Prusa i3 Y-axis lead screw upgrade

by BetterHalf Nov 24, 2014
Download All Files

Thing Apps Enabled

Please Login to Comment

Dear BetterHalf,
I will be very thankful if you could you send me please (my email: [email protected]) ,the ball bearing clips ( .stl file) for the Y axis bed ( more exactly for the LM8UU) ? I have the Prusa i3 Hephestos 3d printer. I found many LM8UU clips on the internet, but I need with the exact distance from rod ( on which the ball bearing moves) until the underneath surface of the Y axis moving bed. I've downloaded the "Y-Bushing.stl" file (from the following link : https://github.com/Xssink/Prusa_i3x/tree/master/STLs) , but the distance between the holes is 28mm, and the required distance (in my case) has to be 20 mm. I guess you don't have a .stp or *.step files of this part. Hope to hear from you soon.
Thank you in advance. Kind regards, Emi

Hi I've printed all the parts but I've ran into a slight problem, the bearings dont fit in the housings any tips in resolving this?

A few things come to mind relating to calibration. The hole might be out of round OR both X/Y could be running even but a bit undersized. Print 100mm x 100mm x 10mm (or similar) calibration piece to check your XYZ calibration. You may also be over extruding a bit, which will certainly affect ID of holes, and that effect is more pronounced the smaller the hole is. Mark your filament and extrude 100mm, compare actual extrusion to target, and adjust accordingly. 608 bearings should be a snug press fit into the housing.

Thanks for that I'll print a test piece as suggested thank you

Hi, I printed all the parts for this and they are terrific. Can you please elaborate as to what you did to the firmware to accommodate the T8 leadscrew? Did you have to recalculate the movement per revolution for the leadscrew? Thanks.

Yes you will need to set the steps/mm in the firmware.

  • the simple answer for most people is 400 steps/mm for x and y if using the common TR8x8 metric lead screw and 1/16th microsteps.

  • The specific answer involves understanding the difference between pitch and lead.
  • pitch is product of the angle of the inclined plane that forms each thread
  • lead is measured as the total distance traveled in a line parallel to the screw if the screw is rotated 1 revolution.
    the ratio between lead and pitch is determined by the number of "starts" or independent thread profiles on the screw.
    a TR8x8 lead screw has 4 starts, visible as the 4 independent threads at each end.
    ~so a 2mm pitch multiplied by 4 starts = 8mm/rev lead.
    ~using the basic formula: steps_per_mm = (motor_steps_per_rev * driver_microstep) / thread_lead
    400 = (200 x 16) / 8
    NOTE: if your motor reads 1.8 degree per step then its 200 steps per revolution (360/1.8=200)

Thank you. In your tests, what sort of motor worked best for you? I imagine you would need a faster turning motor, or perhaps one with higher torque or some other parameter than just using the existing Y axis stepper. Was there a specific motor that works better or perhaps specific motor requirements in terms of max RPM/Torque? I have learned faster motors are needed (vs. higher torque) motors for something like this. Anyhow, any insight you might have would be appreciated. Thanks.

When this was first posted there were additional tabs containing instructions. Unfortunately they have been removed due to some sort of format change in the website. So here are some important bits of information about stepper motors.

  1. you will benefit from a larger/higher torque stepper motor. Speed is important to a small extent, but you will not likely hit the limit of the motors speed in a 3d printer.

    • http://openbuildspartstore.com/ sells a fairly large NEMA 17 with detachable wiring for less then $20. At the time it was the highest torque value motor I could find at that price and its been solid and reliable, no issues.
  2. Micro stepping..... you simply do not need 1/16th for leadscrews. At 1/8th you have a higher resolution on a TR8x8 compared to a standard belt configuration at 1/16th (200steps/mm vs. 80steps/mm)

  3. Jerk and Acceleration. This one caused problems for a long time. This goes back to torque and the effects of torque when changing direction.
    • Start with a jerk value of 10 (on X and Y) and go down as you tune to run at higher speeds. (@60mm/sec print and 80mm/sec travel we use a jerk value of 8)
    • Acceleration values in the range of 800 - 1000 mm/sec seem to work well so far. Ringing is minimal but there are slight rounded corners. more testing is needed.

*Note- Its been difficult to find a clean explanation of exactly what Jerk is, so here is our understanding: "The speed at which the firmware assumes a change in speed can be done instantaneously or without acceleration."

  • Lets break that down a bit with some assumed numbers: the axis is traveling @ 50mm/sec and has to change direction with an Accel value of 1000 and a Jerk value of 10.
  • The axis must decelerate to 1/2 the jerk value at a rate of 1000mm/sec/sec. So, unless my math is completely wrong that takes roughly 0.05 sec
  • The axis is then traveling @ 5mm/sec where is can make a change from +5mm/sec to -5mm/sec (direction change) for a total change in velocity of 10 (the jerk value)
  • The axis then accelerates to 50mm/sec, again @ 1000mm/sec/sec in roughly 0.05 sec
    *Regardless of being lower than belt drive values this is still extremely fast!
    -its generally good to use proportional values for jerk and accel when your looking for a starting point: 10 and 1000 or 8 and 800
    -when fine tuning change only 1 variable at a time and print the same model repeatedly.
    -fine tune your extruder before you worry about printing at higher speeds. A poorly tuned extruder makes everything look bad.
    Hopefully this helps with your understanding.

Thank you for this. Did you find an antibacklash nut was necessary or the ordinary T8 brass nut was good enough?

Comments deleted.

We have never attempted to use either type of brass nut.
We used the delrin nut plates from openbuilds.


Again, at the time they were easier to get from a reputable supplier. But we are on our 3rd set in 2 years, which is pretty decent considering the printer runs almost every day.

Thank you. I ordered a motor and a pair.of nuts.

I have been looking to do this exact same thing, unfortunately I have a "sunhokey" variation so I would either need to replace the front and rear acrylic with the threaded frame rods, or create my own setup. Same with the X axis build since I already have the acme lead screws for my Z axis. Thank you for such a great build though.

Don't see this question will this work for the Wanhao i3?

Not in its current design, as the bearing assembly for the Y lead screw would not directly mount onto the steel chassis of the Wanhao.

Hello! You've made some really cool designs! Thank you both!
I would like to try and use ballscrews instead of leadscrews. Do you think there would be added benefits in regards to backlash? I'm not sure if it would make much difference considering the low forces involved in a 3D printer vs CNC, and the fact that leadscrews already work and are cheaper.
The main advantage of ballscrews is that they're quiet - worlds apart quiet. But I'm interested if you have any experience or opinion regarding the ballscrew use in this case.

I do not see any drawbacks to using ballscrews, other than cost, but also cannot see any major benefits given the cost increase. If one were going for a "combo" machine with an interchangeable head, so that it does both additive and subtraction manufacturing, it would most certainly be beneficial. Ultimately, if your budget can afford the use of ballscrews and you feel like experimenting, I'd say "go for it!"

My stepper motors are still losing steps after various calibrations. Which motors do you have? 1.5 A or 2.5A

I have 3 axis trapezoidal screws (Tr8x2) my actual config is :

define DEFAULT_AXIS_STEPS_PER_UNIT {100,100,200,750} 1/4 microstep

we are currently using a nema 17 motor from openbuilds
it is a 1.68A motor/ Torque: 76 oz-in - 5.47 kg-cm (the largest nema 17 we could find)
stepper drivers are the standard DRV8825 w/ heatsinks
these drivers are a bit difficult to adjust without a multi meter.
for maximum torque your motors will be very warm but not hot to the touch. (approx 90-95% of rated current constantly)
for adjusting the 2 most common stepper drivers:

Where did you get your Y-Corners? did you design them yourself? are they available anywhere? they look pretty cool...
Also are they the 2xM8 and 1xM10 version? or the 3xM8, or 3xM10?


They were designed with the upgrade, are avaliable under Thing Files "frame foot". They fit the full M8 rod frame design. Due to the compact hole spacing, it is not able to be altered to accommodate the M10s.

oh, well they still look pretty cool... I just have to find M8 rods of that length... still deciding which version to make.. your design overall seems quite great. Thanks a lot

Not sure if you have the means or not, but if you're having a hard time finding individual M8 rods of that length, you can buy threaded rod "stock" (500,750,1000mm, etc), cut them down to length, and clean up the threads with a file. That's the beauty of the maker movement is making it work! :)

Yup I was considering that.


first great idea!!! Question, how fast is this solution?
I am building a very larger printer, I mean very large 800mm x 1500mm and seams that using this solution is far better than having a belt.
I am going to use larger robs diameters and will redo the design, but I was wondering what kind of speed could reach.
Maybe a stronger motor >48N/m would help?
Any ideas and suggestions?

Currently, we get around 60-70mm/sec (stable with 48mm Nema 17s). I would bet that a Nema 23 (vs Nema 17) would give you the torque needed for such a large build.

will look into the motor torque suggestion.
Do you have the files in Openscad?

Unfortunately, no. All our files were created in Autodesk Inventor (IPT format). Source files are available for download in Thing Files (zipped folder).

Printed the parts and the holes are too small. Could you be sweet-talked into doing a version for a Prusa with M10 all-thread throughout? I bought a frame/hardware kit from an eBay guy and really like the quality and how it went together. Now, I would really like to upgrade to this screw drive but need bigger holes.

Whadya say?

With the dimensional spacing in the current design, the M10s would simply not fit without a major overhaul. Could you by chance be sweet talked into buying some M8 stock and bolts and giving this design a try anyway? Whadya say? :)

I'd say..."well okay, ya silver-tongued devil". I found a local vendor for a couple of 1 meter M8 all-thread sticks with a handful of nuts and washers, and am going to give it a shot this weekend. I'll let you know what I end up with. (as of this writing, I have everything I need to do both X and Y upgrades).

Thanks for your hard work!

  • a -

Haahaaaa! Hell yeah! Looking forward to hearing about your progress! Please feel free to message us with any questions, we're here for you.
Also, once your upgrade is complete, we would appreciate if you published a "make" of both the X- and Y-axis "things" to encourage others who are interested to take the lead screw plunge! :) Thus far, we have has one user from Europe (johnr24) who has made a successful conversion with our design, and we are excited to see others as well! Keep us posted.

I'll be happy to publish my make when done. BTW, I sent you a message about an alignment problem. Did you see it?

We sent a message back about a half hour ago. It should have made it through? Maybe there's some lag in the messaging system, it's a little lack luster. But, I'll go ahead and respond here as well just in case. Basically, I think the issue is the difference in profiles on the linear bearing mounts (sometimes called bushings) between your kit and the Prusa i3x kit this is based on. We suggest printing off bushings from that kit, here's a link to the STL list (from reprap wiki Prusa i3x): https://github.com/Xssink/Prusa_i3x/tree/master/STLs
(file name "y_bushing")

Thanks for the quick responses. I guess I don't know how to get direct messages here. So, I got the file and am printing the bushings as I type. I had made my own bushing holders when I got the hardware kit and hadn't really looked at how the community did it.
Also, question about the feet; are the long-ways rods M10 and the crosswise M8? I notice that M8 rod under the bearing runners is a bit loose.

Thingiverse doesn't have a message system, per se, they just forward the message to your email and instruct you to visit the users profile to "respond". Yeah...lack luster.

Anyways, all the rods are M8. The holes may run a bit lose, as they were toleranced for a slip fit versus the tighter press fit of the smooth rods. Also, there's a bit of dimensional play in the ID's of any holes due to printer to printer variation. Should be alright, just position and verify parallelism and perpendicularly of your assembly before fully snugging down the retaining bolts.

So, it's together and here's what I had to do; The standard part you linked to doesn't fit the machining on the frame/hardware kit I got. But, it did give me the distance needed from the frame to the bearing. I drew up a shim in Sketchup and got everything to fit with a gnat's eyelash clearance between the frame and the bearing mount.


Does 3.00mm clearance between the bearing and the frame sound right? I just measured the printed part, but my printer is an inexpensive model and not necessarily down to really fine tolerance.

I ask, because it seems like it takes a lot of torque on the lead-screw to move things. Can't really tell where things might be binding.

Nice making it work with what you had! Understanding more how your platform is, we would have done the same in your situation! :)

Okay, as for the spacing between the bottom of the platform and the leadscrew bearing mounts, yes, there is indeed just a couple mm of clearance (according to my assembly file measurements, it's quite literally 2mm from bottom of plate to top plane of the mount...actual dim maybe +/- 0.2~0.4mm due to possible print layer height rounding error.)

Even with everything lined up perfect, the nut plate will take a bit to "break in". If there is a notable difference between resistance of turning the nut plate by hand on the leadscrew (just spinning the two components) versus hand turning with the entire assembly together (with a disabled stepper of course), then that would point to an alignment issue somewhere. If there isn't a large difference, it just may need to be broken in more. Ideally, during assembly, the lead screw should just "fall in" to the front and back bearings.

We used a Teflon based spray lubricant to get things going while hand turning the first few passes, then started running paths with the stepper. Given the tolerance of the lead screw itself, there may be rougher spots that the nut plate will find more resistance to as its breaking in. These spots may cause some skipping/stalling of the stepper.

If after ample breaking in you still see some stalling during motion, you can increase torque of the existing stepper by decreasing your microsteps (from 1/16 to 1/8 or 1/4). The more weight your entire bed/platform assembly is, the more torque will be required to push it. There's always the option of upgrading to a beefier Nema17 stepper (we currently push the Y axis with a 48mm vs the stock 30mm that came with our kit). They're relatively inexpensive at like $20 at most suppliers.

Finally, make sure that you have the voltage reference adjusted on the stepper motor drivers. Here's a helpful link from reprap: http://reprap.org/wiki/Pololu_stepper_driver_board

So, I got the Y axis done and the motor drives the bed frame from end to end smoothly. I had a couple of the 48mm motors on hand, so swapped out the original 'wimpy' motor before I hooked it up to the RAMPS and tested.

Now, I've jumped into X and have some new issues. I'll post them on your X Axis page for comment.

Thanks,000,000 for your help!