z wobble, but not bent?

My new CoreXY build is plagued with Z wobble, but I'm pretty sure it's not from bent rods. What else can cause this?

The design uses 3 8mm leadscrews, driven by a loop belt and one motor. They are supported at both ends (I already tried supporting at one end only). It uses 5/8" smooth rods (2) to constrain xy motion in the build platform. The platform rides on these rods with steel roller bearings.

At first, I suspected that my leadscrews were imperceptibly bent, or interfering with each other in some way. Version 1 had the leadscrew nut connected with a rigid tab to the platform. So I changed to V2, in which the leadscrew nut is supporting the platform with a bit of spectra line. This allows the screws to exhibit any movement they need to, but still support the platform in the z axis.

HOWEVER: None of this has helped with z wobble. What else can it be? The bearings don't seem to allow any movement, but they aren't real rigid... I figure with 3 points of support, they don't need to be. Am I wrong? Do I need to upgrade to linear bearings?

This is definitely not a bent leadscrew. The wobble seems to be random and not match the leadscrews threads. The only thing I can think of is the z axis skipping or extruder skipping. When you print a cube are the dimensions accurate? Another possibility is a skipping extruder. If the extruder skips, than the extrusion will be random thus the random layer widths.

Also if your using a E3d clone, That can jam and cause skipping. Buy a polished heatbrake or genuine heatbrake to prevent this. If its not a clone, ignore this.



UPDATE: I noticed near the top of the benchy (Steam pipe) that the artifact becomes less extreme. This is starting to make me think its a y/x axis issue. But Its core-xy so its one in the same. It looks as though when it needs to travel a further distance, per layer, that the artifact becomes more prevalent. If it is the x/y axis, Its not skipping steps because it recovers, and does not shift layers permanently.

This is definitely not an extrusion problem. Hanging the bed from strings (dyneema or gold or unobtanium or any other material) is not helping. Print quality problems are caused by unwanted motion of the extruder relative to the bed. You have to eliminate all such motion. Try wiggling the bed. If it wiggles, fix it. Try wiggling the extruder. If it wiggles, fix it. Fixes may mean replacing your end supported guide rails with fully supported guides. It may mean using two bearings instead of 1 at each rail. Try to get rid of printed plastic parts in the motion control mechanism. Plastic, especially printed plastic, flexes too much. Flexing is bad. And get rid of zip ties holding mechanical parts. Zip ties are the equivalent of duct tape. They are good for holding cables but not for mechanical parts.

Edited 3 time(s). Last edit at 12/30/2016 07:08AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Would you post pictures of how your Z axis is setup too? From the pics we can see the 3 lead screws run by a belt. But not the guide rails, or really the mounts to the bed.

Edit: It kinda looks like you have one vertical rod guiding the bed in the back left corner, and one in the front right. If that's it, that's not so good an arrangement. I'd expect it to see/saw on the unsupported corners. You may be best moving both of those rods to one side of the bed, say the back. Put one lead screw between them. Then one at each corner on the opposite side of the bed. You'd be even better off with a 3rd vertical guide rod between the 2 lead screws on that side.

Edited 2 time(s). Last edit at 12/30/2016 10:45AM by FA-MAS.

Thanks for your replies.

I tried printing some test cubes. I won't post the pictures, because they were FLAWLESS, and we've already seen plenty of perfect cubes. Now I'm really scratching my head. I'm printing a more complex test cube now. (Half done, and printing perfect...)

Zastin: it is an e3d clone, but with a polished heatbreak, so I don't think that's it. Your remarks about loosing steps made me turn up the vRef of each stepper. (it's louder now... we'll see if it's better).


DigitalDentist: First, let me say, your instructable was an influence on my design for this printer. Yes, it has printed parts, but I don't have access to a milling machine so plastic it is for now. I can live with those limitations. The zip ties... I'm ashamed, but I tell myself they are temporary. (edit: wrote "milking machine" by mistake don't have one of those either!) I did go around and tighten everything and found a little slop in one of the x-axis wheels. I know the strings weren't helping, but they were an attempt to satisfy myself that a misaligned leadscrew was not the problem. But the bed doesn't wobble... Not sure what else to do...

Fa-mas: Yes, there is one rod in the rear left and one in the front right. I really don't feel comfortable moving their locations, because they are perfectly parallel as they sit now. I used a "match drilling" technique to position them during the build process. I think I need to order some 5/8" linear bearings and design some heavy duty mounts for them. They aren't cheap though, and I want to isolate that that's the real cause before I purchase them.

In addition to this, I also switched filament (from matterhackers pla to Hatchbox pla, but both spools have given me great prints on other machines). I also ran PID autotune and updated firmware.

Any other thoughts?

IIRC, the Prusa Mk1 also had multiple Z rods powered from one motor via a belt.

It didn't work consistently due to slippage. That's why further designs drove the rods directly from multiple motors.

EDIT: Found a pic, not great, but you can see the right hand rod has a belt on it going towards the left Z rod.

[images.duckduckgo.com]

I salute you for trying to make the idea work, but it didn't then, and it looks like it still doesn't now.

Mind you, it might be worth trying toothed gears and a steel or carbon fiber reinforced belt. That should help, and might eliminate the issue completely. The SNAG is getting a belt just the right size... Might be a custom built job.

Edited 4 time(s). Last edit at 12/30/2016 02:40PM by DragonFire.

Dragon: I copied the design from the Fusion3 F306, which uses 3 leadscrews and a single motor in a similar configuration. They also use unsupported linear rods in 2 opposite corners of the bed.



Edited 1 time(s). Last edit at 12/30/2016 02:42PM by SlimJim.

Well, seems to be more than just Z wobble - notice how the Benchy doors aren't straight?

Your X and Ys both seem to be drifting too.

From what it looks like in the photos of the first post, that bed is swinging in the wind. It needs more than one guide rail or rod keep it from moving.

The single motor driving multiple screws is a very workable and much better approach than using multiple motors. If someone previously abandoned the idea due to "slipping", they made a big mistake. Keeping multiple motors synced is a problem. Keeping multiple screws synced with one motor is easy. The real reason it isn't done more is because adding a second motor is cheaper than adding pulleys and a closed loop belt, and cheapness is king in the 3D printing world (at least in the US).

My printer has had two screws connected by a single belt for years and has always worked perfectly. The only time I've ever had to adjust the X axis is when I first set up the machine and on rare occasions when I've taken it all apart to make major changes and then put it back together. I have actually taken the X axis off the machine completely and put it back without having to make adjustments (because the Z screws stay in sync because of the belt).

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
The single motor driving multiple screws is a very workable and much better approach than using multiple motors. If someone previously abandoned the idea due to "slipping", they made a big mistake. Keeping multiple motors synced is a problem. Keeping multiple screws synced with one motor is easy. The real reason it isn't done more is because adding a second motor is cheaper than adding pulleys and a closed loop belt, and cheapness is king in the 3D printing world (at least in the US).

My printer has had two screws connected by a single belt for years and has always worked perfectly. The only time I've ever had to adjust the X axis is when I first set up the machine and on rare occasions when I've taken it all apart to make major changes and then put it back together. I have actually taken the X axis off the machine completely and put it back without having to make adjustments (because the Z screws stay in sync because of the belt).

Quoted for agreement.

My latest printer design has a single motor for the two x rods and uses a timing belt to drive them. I made the same experience as the digital dentist, i was able to remove and reattach my complete Dual X assembly without loosing adjustment. It wasn't even difficult. It needs a bit more space, but it also decouples the motor from the rods and allows for a better design than using the radial motor bearings for axial load.

As for the problem, don't suspend the bed this way. This easily accounts for the bad print.

Edited 1 time(s). Last edit at 12/30/2016 03:52PM by Srek.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

THANKS for all your help guys!

I ran PID autotune, and tightened up the x-carriage rollers. And it's the best Benchy I have ever made.

I have ordered a new set of 5/8" linear bearings, and am planning a redesign of the Z-axis structure.

SJ

Now THAT is a nice Benchy. Well done.

If I may add a note about using Delrin wheels it may help someone in future.

I am using 'double V' delrin wheels on MakeSlides for my i3 printer.

The delrin wheels have a ridge in the center between the bearings, and the wheels kits are assembled with a washer / spacer between the bearings.

I noticed play on a few wheels and thought that the bearings were starting to fail.

Turned out that the spacer was 1.1mm thick, and the ridge in the wheel only 0.9mm wide, so the bearings were fixed solid 1.1mm apart, and the delrin wheels were able to float between the bearings.

Replaced the washer / spacer with a 0.9mm washer and eliminated the play. Print quality improved.

2c