My "2 Wifi" is un-baby-stepping

dc42

@gnydick said in My "2 Wifi" is un-baby-stepping:

@dc42 well I tried the latest beta, it's still happening. Is there any debug mode I can do for you? Maybe hooked up to the serial terminal?

That's odd, I took your config.g file, made minimal changes to allow it to work on my bench system, and i was unable to reproduce the problem.

Does the un-babystepping fail only in one direction; for example, if your machine has a lifting bed, does baby stepping to low the bed always succeed and lifting it sometimes fail?

gnydick

@dc42 IIRC, it fails in both directions. Sometimes it's a slow winding back to where it was, other times it's the complete sum of baby steps e.g. if I click it 5 times and it moves .05 x 5times, then it will undo it .25 all at once.

Phaedrux

Are you by chance baby stepping below your M208 Z axis minima?

gnydick

Well, yeah, that's kinda assumed to be one of the use cases for baby stepping. But it's not always the case.

wilriker

@gnydick AFAIR baby stepping honors axis minima (and maxima). But it should never even try and later undo it so that should not be the source of this issue.

Phaedrux

Reason I ask is because I just had baby stepping kind of freak out on me, but not exactly how you were describing. I was trying to lower it quite a bit, (-0.3ish) and it started to throw G1/G2/G3 errors something about moving outside machine limits and the XY movement would kind of reverse. It was doing the skirt at the time. When it started moving unexpectedly I killed the power and lost the error messages.

gnydick

@phaedrux I've seen those too. Sometimes the board actually locks up or reboots the web host portion.

gnydick

@dc42 OK, Now I saw it happen with a regular move, no print was happening, I was just calibrating Z.

I homed it, moved my head to the center of the bed, brought it to zero. Then I +.5 to see if the head dropped at all (indicating that the bed was pushing up on the head) and it wasn't pushing on the bed, so I -.5 and the bed came back up to meet the nozzle, and then it, all on it's own, the bed moved back down +.5.

I CAN'T be the only one experiencing this on multiple revisions of firmware. And I'm pretty sure I saw it on my Maestro too.

This has to be looked at. I'm not very happy that you just kinda ignore it because you think I'm the only one experiencing it. To be totally honest, I'm not your average person. I tinker with things, explore all of the ins and outs, as well as am more observant than 99.999999999% of people. It's a curse. In my career, I'm constantly pointing out things that are broken that nobody has seen, sometimes for decades.

You can take it from me, if I say something is broken, it's broken.

Now, I'll ask again, is there any debugging help I can provide?

gnydick

I have a possibly interesting update....

One of my Z motors was apparently dying. It finally died today. Is it possible that a faulty motor in a 2 motor series could cause this behavior?

Phaedrux

@phaedrux said in My "2 Wifi" is un-baby-stepping:

Were you able to figure out of your motors were cooked? Did you try any of the other suggestions above?

I guess that answers that?

dc42

@gnydick said in My "2 Wifi" is un-baby-stepping:

I have a possibly interesting update....

One of my Z motors was apparently dying. It finally died today. Is it possible that a faulty motor in a 2 motor series could cause this behavior?

It's unusual for stepper motors to die. In what way did it die: an open circuit coil, or something else?

If a coil of one Z stepper goes open circuit, then power to the series-connected coil of the other stepper motor will be cut too. When this happens, it is quite possible that commands to move the bed up will move it down instead. Whether it does that or just vibrates will depend on the weight of the bed and other factors.

When you have replaced the faulty stepper motor, let us know whether you are still having this un-babystepping problem.

gnydick

@dc42 oh well, I replaced the motor and it's still happening.
I'm not sure how it died. I think it's bearings though. It was really hard to turn the motor by hand while it was disconnected, and made a soft grinding sound.

gnydick

@phaedrux I've tried everything.

dc42

@gnydick said in My "2 Wifi" is un-baby-stepping:

@dc42 oh well, I replaced the motor and it's still happening.
I'm not sure how it died. I think it's bearings though. It was really hard to turn the motor by hand while it was disconnected, and made a soft grinding sound.

That can be caused by swarf in the motor. If so, it can usually be cleared by dismantling the motor, blowing or brushing out the swarf, and re-assembling it.

gnydick

@dc42 yeah, I disassembled it and cleaned everything as much as possible and it won't even budge now.

dc42

@gnydick said in My "2 Wifi" is un-baby-stepping:

@dc42 yeah, I disassembled it and cleaned everything as much as possible and it won't even budge now.

I've found that rotating the back half of the motor 90 degrees with respect to the front half before screwing them together sometimes helps.

gnydick

@dc42 thanks for the advice.

gnydick

I'm getting more than furious. I've used 2 different Maestro's, and now a 2-Wifi and it still un-baby-steps. It happened with 2 z motors, 3 motors and 4 motors. This cannot be a fluke.

I'm also getting ridiculous z-positions.

With the machine perfectly leveled, z-homed to perfectly 0, I start a print, and the first layer immediately starts at at-least 1mm off the bed. I will then re-home the machine and the Z height is perfect. Start another print, and it starts over 1mm off the bed again. This is with no baby-stepping at the beginning.

The other bizarre part is, when I pause the print after I see that it's 1+mm away, I manually G1 Z0 and it goes right to the bed. I G1 Z5.2 and it goes back to the previous position. I G1 Z0.2 and it goes to a C-hair away from the bed. I resume the print and the bed moves down and up, and the position reads 0.2mm, but the bed is still over 1mm away.

These behaviors seem to be intermittent. I'll have a streak of awesome prints, then it'll just go bat-shit crazy. I do leave my printer on for very long periods of time.

My prints, when successful, have the correct layer height, and when I measure 100mm of Z movement with calipers, it's correct.

There is something wrong.

Also, when I do baby-step, the Z position on the screen reads positive, when I baby-step negative in a 1-1 ratio. I baby-step -0.8mm and the Z position reads 0.8mm.

WTF is going on????

I've uploaded video, I can reproduce it anytime I want, why is this being ignored like it's some kind of fluke?

dc42

Have you tried reducing Z max speed (M203) or acceleration (M201)? It could be that the initial move to Z=0.2 (or whatever your layer height is) is skipping steps.
Have you checked that your slicer isn't inserting M201 or M203 commands into the GCode? Slicers should leave these settings alone because they are used to set machine limits, but unfortunately at least one slicer tries to change them.
Try running the Z stepper(s) in spreadcycle mode. Stealthchop mode is much more likely to result in skipped steps.

gnydick

@dc42 these are not skipped steps, that would result in the bed not being far enough away. It would also not explain any of the symptoms.

Speed and acceleration are not high at all.