Stepper Chatter

Assembly questions and answers for the Shapeoko 3 CNC Mill
Post Reply
rickmellor
Posts: 7
Joined: Sun Sep 30, 2012 5:07 pm
Location: SF Bay Area

Stepper Chatter

Post by rickmellor » Sat Nov 12, 2016 7:08 am

Hello. I recently assembled my SO3 and have been having issues with the motors/controller from the start. I thought I'd resolved the issue by swapping out the power supply but after that initially seeming fine, it now looks like that was a red herring and I'm still having issues.

The first thing I noticed was that when I was doing a rapid jog to one of the preset points in Carbide Motion that the SO3 would seem to skip along the way. This seemed to happen only when moving the X & Y axis at the same time... moving a single axis didn't skip. I would also hear this skipping sound when I was running a program and since I was milling aluminum, I could tell it lost its position because on the next pass the endmill started making a lot more noise as it cut through the misalignment. Here's a video that shows the skipping (slight jerk sound) at about 14 seconds:

https://www.youtube.com/watch?v=gCqFr8dzRyE

So I tried to slow the machine down in case the grbl board was getting overloaded or something. I did a $130=150 and $131=50 but that didn't seem to make any difference in the speed of the machine. However, the machine now makes awful noise when it's holding position and the jerks in a single axis are much more pronounced. He's a video of the disaster:

https://www.youtube.com/watch?v=BOzDIyG1LMk

I hooked the power input to a scope and I can see that the voltage supply is super smooth. And since before I made the grbl speed changes the machine was working fine in single axis jogs but not dual axis, I assume the problem is somewhere in my grbl config, but I'm too noob to pin it down.

I'm running Carbide Motion 366 and the grbl version seems to be 0.9. I also tried using the universal g-code sender but the machine showed the same issues so I don't think it has anything to do with Carbide Motion.

Mechanically, I checked and all the belts are nice and tight. I even loosened all my v-wheels so there's no tension in there anymore... didn't change anything.

Your suggestions are greatly appreciated.

rickmellor
Posts: 7
Joined: Sun Sep 30, 2012 5:07 pm
Location: SF Bay Area

Re: Stepper Chatter

Post by rickmellor » Sat Nov 12, 2016 7:13 am

Looking at the docs it seems I used the wrong grbl commands to set the max speed. I should have used $110 and $111 instead.

These values started out at 5000mm/sec so I tried moving them down to 2500 and 1000mm/sec and at both speeds I still had the skips. So, it looks like the problem isn't the grbl board keeping up. Hmmm

WillAdams
Posts: 8619
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Stepper Chatter

Post by WillAdams » Sat Nov 12, 2016 12:26 pm

I believe that you need to:

- check and increase your belt tension --- http://www.shapeoko.com/wiki/index.php/Shapeoko_3#Belts and http://docs.carbide3d.com/article/42-ho ... lt-stretch --- in particular, you need to get the tension for the Y-axes belts even --- I've had symptoms similar to what you describe when one was looser than the other
- check and verify your pulley set screw placement --- http://docs.carbide3d.com/article/169-s ... set-screws

There is a bit of jitter sometimes on some machines when jogging in Carbide Motion --- anyone who experiences that needs to pass it on to support@carbide3d.com along w/ their machine specifics so that they can (if possible) add it to the testing pool. On the bright side, Grbl 1.1 will be released soon, and there should be a new Carbide Motion to go with it, which ought to be a marked improvement.
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets

rickmellor
Posts: 7
Joined: Sun Sep 30, 2012 5:07 pm
Location: SF Bay Area

Re: Stepper Chatter

Post by rickmellor » Tue Nov 15, 2016 7:27 am

Thanks Will. I carefully adjusted the belt tension using the 1/4" instruction in the assembly guide and checking the relative tensions now feels like they're both about the same. I've ordered a fish scale and when that arrives I'll measure the force required to lift the belt a measured distance at the midpoint of the span. If it matches side to side then I'll be confident that the tensions are the same. This assumes that each belt has the same tensile strength... but I'm sure they're close enough to not matter.

I also moved the gantry to the midpoint of the Y axis, marked the Y axis pulleys and the belt and then ran the machine back and forth many times and let the skip happen several times. I then moved back to my starting point and checked my marks. The belt and pulley marks matched so I don't think we're skipping teeth on the pulleys. It's possible that it's popping off the side of a tooth and settling back into the pulley, but I'm not seeing any black belt dust so I think there's good engagement between the belt and pulley. If it was misaligned I'd expect the belt to get chewed on a little bit by now.

I also checked the set screws on the pulleys and they're turned down onto the flats so I don't think there's any slippage there.

I also leveled and re-squared the gantry. I needed to add a bit of shim between the left Y axis carriage and the left side of the Y axis extrusion, in the rear, to get the two carriages to rest flush against the front frame plate.

I also talked to a EE at work about it and he suggested that I replace the power supply wires with a heavier gauge, in addition to the PSU swap I've already done. So I rewired the mini DIN connector using 14 AWG stranded and kept the length to about a foot to reduce any inductance buildup in the wire. I've also got a 24v 25a PSU driving it so there's no question that there's plenty of power getting to the controller. I think I can rule out electrical. I'm tempted to hook the steppers up to the oscilloscope to see if there's any anomaly in the stepper signal when the skip's occurring, but I'll save that for later when all else has failed.

So with all that done I'm still having the skips... what the heck!?

So in desperation I just started staring at the thing and then I noticed something: an uneven wear pattern is developing on the inside of the v-wheels. I looked more closely and I was able to see that the v-wheels aren't riding true on the extrusion rails. Could this be it? Could stress be building up in the v-wheels as the gantry moves back and forth and it periodically releases? Or could the misalignment be causing the wheels to ride higher and higher up the extrusion rail until finally they slip and fall back down? I'm not sure but there's definitely an issue here.

Here's the worst one... the left rear Y axis upper v-wheel. You can see a definite misalignment and the uneven wear on the v-wheel:
Left Rear.JPG
Left Rear.JPG (86.14 KiB) Viewed 1064 times
You can see it on the right rear too, but not as bad:
Right Rear.JPG
Right Rear.JPG (89.89 KiB) Viewed 1064 times
Here's the right front. It's barely noticeable here and is probably why I didn't catch this until the uneven wear pattern emerged. It's similarly not as bad on the left front.
Right Front.JPG
Right Front.JPG (89.06 KiB) Viewed 1064 times
So I loosened the Y axis extrusions and tried to adjust them but there's no play in the end plate holes so I wasn't able to correct it. I whacked them a few times with my rubber hammer and they wouldn't budge so I think the screws are flush with the hole walls. I'm thinking I need to remove the rails and use a circular file or reamer to open the holes up so I have some play for adjustment. Since both Y axis extrusions need to move outward, I'm thinking that for some reason my rear end-plate holes might be slightly out of tolerance.

That's the best I've got to go on right now. Any other suggestions would be appreciated as taking this apart is going to be a chore. If I do it I think I'm going to be entirely neurotic about it and make sure the thing is perfectly square as I reassemble. It'll take some time... something I don't have a lot of with an infant in the house. :roll:

WillAdams
Posts: 8619
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Stepper Chatter

Post by WillAdams » Tue Nov 15, 2016 10:34 pm

If you're going to go down this rabbit hole, I'd suggest getting enough precision shims to use everywhere that they are dimensionally significant. Also see twforeman's spreadsheet for calculating those dimensions.

The heavier wiring is an interesting point --- mostly on the SO1/2 we used 4-conductor security wire which is heavier.

Do the skips occur w/ all comm / control apps or only one?
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets

Post Reply