Loading
elmerohueso

Ender 3 Lead Screw Stabilizer

by elmerohueso Sep 2, 2018
Download All Files

Thing Apps Enabled

Please Login to Comment

Anyone who designs this kind of constraint DOES NOT UNDERSTAND that the top of the z-axis is MEANT to be free-floating to IMPROVE print quality. RepRap designers had spent YEARS coming up with the free floating design.

WATCH THIS:
https://youtu.be/2Z7mZVvPlc8?t=480

I used another design to "fix" this on my Anet A8 and print quality went down the drain.
Better not do this.
These below-200-bucks machines are not constructed precisely enough for this kind of thinking...

I put one on my Ender 3 Pro and quality improved. Less vibration at all heights and it improved the stability at higher Zs. Just align it properly. Probably a bad idea on an older machine though. Start with a straight z screw or you might have issues.

i used a fidget spinner bearing

You can't do this! It is supposed to be unconstrained. You just ensured any imperfect screw movement gets transferred into the print!

The fact that this got 3000+ likes means people who buy Ender 3 know very little about mechanical design.

I was thinking exactly the same when I saw this. The way this part is done is overconstraining the mechanical design of the Z-axis.

Video that explains quite well why this "stabilizer bearing" is a horrible idea, despite the thousands of likes:

https://www.youtube.com/watch?v=2Z7mZVvPlc8&t=600

all that video illustrates is the need for straight lead screws and why this isnt a solution for bent or poor quality lead screws.

I was just speaking with the head programmer of a local machine shop (suggestions for schooling as I am looking to get into the field) and he was showing me the capabilities of the 5-axis mill. Every ball screw on the machine is captured at both ends, yet it delivers positioning accuracy of less than .0005". If your statement, or any statement implying a captured lead screw is an inherently poor design, was credible, I tend to believe a different system would be used on a million dollar 5-axis mill.

With a straight lead screw of good quality and proper alignment of all 3 points of contact (stepper, nut and bearing) there is absolutely nothing wrong with this design. The only limiting factors would be if your components have garbage tolerances from poor manufacturing or improper installation where all components are not properly aligned and concentric to rotational axis of the leadscrew.

To be clear, Im not referencing this part, but the concept in general

There is a difference between a precision machine and the ender 3. The issue is the the motor holds the end of the shaft a certain distance from the vertical rail, the nut does as well. If the distance between the nut axis and the rail, and the motor axis and the rail is not same (which it isn't even with careful adjustment, and it will be near impossible to achieve as the parts are not straight enough for that), the screw will not be parallel to the rail. As the carriage moves up and down, that will cause the opposite end of the screw to get closer and further from the rail. If you then fix that end it must bend the screw as it travels.
I suppose if the screw and rail are pretty straight and you installed this bearing and tested the distance to the rail with a stack of something hard and flat, with some feeler gauges and shimmed the bear stand until the gauges fit between the screw and the rail the same on both ends. Then run the carriage to one end and tighten the nut mounting screws. The might be a chance it does not bend the shaft significantly.
In the end though, what benefit does this provide? I don't like the end being free the way it is either, it just doesn't seem right, but thinking about it more, as it is, is the correct design on a machine of this ones precision.

Another option would be to fix the end with this bearing, and loosen the nut mounting screws permanently so that it can float side to side. You would have to design a way to keep them from getting loose, but still keep the z axis from having play.

first, I do agree with you for the most part. But I have a machine shop of my own and have been either reworking or remaking anything I dont like on my Ender. Not everyone is going to have a mill and lathe of their own, and I do realize this. But since I do, and I like to tinker with stuff, my Ender 3 has been going through some changes, and I have my own ideas for how I feel a printer should be built. I could have bought another printer with what I spent on linear rail both on the printer and what I have for planned upgrades. Is it necessary, or even worth it? probably not, but its going to be fun anyway!

Anyway, I solved the misalignment on the Z-axis with an adjustable bearing mount I designed (will post here once I make some small changes), an adjustable stepper mount I found here and a flexible coupler. The coupler wasnt really necessary, but I had it and used it anyway. I checked for straightness on my surface plate and the screw was really close, and after adjusting the bearing and stepper, its within .010" of parallel to the extrusion and it stays within that .010" while traversing the Z axis with the carriage. I think adding a damper to the leadscrew nut might help to absorb any force exerted by the slight bit or run-out in my leadscrew, but its probably not going to make much difference. I plan to machine a new back plate for the carriage, as I dont need the extruder mount and its a bit on the shoddy side anyway.

I cant say for sure if this has improved print quality, as I changed so much when I added the bearing its impossible to know if it did anything. But it does print far better than stock, though nothing on the X-axis is original at this point anyway. Linear rail, mount I designed for a pancake stepper and BondTech BMG, E3DV6 with microswiss heatbreak and copper heater block and nozzle, and dual 5015 part cooling fans on a mount I downloaded here (needs some rework, but its ok for now).

Not every Ender 3 is still inaccurate Chinesium ;-) Sane people probably just buy a better printer to start. But I had this already and its more fun when you start at the bottom and build it as you go! I cant wait until I have the time to add the dual Y axis rails and new bed. That will be the end of bed leveling for me. No Gcode, no touch sensor, just an adjustable Z-axis stop to set nozzle height. Hopefully it all works as it should.

Even .010" will cause bending of the screw. Basically if the out of parallel exceeds the clearances in the nut, bearing, and motor bearing (which are very small, all combined are probably .003") it will start bending it. It probably wouldn't matter functionally (you probably have it working without issue), but would cause a bit more wear on the nut and bearings and a larger load on the motor due to the extra friction.

Some friction could be beneficial though, particularly on the x and y. It would damp any overshoot, and improve print quality. The z won't see much overshoot because gravity is always keeping a force on the nut in one direction. I see overshoot in my prints at the corners (a bump that is out of plane with the rest of a flat surface right at the corner). In theory the software could drive the motor backwards (opposite the direction of travel) for a fraction of a second and just the right time to compensate for overshoot, but I doubt it does. It would have to be calibrated for the mass of each axis too, to know how long the back pulse should be.

In a technical sense, .010” of runout over 15” of leadscrew isn’t exactly horrible, and not near enough to cause issue with a leadscrew nut as sloppy as this one is.. With just the nut on the lead screw, it rattles when I shake it. I can twist the nut side to side by at least 10 degrees. An anti-backlash nut has more friction than what I have by far. The “precision” label on this thing was strictly a marketing term and in no way a description of the actual product. But it works, so I leave it alone. I have a pair of screws for the dual z axis upgrade I hope to get done at some point, and they’re definitely a bit better.
And .010” is being generous, it’s probably less. It’s not as though anything on the entire machine is within a .005” tolerance anyway. I would almost guarantee that the extrusions aren’t any better. Even US made stuff from 8020 is only .005” per 12” if I recall correctly .
That said, the less you have moving around on the printer and causing vibrations, the better your prints will be. You’d need a significant amount of lead screw runout to cause issue if the 3 contact points are in alignment. I’ll take whatever binding is happening over misalignment that visibly moves the z-axis screw as the gantry travels up and down, which is exactly what mine did. Between that and the pendulum effect of the screw being free to move, it wasn’t helping print quality by any means. And print quality out of my Ender was a major disappointment from day one. It’s worlds away from where it was, and I’m very happy for that. It prints near flawless 90% of the time, with the other 10% being due to someone not checking belts and rollers often enough, and probably a little to do with the Lack enclosure not being the most stable thing on the planet by any means.

I’m going to put my bearing block up in a day or two when I have time to fix the holes I made a touch too small. I’ll let you know so you can see what I’ve got going on and let me know what you think. Regardless if you agree with me or not, another brain sharing it’s opinion is always a good thing.

Is your overshoot on a belt driven axis or leadscrew? I notice that once in a while myself, though not so much anymore. I have been keeping my belts really tight, probably bordering on too tight, and it’s been printing better. And I’m not worried about stepper bearings, they’re cheap enough not to worry about too much. And I bought a bunch when I killed my extruder motor.

Lol, stabilizer doesn't even come close to fitting on my Ender 3. Upright to center of my Z lead screw is 18mm. Contact of upright to center of the stabilizer bracket bearing is 23mm. 5mm too far out to work..

In this application, Creality used the lead screw correctly. The lead screw should not be restraining the bed, which is why it's only restrained at one end. By adding this mod, it will be over-constrained. The lead screw is only meant to hold the bed up, not effect it's position. If this does in fact improve your print quality, something else is wrong. I do 3D printer design, I'm not just some uninformed user.

I totally agree.
In my case the real problem was the misalignment of the parts motor / screw / T8 nut.
Lead screw alone was really straight.

With this part you solve the symptom not the disease !

You are actually solving the problem of the lead screw wobbling while printing at high speeds, and it should not be used to correct bent lead screws or improper alignment of components.

Printers are better if they are more rigid, this is not a mystery. If your Z-axis lead screw is wobbling from vibration transferred through the printer, it will impart force upon the carriage in the X-Y plane. if it is constrained at both ends, you have eliminated this issue (in a practical sense). How much of an issue is it? I dont know for certain, and not enough for me to put forth the effort and figure it out.
That said, a properly implemented captured lead screw is certainly not going to have any negative effects, and could provide an improvement in print quality.

Could you share links to some of the printers you've designed? Some pics of the finished builds? Maybe a vid of them running? I'd love to take a look.

Agree, this should not be used. Adding flex fitting at the bottom could be another thing to reduce over-constrained system.

If the screws aren't bent and you take the time and have the ability to align things properly there is zero negative effect in doing this. I modeled up some for my CR10S-Pro and they work perfectly.

Great if it helps you. In Tronxy X5S I had to remove factory supplied top Z mount to eliminate Z wobble and saw noticeable improvement in prints after that.

Hadn't thought of it that way, thanks for posting this. I noticed my lead screw is probably 1mm off along it's length and thought this would be necessary, but now I'm going to leave it alone

Absolutely correct. Other 3D Printer brands such as Prusa don't do this either. A small amount of movement on the top end of the lead screw is normal. If the movement is large then there are other issues at play :)

I am trying to print the screw stabilizer, and before I do that, I look at where it goes and see no screw hole. Could you tell me how you put the stabilizer in?
Thank you!

As mentioned in the description, hammer-head nuts, also known as tee-nuts, or drop in t-nuts.

you have to push it really hard

Yup. And all the pictures and "makes" are photoshopped.

/sarcasm, it's been a long day

What exactly doesn't fit? If it's the bearing, you may want to confirm you're using the right kind of bearing and that your printer isn't squishing layers to make the fit even tighter.

It can or cannot improve your print quality, or even make it worse. If your lead screws are really bend and are moving alot at the top it's better to not fix them as the irregular movement will get transfered to Y axis and result in a bad Z wobble. Great model though :D

Yah, this overconstrains the setup.

Will it work for the CR-10S ?

Specificky, is it a better version of this?
https://www.thingiverse.com/thing:3377129

Z-Axis Rod Support for Synchronizer (Anet E12)

I didn't base this model off the one you linked to, but if the CR-10S uses 20mm extrusions (for proper mounting) and its lead screw extends past the top, it might work.

This should not be needed, might even hurt.

Comments deleted.

I need some help with this one.

So -- maybe I don't understand. I've got the bearing, had to hammer it into the support (it was tight but not too tight for some sanding and a rubber mallet)

Tested the fit, and it fit fine.

However, the internal part of the bearing isn't rotating (Might've busted it, hope not I'll try again I suppose)

But it also seems like the Z axis screw wouldn't rotate it anyway.

Am I unclear as to how these components interact? As a curiosity, is the bearing intended to spin rapidly or freely or is it just supposed to be able to turn if necessary?

Please advise.

The inner part of the bearing should spin freely. If it doesn't, you may have broken it when hammering it into the mount.
Note that, if the fit was so tight that you had to use a hammer to get it into the mount, you likely have some settings wrong on you slicer or the printer itself.

I do believe that these leadscrews are meant to have play because of the quality of the screws

If your leadscrew top is clearly moving around, you definitely have a problem. That is fixed by replacing it with a straight lead screw.

If you use a design like this with a bent lead screw, that basically "forces the bend" into where it connects to X axis, potentially making print quality worse.

If your leadscrews are ok, I don't think this should make any practical difference.

If you have ever looked at the CR-10s lead screw design you would see that it uses a similar design to this, except the CR-10 has a slot to allow play. You're right about using this design with a bent lead screw, but most of the lead screws used on 3d printers of this capacity have a slight bend in them. I think about how I couldn't get a good print with this printer for 2 months because of a cupped build plate until I finally upgraded the bed. This is a basic, intro level hobby machine and I don't believe Creality invested in premium screws. For some of the people that lucked out and happened to get a perfectly straight screw this will probably help, but for the rest of us it would be of much detriment.

It made a huge difference in the side surface quality of my prints. If it does for you, great. If not, my feelings won't be hurt.

Comments deleted.

Excellent (even it's not easy to put in place)! Many thanks!

Absolutely brilliant - I never even considered this being an issue, but as soon as I saw this it was obvious. Thanks!

Comments deleted.

Am I supposed to drill holes for the screws?

No. As with most things that attach to V slot extrusions, you should attach this to the V slot with hammer nuts.

what type of hammer nut should be used?

... sorry I had a little moment there lol thanks

Does this use Z Stepper Motor shims? Some ENDER 3 printers seem to have the Z stepper slightly mis-aligned, which causes the Z lead screw to run slightly out of parallel with the frame upright.

My printer didn't need shims on the Z motor. Either way, though, the stabilizer is based of measurements from the X-gantry at top position, so it'd probably work regardless of whether the motor needs shims at the bottom.

Is it compatible with https://www.thingiverse.com/thing:2925230 or even recommended to use both? I have no idea which one to take or what to do :/

Ender-3 Z-motor spacer / support

Has anyone found the 608 tobe too tight on the rod

Yeah, cant even come close to getting it down the rod. What did you do?

Comments deleted.

Skateboard Bearing= 608-2RS Bearing
I spent some time looking to buy one until I popped one out of an old skateboard. Worked perfectly with the Z-Rod & the print. Thanks!

Thanks for this I have tons of skateboard bearings and was wondering if they would work.

You needed this improvement? Is there a visible change in prints after installing this?

Yes, there was a very noticeable improvement in surface smoothness in the Z direction. Prior to the modification, I would see slight bulges, similar to a layer shift but uniformly distributed around the perimeter, every few cm. resulting in rougness which was mostly noticeable on flat surfaces. I thought this was due to dirt on the lead screw, but the problem remained no matter how much I cleaned it.

I printed a 20mm x 20mm x 100mm cubic tower before and after the modification and was pretty blown away by the difference it made. Unfortunately, I can't find them now to take pictures.