TAZ 5: Z-Wobble Fix

Hey 3D printing Folks,

I wanted to share my experiences with Z-wobble on the TAZ 5 with you, what the cause is, and how to fix it. I am Dutch, so I will try my best to explain it in English. I experienced major Z-wobble on my TAZ from the beginning. It was really annoying since I also have 3 Ultimakers which don’t have this problem at all. But as I have an education in mechanical engineering, I found a solution to this.

At first I checked what the pattern of the Z-wobble was by printing tall cilinder shaped objects. I found out the wobble was not really a wavy pattern that happens to certain layers, but the pattern was a spiraling effect. I found out the pitch of this spiraling effect was identical to the pitch of the threaded rods. So this got me thinking…

As you may know, the TAZ 5 has two threaded rods which move the X-carriage up and down. The threaded rods are the cause of the Z-wobble, and here’s why:
Threaded rods do always have a slight bend, this can of course be unnoticeable when this bend is very little. But some rods are bended more than others. Another thing is, the threaded rods in te TAZ 5 are fixed with two ball bearings each, on the very top of the machine and on the bottom, close to the stepper motor. Those ball bearings fix the movement from the rods in X and Y dimensions. But this causes problems to the x-carriage. Remember that the threaded rods are bended, when the stepper motors and rods start turning around, this bend also starts to turn around, like a jumping rope on its side.
Because the threaded rods are fixed in X and Y directions, those bends start to apply forces on the x-carriage. So the whole x-carriage starts to move. And this is why the z-wobble happens. The direction of the force rotates with the threaded rod, so the z-wobble effect also rotates which causes the spiral effect.

The fix:
Since the threaded rods don’t have to be constrained in X and Y direction, you can safely remove all 4 ball bearings on top and bottom. The threaded rods only have to fix (and move by rotating) the X-carriage in Z direction. The X-carriage itself is already fixed in X an Y directions with the smooth Z-rods and linear bushings. As you remove those ball bearings, the threaded rods are able to move a tiny bit in X and Y directions which compensates the bend of the threaded rods. The flexible couplings that connects the stepper motors with the threaded rods also allows them to move slightly. Because the threaded rods are now allowed to move a little bit, they won’t apply forces on the X-carriage anymore, and the Z-wobble disappears.

I also switched the linear bushings for linear bearings (LM10UU) since the bushings had way to much play. This also made the prints look even better.

Long story short: if you have a TAZ5 (or maybe an older version?), you can easily fix the Z-wobble by removing all 4 ball bearings on top and bottom of the threaded rods. :slight_smile:

Sven van Dam

Is it the same problem with you as the photos on below?

I guess I have never seen any wobble in either of my machines, the Mini and the TAZ 5. Its called threaded rod but its actually precision rod as here

: http://www.protoparadigm.com/news-updates/accuracy-vs-precision-and-threaded-rod-vs-leadscrews-in-3d-printers/

When I think of threaded rod I think of all thread rod used to hold your wooden deck together or other construction jobs.

What the other poster is showing I don’t think has anything to do with wobble but settings.

@champagne: Yup, that is a perfect example of Z-wobble. Remove your ball bearings and it will be gone.

@wmgeorge64: Indeed lead screws and threaded rods are different. Didn’t know the english word for it though. But although not every machine experiences Z-wobble, some do. And removing the ball bearings will fix this.

@umakeprint Thanks man! The last question is which one i need to remove? All? the two on the top? the two on the bottom?

If your threaded rod is bent, replacing it with one that was straight would be the solution I would take. Replacing the linear plastic with your ball bearing ones LM10UU would make a big difference.
Removing support bearings is going to put a lot of stress on those stepper motor bearings.

@wmgeorge64 Got it~ But …can you show me how are the fine quality of the printed surface by photos? Because i never seen the smooth surface ever when i met my Taz 5 on the first time :frowning:

That is somewhat worse than normal Z wobble for a Taz. I would suspect you have something misaligned or bent in your Z motion segment. The Taz does have a small bit of built in wobble. So does the mini, but it’s less apparent because the mini doesn’t have the metal sideplates on the X axis. Most of the leadscrew / threaded rod based Z axis printers do to some degree. About the only ones that don’t are the belt drive Z axis units.

I’m working on an elimination method that lets you keep your bearings though. https://forum.lulzbot.com/t/anti-z-wobble-z-nut/2064/1

@wmgeorge64: Trust me, it will not damage your printer. There would have been forces applied on the stepper motors, only IF the flexible coupling wasn’t there. But the flexible coupling that connects the steppers to the z-rods compensate the forces by moving a little. The rods with the bearings are called overconstrained. Explanation according to wikipedia: https://en.m.wikipedia.org/wiki/Overconstrained_mechanism

I have no intentions of removing bearings.

I know this is an old thread, but some of this makes sense. My non Lulzbot printer is a Makeit pro. Their Z is driven truly by threaded rods, not precision lead screws. How they run the z-axis is a little weird, but this discussion tells me why they did it. The threaded rod (looks like about 10-32) is attached to the motors with a rubber tube and zip-tied. The threaded rod has a simple nut on each one that the carriage sits on in a recessed hole. The rods are really loose, but that printer gives me nearly perfect walls.

That is one way of addressing Z wobble. It leads to other inaccuracies though, especially at the far ends of the print. THe Taz used to use that exact setup actually. The problem with an unconstraiend far end is that the tube ends up acting as a rod height setting member so you can end up with a variable Z layer height by a few tenths of a milimeter. You can minimize that by putting the leadscrew directly on the motor, or resting on a metal ball but that still doesn’t address the issues of Z layer variability at the far end of the leadscrew due to shift arc. The best way to implement a leadscrew is the Fixed / Floating approach with an anti wobble device if necessary.

Thanks Piercet. I’ve started to gather the parts for your z-axis anti-wobble. Once I get everything, I may have some questions on actual disassembly/reassembly.

Here I am at the end of 2020, deciding to fix my Lulzbot Mini 1 z-wobble on the two y-sides of a calibration cube, aligning perfectly with the threads in the threaded rods, being thankful for you posting this back in 2015!