on an SLA printer.

can u sell me the dxf file all done on the vinyl pls

I'm working on a different way of doing the letters and numbers that won't require the vinyl. The vinyl isn't something I can easily reproduce and sell because the shapes are really too small for the machine. It tries to cut them, but fails on the small bits -- especially with numbers like 0 with an inside and an outside. I had to cut multiple copies for each set of numbers and combine the bits and pieces together. The vinyl also gets pulled and torn in places so it doesn't sit flat after cutting and won't ship very well.

bearing plate and zero positioning disc can u split in half so don't have to use up lots of support

can u sell me the dxf file all done on the vinyl pls
i have some pic of what i printed so far
Thank you

I am working on an easier system for getting the letters and numbers on the Curta. I plan to utilize a pantograph and either a Dremel with a flexishaft attachment and engraving bit or a pen style engraver (http://amzn.to/2xLiL2I). I have the Dremel rather than the pen style engraver -- not sure if it would be underpowered or not.

With the new split step Drum parts can you tall me the settings for infill, resolution & support please? I have probably 5/8 parts printed

Lower part: 20% infill at 0.2mm layer height. 30% support infill.
Upper part: 30% infill at 0.2mm layer height until 16mm height and then 100% infill from there on up.

Printing now,
I just want to thankyou again for making such a wonderful model available for us on Thingiverse!

Thank you!

how do i get the right size for the dxf files thanks

The dxf files should already be scaled to the size they should be. However, I noticed that when I imported them into the design program for Cricut, that they didn't look right, so I exported them as a highly scaled up png image. I then imported them to the Cricut design program and scaled them back down.

what can i use to print the dxf files on normal printer thanks

I used qCAD to create them (https://github.com/qcad/qcad) to just view and print the files, you can use https://viewer.autodesk.com/

I was just looking at printing and realized that at least one of the files I uploaded has lines not visible against a white background. I will update those soon.

can u sell me all the hardware needed pls

Do you mean all of the non-printed parts? The Bill of Materials spreadsheet contains links to where those parts can be found.

hey i have a issue i can print one part too big how much for u to print this thanks Main_Axle_and_Step_Drum.stl

how much for a finishedone if u sell it thanks you i can put together so is it cost less

Thank you

I put up an alternative to the Main_Axle_and_Step_Drum.stl and main_shaft_-_main_shaft_bottom.stl recently for supporting smaller printers. They are Step_Drum_upper.stl and Step_Drum_lower.stl connected with three copies of step_drum_joining_pin.stl with some superglue (or any cyanoacrylic glue)

Thanks for taking the time to create this masterpiece and extra thanks for sharing it. I never heard of this type of calculator until I stumbled across Adam Savages Youtube video on it. I especially like the box you made to house it. This is being added to my 'must do' list after I improve my 3D printing skills and 3D printer. Thanks again. Keep up the great work!

Thanks, I appreciate it!

The box was actually made by https://www.etsy.com/shop/cocobolostudio. I didn't intend to take credit for it and I noted this in the comments of the video.

that's an awesome work you did, i always wanted to have a curta on my own, and i never could. so i'm drooling for your model since i've seen it on adam savage's tested. it actually was the final argument making me buy a 3d printer
i'm looking to make an actual lifesize of the curta not a giant one, as i wish it to be easily carried around, so is it possible? or will it be difficult to scale down because of tolerances? and, if doable, how should i adapt both printed and non printed parts?

anyway i'm gonna make one giant too because i'm craving for one.
again this is an incredibly awesome work and you're a dedicated and talented awesome guy :)

Thanks for all the compliments. It wouldn't be possible to build a working 1:1 scale Curta by pushing plastic. One might be able to go down to 2:1 with an SLA printer, but scaling down isn't as simple as directly reducing the size in your slicer (search the comments here for more info).

To go 1:1 the Curta would need to be built with the original materials or something with equivalent strength and rigidity (not sure what modern materials could approach that). The parts are incredibly small and tolerances are very tight, so AFAIK today's CNC machines wouldn't cut it unless you used a watchmaker's CNC machine which is incredibly expensive.

I would like to attempt it and I do have some ideas for approaching 1:1 scale, but I would probably still need a home machining shop for some of the work.

Thanks you for the answer it proves how dedicated you are for your project and that you diserve the compliments :)

I get it, it's not easy to scale it down, as expected.
I'm not aware of all technical subtilities now, but i'm about about to get a formation in machining, so, if i'm lucky enough to get it and lucky enough to be allowed to use the CNCs for personnal project, i'll maybe try to port your project to full metal and eventually scale it down to 1:1.
Be sure that if i do, i'll happily update you on the work and give you credit back

If you are able to cut a 1:1 Curta on the CNCs, I bet Mr. Savage would be interested in that :)

No doubt he would! Didn't thought about it but now you got me twice as much motivated to do it. :) Even if it is not doable it'll be a lot of extra learnings to show at the end of my formation.

Thanks for everything and keep up with the good work!

learn auxiliary verbs

My next major 3D printing project will be your superb Curta Calculator design. I see in the BOM the resolution things need to be printed at - but it is not clear what nozzle size I can use throughout. Can I use a 0.4mm nozzle to print EVERYTHING or are there pieces that need a smaller nozzle size?? I am assuming that everything can be in PLA - again, please let me know if this is not the case. I intend printing the outer case in black PLA to minimise the painting required :)

Yes, a 0.4 mm nozzle should have you covered. I did have to reduce extrusion width on a few parts to help S3d with it's slicing. PLA is what I used for everything.

If you are going with black PLA to reduce painting, I suggest also printing the selector knob, the upper portion of the selector shaft, the carriage knurling, the carriage cage, the results dials, the clearing cover, the ring under the crank collar, the crank collar, the crank itself, the crank handle, and the pin screw for the crank handle all in black.

Many thanks for getting back on this so quickly :) This is going to be a GREAT project. I already have the black type 1, the black type 2, and the grey type 2 Curta calculators - so I HAVE to print out your version to complete the set :) As I have original Curtas I am of course aware of the effort you have put into this 3D project. To say that I am totally amazed is a massive understatement!! Thank you for all your hard work.

This is just insane. The amount of work you have put into this is just awe inspiring. Until the Adam Savage video I had not even heard of this device but now I really want one, real or 3D printed.

I know the feeling. I got started on this project after I saw https://youtu.be/loI1Kwed8Pk and had the same desire. I checked eBay and realized that the prices were higher than I wanted to pay for what is essentially a fun conversational piece / paper weight. Sure it's functional, but I wouldn't exactly carry it around to do math. I decided to print a working portion of the device, but later I got pointed to the engineering drawings which allowed me to build the entire device with all of its functionality.

I've looked through both your videos and the unboxing with Adam Savage, and I never saw the use of the upper ring of the calculator (the one labelled 1,2,3,4,.....) that controls multiplications of 1's, 10's, 100's,....Is this feature in this model? The real overarching question is, does this 3D printed model truly replicate the true Curta Calculator? Can it do all the same operations?

Thanks in advance...absolutely love this idea and will be printing my own for sure!

Yes, that feature is in the model. The model has all functionality of the original including the decimal place based multiplication (on computers this is called a shift operation since the digits get shifted over) as well as the safety features: the anti-reversal pawl (preventing reverse rotation), the ring on the main shaft that prevents the calculator from shifting between addition and subtraction mid-operation, the ball bearing that prevents the carriage from being rotated mid-operation, etc (too many to list).

Wow...just wow...so impressed man, can't wait to start this project!

I wonder, to get down to a 1:1 scale, if you had all the delicate and small shafts, axles and gears metal sinter 3d printed through a company like shapeways, then do the rest plastic on a home 3d printer, you could end up with something very much like the original without the several hundred to thousand dollar price tag of one on ebay.
Actually with modern cheap sinter metal parts, and robotic manufacturing and assembly, a company could probably make novelty replica curta calculators far cheaper then the originals without inflation. Now wouldn't that be cool, maybe even as a crowd funding limited run or something.

Good idea, and I looked into it before. The only material Shapeways offers DMLS is Aluminum which isn't cheap. Just the main body is over $100 with that process. Also, some of the parts require steel. The steel Shapeways offers is steel powder with a glue binding. That binding is then replaced with bronze. The result is about 60% steel and 40% bronze. They also state limitations about size that might be problematic.

Ive been playing around with uploading some of the different STLs at a 1/3 (so 1:1 scale) up to your scale and with the steel a lot of the little counting gears and shafts are anywhere between a couple of dimes to a couple of dollars per part before you add on labor, and only the few big parts like the step drum (which could be separated from its axle) where closer to $20 dollar material cost (at 1:1 scale). Most of the price is a labor cost per part which can be mitigated by nesting all the part needed on one or multiple big mesh's (like how plastic model kits come where you have to cut all the individual parts away.) I think all the critical part could be gotten at good enough quality to value ratio around the 1.5:1 scale. and with metal for the moving parts at its core it would be much more durable. the housing, handles, main counting body and other such large or static part can be printed on a 0.15mm or 0.25mm nozzle. I think I'll try to make that work after I get a large plastic one done first.

That would certainly be an interesting build. Let me know how your progress goes.

Is a 0.25mm nozzle required as specified in the title of your OnShape document? Can we get away with a standard 0.4mm nozzle?

Have you considered adding bearings/bushings and changing out some of the 3d printed cylindrical parts that act as bearing surface for metal (aluminum?) rods to reduce friction and ware?

The 0.25mm nozzle was originally required, but I made changes to the few parts that needed a nozzle that small and managed to make it printable with just a standard 0.4mm nozzle.

Which points in particular were you thinking of?

I went back through your build video, and I can't seem to find the bit where I originally identified the area where I thought bearings would be a good alternative. Maybe I'll come across it again. However, my general philosophy is not to over-rely on 3d printed parts when I can buy stronger, more precise standard hardware instead. I was thinking the support shafts/columns could be 4mm threaded rods glued into a 3-d printed cylinder so that you don't risk over tightening the screws. I feel like the zero-positioning pin could be aluminum as well. There are numerous places where you screw bolts directly into plastic (tens bell spring, anti-rotation-plate, transmission shaft lock ring, etc). While it may be overkill, I'd have used heat-set inserts in these areas rather than tapping the plastic itself.

Additionally, the main crank should be split into 2 parts. If you remove the shaft that the crank handle attaches to and make it a separate part, you can make the handle MUCH easier to print; you can put the top surface of the main crank on the print bed and print it upside down without support. If and when I make one, this will DEFINITELY be something I do. (I may also make that shaft for the crank handle out of aluminum, but I realize that's not an option for everyone.)

Would you be opposed to me forking the design?

My goal was to keep the Curta as close to the original design as I could. I have no problem with you forking the design.

That's fair. Thanks!

Comments deleted.

What 3D printer did you use?

For the first Curta I built, I used a gMax 1.5 printer that I sourced and built myself (my first printer, but the second printer I built). The second Curta was built on a Triple C-Bot I built (very similar to the D-Bot here on thingiverse -- https://www.thingiverse.com/thing:1001065), but with three lead screws driven by a single stepper (the triple c-bot is here: http://openbuilds.org/builds/triple-c-bot.1757/).

D-Bot Core-XY 3D Printer

Simply amazing work, congratulations from Brazil !

Thanks, I appreciate it!

Just watched the 'Tested' package delivery/opening for Adam, I had been following your progress for quite a while and was happy to see you made one for Adam. What a masterpiece and a nice touch to deliver in the beautiful wooden box! I too would like to give this a shot with SLA at 2:1, but know once I get started, I will be in for a long project. I can't imagine the time you put into modelling the 600+ parts it took to build this beautiful computer, I just wanted to thank you for sharing such a project! Stunning work all the way around, maybe the best 3d printed project I have ever seen! Thank you, -j.

Yup, I have made two now. The first one for myself (currently unpainted), and one for Adam. The box was an idea I had after watching him get excited about weathering a box he had custom made to fit something. He talked about his love for boxes, so I knew I had to deliver the Curta in one.

2:1 should be a challenge -- I look forward to seeing someone log their success or why they fail. Failures can be very enlightening.

Modeling took about a month. Luckily for me, many of the 600+ parts from a Curta are duplicates or near-duplicates. There were only around 100 unique parts to model and many of those were pretty simple. Some that I enjoyed doing were the teeth of the step drum -- I made one sketch with all ten teeth and then extruded each slice of the step drum by selecting a different number of teeth from the sketch for each extrude.

Thanks for your interest and good luck on the 2:1 scale Curta. Please keep me informed with progress!

Hi, how much it weight ? Just to approximate the number of spool I'll have to buy..

The weight by itself (3 lbs or 1.36kg) is a little bit misleading -- the weight includes a bunch of non-printed parts (screws and nuts mostly) and doesn't include a bunch of support material that is required. It can be done in about two spools, but getting three would be safe.

What type of plastic did you use to print this? Just standard PLA?

Yes, standard PLA

Is there a description somewhere of your hardware tools, software tools, and process?

Yup, check the build manual listed under the files. There is also an instructable with basically the same content as the build manual

Just seen this on a Tested, what an amazing build Marcus. I'm proud to say I own an original Type 2 Curta in my collection of oddities. I tried to download the BOM but get a 404 error, any chance of a new upload of the BOM please? Many thanks.

Thanks! The link in the description was old because I had updated the file. The BOM in the files list was correct. I've updated the link.

Many thanks, I should have just downloaded all the files in the first place.

When I get my 3d Printer back to a working state then I'll do this!

How did you get that perfect finish on this?

I mostly followed Joel Telling's video on painting Harry Potter wands (https://www.youtube.com/watch?v=pnqntteQx80) and added a bunch of extra sanding between coats of primer paint the actual paint, and a clear coat. Between the black and the clear coat I did some wet sanding with 1500 grit sand paper.

Firstly this is awesome, wanted one of these since I knew they existed. Thanks for putting in the hours of designing and sharing

I was wondering what is the odds that I can print these parts at 60% scale? I have a custom prusa i3 style printer set up for extreme tolerances and small parts so using a 0.15 mm nozzle I am confident that I can print the parts to fit, the only concern I have is with ratios and the moving parts. If my logic is correct, it should work as long as everything is scaled equally?

once again thanks for sharing!

60% is close to 2:1 scale instead of the 3:1 scale it is currently at. The hard part about scaling is that you would want to scale everything except the tolerances I added between parts -- otherwise there would be a lot of work to manually fit parts (there already is). To do that you would need to visit my OnShape 3:1 Curta document and make a copy to scale it down. I modeled the parts to 1:1 scale, then I used formulas for most of the tolerances and finally applied a 300% scale transformation so it shouldn't be too hard to alter those tolerances and the scale operation to be for 2:1 scale... not hard, but a lot of repetitive work.

Some concerns I have about scaling it down are:

  • Part strength -- particularly the main shaft. At 3:1 scale I broke the main shaft a few times when I over-stressed it (mainly because parts were misaligned, but still...). At 2:1 scale it will be easier to over-stress it. Also at 1:1 scale some of the parts are too weak to function properly in plastic such as the selector shafts. I'm unsure how that would work at 2:1 scale.
  • Transmission Shaft rigidity -- The transmission shafts can bow under enough pressure and skip teeth on the step drum. That causes the result to be off by however many teeth are skipped for that digit. This usually happens during carry operations -- especially ones that cascade through the digits such as for subtraction. At 2:1 scale the transmission shafts will be thinner and even more prone to bowing under pressure.
  • Non-printed parts -- I spent a lot of time finding the right springs. The spring rates don't always seem to quite scale directly and I had difficulty finding acceptable equivalents at 3:1. You may need to experiment to find the right ones at 2:1. I also had trouble finding some of the balls used and ended up using a different size for the ones I used in the selector knobs in order to make it easier on others which required altered models. Finally, the screws and nuts used are not scaled exactly 3:1 since for many of the screws it would have put them between standard metric sizes. At 2:1 you'd have to walk through the models and engineering drawings and alter sizes to match the closest metric size to the scaled screw size. For instance, the nut at the top of the reversing lever is M1.4 on the original. At 3:1 it would be M4.2 so I altered the models to be sized for M4. At 2:1 it would be M2.8 so the closest would be M3.

All that said -- don't let me stop you if you want to try it. There were a few posts around when I got started on this project that said 3D printing a Curta couldn't be done.

Thanks for the quick reply, it would be a challenge to print a smaller one, ideally, I would like to print a 1:1 scale Curta but I'm pretty sure 3D printing will not be capable of that for another few years. The strength of the parts you mentioned would always be an issue and thanks for sharing your experiences.

As for the springs, I have a few suppliers that should give me enough variety to find the correct springs. The other hardware such as the nuts and bolts, I am considering to adjust the scaling, for example, scaling to 65% instead of 60 to compensate.

thanks again

Yea, I think a 1:1 scale Curta may require the materials that it was originally made with. Some of those materials can be 3D printed today, but I'm not sure it could be done at the size with the precision that are required. Even if it could be done today, those machines are way out of my price range. If the tech isn't up to par today, you are correct that it might be in a few years, but even then the price would likely still be out of range. I do have some ideas regarding a way to approach 1:1, though. It requires a combination of techniques / processes and only a part of that is 3D printing.

McMaster Carr is the main place I knew of to look for springs. I ended up finding a couple on eBay. The hard thing about McMaster Carr is that it's all in imperial units (I'm an American and I really wish we'd just get on the darn Metric bandwagon). I did find some places on the net that had springs closely matching what I needed, but they only supplied in bulk... grr...

For the nuts and bolts I recommend thinking of it as calculating what standard size is closest to the scale you're attempting rather than scaling them to a particular percent that matches a standard size. It may already be what you were planning and it's a small difference with the same result. However, it makes it easier to think about which means less likelihood for a miscalculation.

I just saw the video of Adam Savage receiving one of your Curta computers and realised I haven't responded to your comments.

I am currently still studying engineering so I have access to some interesting equipment and people with the skills to reproduce a 1:1 scale Curta, although this would take a fair amount of convincing and I would need funding. So, for now, 3d printing with some machining is the best I can do.

I am in South Africa so large chain hardware stores are almost nonexistent but smaller owner run and own hardware stores are generally quite useful for finding the correct springs so it will take some shopping around but I think that part will be possible.

I was considering scaling all parts before starting construction and possibly assembling the calculator in CAD software to check for clearances, tolerances and ensuring the nuts and bolts will fit. This way if I print all the parts after scaling and the virtual assembly I should avoid miscalculations.

Yes, a virtual assembly helps a LOT. I avoided many time consuming reprints by doing that... although I still had some time consuming reprints I had to do, just not as many :)

I bought a 3d printer just to do this project. I have a little more work to do to upgrade the printer before I'm happy with it, and need more practice before taking on a project this big, but I'm anxiously awaiting the day I start printing...

Bravo on taking on such a big project for a first experience! I'm very happy to have inspired so many people.

I'll be watching the Curta thingiverse discussion group. If you run into issues you need help with, I have a troubleshooting topic or you can just pop in to show off progress or an end result.

I am soooo doing this one. I've been fascinated with those for years, and unfortunately they are quite expensive. Thank you for your work kind sir. As soon as I finish printing these accessories and prusa printer parts.

Awesome, I look forward to another Curta make!

It's people like you that cause unrest. I am obsessed with the Curta calculators and now I have to print this and play with it. I can't stop thinking about it. BRAVO!

Here via the Make:magazine article, and from there the reddit post (which i sub to but apparently missed when it first went by). Just saying hi and congrats for a great model, it's truly a demonstration of how far the tech has come and what patience and determination can do to your hobby. I hope to get around to building it one day!

Welcome and thanks for the compliments!

Hey! This is probably my favorite thing on thingiverse, and my math professor dad (who I want to give this to as a gift) loves stuff like this, do you think that you'll ever make a version that is the original size?

If I do, it almost certainly will be made from the original materials -- I ordered a 1:1 scale selector shaft and selector knob from Shapeways. The parts work, but the fingers on the selector knob are so small and fragile that it wouldn't hold up to use. The selector shaft is like a toothpick made from plastic with the helical groove and divots in it, so I'm afraid that would snap in time too.

I have a 200x 200 build plate with a 180 height, what items need the 220 height so I may find a way to print diagonally or outsource to get the full height?

The update is now up. There is now an option to print the step drum in two parts -- Step drum upper and lower. There is a pin to combine the parts (print three of them). Just use a bit of superglue to combine everything.

I have an update coming to reduce the height requirement on the printer from 220mm to 170mm. See https://twitter.com/marcuswu/status/904507501181358080. Unfortunately, it will require another day or two to work out an error I introduced before I can test and then release those updated files.

I have a extra large height with a new printer, but now I'm just going to be closely watching while I gather resources and start fall break when I can get a nozzle set for small parts and have the printer steps calibrated. The thing is once they are, I should not adjust it until it is fully printed.

but bringing the printer settings closer to the mean will increase popularity.

I wish I got to have my own tv show for years then get to buy whatever I want fr my own, slightly smaller tv show.

The models should be printable on a 0.4mm nozzle which isn't anything special. The design did originally have a few parts that required a 0.25mm nozzle, but I updated the design between my Curta and Adam's to remove that requirement.

The main shaft and step drum requires the height. The upper and lower housings require the full width and depth specified while the lower housing requires 170mm of height. The transmission shafts need 170mm of height as well. After the lower housing the next widest / deepest part (the upper sleeve) requires 143mm of width and depth.

Fantastic project! I too came here from Adam's unboxing. Question: Is the main axle and step drum supposed to be printed as one item? I have started printing the parts now, and have completed the bottom housing (almost 50 hours) and about 15 other pieces.

Currently the step drum and the upper portion of the main axle are combined into one part. The lower portion of the main axle is a separate print that slides into the bottom of the step drum.

The main axle and step drum combination are apparently taller than many 3D printers can print, so I have a version where the bottom part of the step drum is combined with the step drum and the step drum is divided in two near the top. This would reduce the required printable z-axis height from 220mm to 170mm (see https://twitter.com/marcuswu/status/904507501181358080). That design also significantly reduces the amount of support material required for the step drum which also reduces the print time. I think the original design required around 12 hours of print time for me. It is now around 6.5 hours.

Unfortunately, while putting it together yesterday I noticed that I introduced an error into the angle of the main axle when I combined the two so I cannot release those files yet. If you can print the existing models, I suggest you do that unless you can wait another day or two on the updated models.

I use S3d and separated the two pieces and printed them individually but I think I will wait for your revision and see which works best.

How many individual parts make up this print?

Dare I ask how long it takes to print all the parts ?

I calculated it out and it's around 9 days, 14 hours, and 44 minutes non-stop printing on my printer. That's assuming no failures and that someone is there to immediately remove one print and start the next. Filing and sanding the parts to fit properly and assembling it takes longer.

Very nicely done. Really beautiful work.
That's a lot of parts surely requiring a lot of work to get things to function together. I will be hoping for some printing time soon to have a go at this myself.

Awesome, thanks!

Note to everyone, I have a thingiverse group (https://www.thingiverse.com/groups/curta) where I would love to see progress photos and any questions people may have building theirs.

Hey Boss,
I'd really like to print one, but I'd like to print the most up-to date version. Any chance of an update soon? I'd be more than happy to work on some issues you're having. I'm proficient with Fusion360 and have 3 printers of my own. Currently work as an R&D engineer and just love the history behind the curta. Ever since I saw your post on Reddit, I've wanted to print one and the time is now.

Thanks again for your work, Luke

Just uploaded an update. The BOM is updated, paint masks are added, and an initial draft of the build manual is added. As far as I know, these files should now be correct and the main changes left are to the build manual as I find things that need better wording and as I am able to add more images to it. Feedback on the build manual is desired and welcome.

You are the man! I can't wait to get started. Thank you much!

I do have some updates coming soon as well as paint masks for numbering and lettering on some of the parts. I was out of town for a bit for a wedding so those have been on hold. I will try to get those uploaded today.

If Thingiverse counted multiple likes, I'd click this project every time I went on line. Amazing!

Thanks, I appreciate it!

I've updated the description of this thing to list a poll (https://docs.google.com/a/digitaltorque.com/forms/d/e/1FAIpQLSeJQcq5njLh1_Z-IPOROXWmYAR7QqbMc7tr924hxTVYikhf6Q/viewform?c=0&w=1) on how you would like to see build instructions for the Curta. Please fill this out -- it's just one question.

Wow! This is incredible.

I am waiting until the instruction set is complete, but I am definitely watching this. If you have a video I would love to see it in action.

I finished the second Curta and took tons of pictures on interval while working on it. I will be using those to document the build process soon. Here is a vid I took recently: https://youtu.be/ShFkJgck6Pw

Really great that you have made the effort.
I was trying to sort the files according to the POM and have few problems. I couldn't find few files (and obviously had few left that didn't fit into the POM table).
item # 2 --> I found 10061, but 10029 not really
item # 70 --> is it the "digits axle"
item # 71 --> ?
item # 72 --> is it the "main shaft pin", there is no number so I can't compare it with the original drawings

I'm thinking to build it or at least give a try. I have ordered almost everything except two springs.
How strong should "torsion spring for zero position lever" and "torsion spring for anti-reversal pawl" be? I have real difficulties to find the matching ones and was wondering if I could bend something similar by myself.

Do you have more pictures from the assembly process? Clearly it will be a huge effort to write instructions, thus I would appreciate already when I could see some pictures of the assembly process.

Thanks a lot!

10061 is the selector shaft and 10029 is the cylinder that goes around it with the number markings on it. I combined the two and then split the selector shaft in two (to reduce the support required). The selector shaft is named Selectorshaft-_selector_shaft_bottom.stl. The top of it is called Digit_Selector_Axle.stl.
Yes, #70 is the digits axle.
71 and 72 are mainshaft-_Main_shaft_pin.stl and Part_Studio3-_main_axle_dowel_rod.stl -- the longer one is to connect the main axle to the crank handle and the shorter one is to connect the zero positioning disc to the main axle.

The two torsion springs were the most difficult for me. I ended up making them myself the first time around, but not in a way that would be easily reproducible. I did some searching and linked to mrspring.com and McMaster Carr to items that should work, though I have not tried them yet. I'm working on building a second Curta -- if they don't work, I'll update the BOM with something that does.

I don't have a lot of pictures from the assembly process yet, but I will be taking a lot in this second build. I'm also working on assembly renderings from the 3D models.

Comments deleted.

What a terrific project. I'm starting now on a 3D print of the 3X version but wondered if you had given any thought to machining the 1X scale model from aluminum, steel, or brass bar stock? The idea intrigues me though it seems like an ambitious machining job.

I would love to, but I don't have the tools (Mill or lathe) and I don't have the experience working on them either.

Thanks - I have both lathe and mill, but as I saw it looks like a really ambitious machining job so I'm starting with a 3:1 scale 3D print first on my Zortrax M200 printer. So far it looks like larger pieces are best done in PLA+PHA because that doesn't warp, and smaller parts are better done in ABS because those print better with my printer. How well does your 3D print model work?

Yes, it is a pretty ambitious machining job, but many of the parts are pretty simple with a few complicated ones in the mix (the selector shafts come to mind). I hope to one day learn to machine and build a 1:1 scale Curta.

Mine was working quite well until I made the mistake of lying it sideways on my desk. It rolled off and landed with a sickening crack. Multiple parts are broken, but I haven't had a good chance to reprint and repair because I am busy working on a build manual and a commissioned copy. I don't think I will do any more commissioned copies for those reading... I didn't think I would even do one, but that is a whole different story.

For me, some of the harder parts to print are those selector shaft parts and it seems like those would be really simple to machine, even at 3:1 scale so I may give that a shot if the printed parts have problems.

Really sorry to hear about your damaged model. I did the same thing with a large modular arm but only a couple parts broke on that so fixing it wasn't too bad. It's especially sad on something as complex and time-intensive to build as your Curta calculator so good luck on getting it repaired.

It must be tough to get a fair price on commissioned copies of the Curta - most folks have no idea how much time goes into the printing and post-processing, let alone the time to design it in the 1st place.

A fair price for a 3D printed Curta with all the time necessary to manually fit all the parts would run (significantly) more than a real Curta. In this case I'm doing it for someone I couldn't say, "No," to (in fact, I was honored to say, "Yes
) and plus I get a trade for a real Curta and some publicity too which will be fun.

I really didn't have much trouble with the selector shaft parts. The peg on the top of the selector shaft and the peg at the top of the bearing that the selector shaft sits on did break a few times. I printed those at 100% after having them break. I have considered (and still am) printing them with a hole instead of a peg and fitting a pin in its place. The pin could be a metal one, but I've had good experience with printing pins horizontally that are pretty strong and that would keep more of the Curta 3D printed.

Not really a big deal on breaking my Curta -- I'll reprint the parts when I get time and I've gotten pretty good at taking the thing apart and putting it back together. Really, the time consuming part has been printing and fitting the parts (more fitting than printing). I am preparing the external parts of my Curta for painting and that also takes significant time to get it right (especially with the knurling on it). I also need to find a good way to get custom made decals or paint masks to get the numbering and lettering on it before I can consider mine done.

Comments deleted.

Nice that you get an actual Curta and publicity out of the deal.

My Zortrax printer maxes out at something like 80% fill (they only have options for low, medium, high, and maximum) so solid is not an option. I also thought about using metal pins on the selector shaft parts if the plastic ones break off. It would be easy to drill holes in the applicable parts with a lathe and there might be an appropriately sized dowel pin available. By printing those parts horizontally, do you mean on their side? I don't think that would work at all well on the Zortrax but maybe you mean something else.

By the way, which brand and model of 3D printer are you using?

Yes, on their side. I can't print a larger diameter than about the size of the support columns on their side due to the angle of overhang at the bottom of the cylinder diameter and I have to use a raft or a big brim / skirt. I also slow down the print to reduce the likelihood of longer cylinders warping with the small contact area. For a small pin like the ones we are talking about or the carry pins for the results dials, I print them on their side on a raft then clean them up with a needle file afterwards because it isn't the best geometry to print that way, but it does make the pin much stronger.

I printed the first Curta on a modified gCreate gMax 1.5 (not the plus version). I am printing the second one with a triple c-bot (http://openbuilds.org/builds/triple-c-bot.1757/) running on a replicape (http://www.thing-printer.com/product/replicape/). Not having a large print bed moving back and forth quickly on the y axis for small features or infill in tight areas is a help. I am also enjoying the reduced moving mass of the core-xy. I am printing at twice the speed and getting significantly quieter operation and a better quality result.

Thanks for the info on your printers. Looks like they (or probably the slicers used) provide some nice capabilities that I don't get with the Zortrax. Oddly enough the 10207, 10208, 10209, and 10216 shafts print vertically pretty well for me so long as the raft sticks to the print bed. I'm having major problems with 10220-410003 and similar sleeves for those parts though. That little V feature that protrudes from the inner sleeve wall toward the center of those parts fails to render in my slicer, apparently because the slicer doesn't recognize a wall less than 0.5 mm thick and can't make a wall less than 0.9mm thick. I'm fiddling with slicer settings and changes to the Onshape part files to see what I can do to fix that but am not hopeful.

I just updated those parts (10218, 10219, 10220, 10221, 10222, and 10230) to be more easily printable with a larger nozzle. On my slicer I did have to reduce the extrusion width some, but it should print without a 0.25mm nozzle now.

Thanks - this comes at a good time for me as I have time to try those parts again and was about to see what I could do in Onshape. I'll report back in a day or two on how it goes for me with the new parts.

I've already printed the zero positioning disc and those little teeth seemed to come out pretty well for me. Time will tell if it works as good as it looks.

Thanks again for the revised parts - I really appreciate the effort.

I have been busy printing a second Curta and have been needing to print those for a while, but my 0.25mm nozzle is clogged and I don't have a tool that small to clear it out (tried some other methods already). So the updates weren't entirely altruistic :).

Currently I'm printing the large lower housing. On my old printer that was a 60 hour print. This one should finish in just over half of that time... assuming it succeeds.

There is no reason the curve needs to be a single wall (I just followed the engineering drawings directly) -- it can be capped and filled so that the outside of the is a cylinder and between the 'V' and the cylinder is filled with material.

In addition, there is also no reason that the spacers that make the part thicker can't reach all the way to the bottom of the part so that there are no thin portions of wall. I avoided that issue since I have a 0.25mm nozzle. If I get a moment, I can make those changes or if you have already made them, if you send them my way I'll update the files and add credit to the description.

In fact, there is only one other part I can think of I really needed the 0.25mm nozzle and that was on the zero positioning disc. The toothed portion of it that combined with the anti-reversal pawl prevents backwards rotation doesn't print with a great edge on the teeth when printed with a 0.4mm nozzle. It may have worked, but I didn't want to take a chance on it. My new printer got better definition with a 0.4mm nozzle, but I'll have to test it to see if it's good enough. I'll be able to check that out soon.

Comments deleted.

You just made my dream ! I have always wanted a curta !

For anyone watching, I just updated the printed files in the BOM to include details on how they should be printed -- layer height, infill percentages, etc. For parts that I said to print horizontally, I did that for strength even though they don't come out as round. Sometimes filing is necessary to smooth those out and fit them with other parts. There are two types of screws that need to be printed -- one for each of the selector knobs and one for the main crank handle. Those along with the tops and bottoms of the frame supports need to be threaded. I encountered a lot of splitting of parts when putting them through a die before I found that if I filed the layer lines down, it eliminated that problem.

I hope to begin writing up an actual build document soon.

Awesome job! I just decided to search for this to see if maybe someone had made this amazing piece of engineering printable and lo-and-behold you put it up just 5 days ago! I know what I'm printing next!

This is Amazing!

Thank you very much mwu for sharing! I've been watching your updates on your blog for months, hoping someday soon you'd release the files.

Mechanical Calculators are hobby of mine, and I've always dreamed of seeing and using a Curta in person, or, by some miracle, owning one myself. Looks like my dream just became that much closer to reality!

This is amazing work! I think people sometimes underestimate the mechanical marvels these devices were, the Curta being the apex of their design ingenuity. Thanks for shining a light back on a page of history many have forgotten!

Many Thanks,
Dan I.

I wish I could have gotten it released sooner -- real life intervenes more than I would like :) I also wish I had a build guide, but that will take longer.
Thank you for your interest and support!

Quick question. Is it possible for the chrome steel 5mm ball bearing that's available on Amazon to replace the 5.3mm steel shot or does it have to be exactly 5.3mm? Regardless, THANK YOU for posting this (I'm Mr. Pute on YouTube who's been bugging you about it)!!!

The 5.3mm ball is used with a spring to snap the selector knobs (used to input a digit) to a number so that it cannot sit between digits. Given that its main function is just to sit in the detents in the selector shaft, the 5mm ball will most likely work.

To be sure, I can update the selector knob's hole to be 0.3mm smaller to eliminate slop caused by the smaller ball. The springs listed for the selector knobs are small enough to account for the smaller hole so that won't be a concern.

Good suggestion -- I was thinking that steel shot would be a problematic part since the only way I could find it was an entire bag.

Awesome! I'll hold off on printing that part until you make a modification. I'll order the 5mm ball and confirm whether or not it works properly after I assemble. I'm looking to make two of these (one as a Christmas present and one for myself) within the next couple of weeks. Thanks again!

It is updated now -- both the model and the BOM to reflect the new ball size.

It's very ambitious to make one of these before Christmas. Significant time is necessary to manually file and ensure proper fit of each part. I also had to do a lot of calibration to ensure carry operations work properly. Good luck -- Also I don't mind answering questions along the way.

Fantastic! Thanks! I'll keep you updated on how it goes.

cant wait to print this

Now I can own my own one. And I still know how to use it. Thanks ;-)

Now I can own my own one. And I still know how to use it. Thanks ;-)

So I'm not the only one who dreamed of owning an over-sized Type I. This is a work of art!

Thanks :). The over-sizing is simply a result of what my printer is capable of. If I had a mill and lathe and the skill, I would have made one at 1:1 as close to the original as possible.

No, it is perfect! I am a huge fan of up-scaled machines and cut-aways, so this is my catnip :)

Simply amazing work, congratulations!