Diagnosing Random Sudden Jerks

[dameronw]

I hope somebody on here can help me, cause I'm pulling my hair out here...

I have a CNC router, bought from somebody in LA, who since appears to have stopped business. I certainly would not say his CNC routers were the best quality, but it's been doing what I need just fine for the most part. For all intents and purposes, it should be considered homemade, and not professional quality. Lately though, it's been acting up.

About the machine - it's a 4x8, with a steel welded frame. Its X, Y and Z rails are as below:

Here is a link to the actual bearings: https://www.amazon.com/gp/product/B0...?ie=UTF8&psc=1

Its rack and pinion X and Y, with a horrendous acme Z axis setup (soon to be replaced).

I have no limit switches - only hard stops. I jog it against the hard stops to square it every time I turn it on.

Running on Mach3.

Right now, as it moves along any of its axis, it will make a sudden and violent jerk, and then keep on going. It gets slightly out of place each time it jerks, and makes horrendous cuts. It appears to do it on the Z axis as well, although it's harder to see, being on an acme screw, and with less movement. It does it on long straight cuts, and also in the middle of curves. When I jog it around with the arrow keys, I cannot make it jerk. Only when its running from gcode. Also, maybe related, maybe not, when I jog it around, sometimes it seems like there's a delay from when I let off the key, until it starts to decelerate. This makes it a real pain to zero it out in a very specific spot, and sometimes when I go to jog it against its stops to square it it runs into them for a few seconds instead of just gently tapping them.

It started doing all this out of nowhere a couple of weeks ago. A month or two back I ran a 70 sheet job through it, and never had it lose its place or act up once. I didn't change a thing, and it is now acting up.

So far, I've tried the following:
First, I read up on proper grounding and shielding cables. My cables all have the foil shielding, and I took all the dump wires, and grounded them to the ground of my power supply, which is the ground going back to the wall. The other end is ungrounded. I also took the router power out of the cable trains, and ran it along the dust collector hose, so that it cannot contribute any noise into the other wires. I ran a grounding wire down the dust collection tube, and grounded it on both ends to try to counter static. But, the machine jerks even with the dust collector off.

I pulled the motors away from their respective racks, and under no load they all appear to turn fine. I can run the gantry or carriage back and forth by hand, with no binding whatsoever. But, just to be sure, I replaced all the bearings. Still no change.

I checked the backlash between the rack and pinions - my setup doesn't have a spring tensioner - you just have to set them by hand and lock them in place. I tried having them tighter or looser, with no luck. I tried brushing out the racks with a brass brush, and lubricating the racks, pinions, and rails with a dry teflon lubricant - no change.

I replaced the parallel port cable - no change.

I replaced the keyboard with a brand new one - no change.

I've tried messing with the velocity and acceleration. Everything from painfully slow, to fast (for me fast is 500 ipm).

Argh - at this point I'm about to start replacing the motors, drivers, and breakout board. I don't know what else to do. I'd rather not spend all that cash, but I need to get the machine going again.

Any suggestions? I'll try to get a video if I can...

Will

Similar Threads:

• Uccnc load file jerks possition
• Need help diagnosing sudden g540 + ess problem
• Need Help!- Gantry jerks hard and skews out
• Problem- gibbs profile jerks the machine table
• Thc 300 Jerks Or Does Not Work?

[ger21]

Very hard to diagnose, but it sounds like something in the PC may be interfering with Mach3?
In the Mach3 folder, run Drivertest.exe, and watch the pulse rate, and see if it stays steady.

[dameronw]

Running drivertest.exe doesn't work for some reason... It just thinks for a milisecond, and then nothing happens. Tried to get it to run multiple times, with no luck. I don't know if it has always been like this or not - never tried to run that until yesterday.

[dameronw]

Here is a link to a video of the router running. At multiple points throughout this video it gives little jerks, although it is hard to see. But, at about 43 seconds, you can see a good one. Ugh. I have no idea if there is even audio on this video - my shop is too loud right now to tell. But it makes quite a noise when it jerks like this.

[ger21]

So it's apparently using a motion controller, and not a parallel port?
If using the parallel port, Drivertest.exe has to run.
If using a motion controller, which one?

I would say that it's either an issue with the PC, or a mechanical issue.
Open Windows Task Manager and monitor CPU usage when it stutters.

Is that a direct drive rack and pinion? If so, that's not a very good design, as it requires perfect alignment, and any dust or chips can cause binding, stalling, or maybe what you're seeing.

[Jim Dawson]

It looks like the jerk is happening in the same place each pass. I would take a close look for crud in the rack.

[dameronw]

Ger21 - it is direct drive, and agreed, it is a horrible design. In the past, I've had minor binding issues, which have always been solved by brushing out the rack and pinion, and lubricating. This time I've had no such luck, but I'm going to try again. Maybe it's something really stuck in there. It does 'seem' to happen in the same place pretty often,,, but then it won't happen at all in that spot, and will happen in some new spot. Also, I tried loosening the backlash in the rack and pinion. In the past, if there was dust built up in there, having a little backlash would let it not bind, but I would have play in the gantry - this was my diagnostic method. I would then clean the rack really well, and it would work again. Also, what are the odds that overnight I would develop binding in the X, Y, and Z axis due to dust in all their respective drive systems?

Also, it runs from the parallel port connection in the back of the computer tower, directly to the parallel port connection on the breakout board. Unless it has some sort of internal motion controller, I think it's running from the parallel port... I have no idea why running the drivertest.exe doesn't produce anything... I think maybe it's a quirk with my computer/mach3/windows/whatever setup. But, as far as I can tell, it's running from the parallel port.

I'm going to try cleaning the rack and pinions super well right now, and see what happens... But in the meantime, keep the ideas coming if you have any, it is very appreciated.

Man, as much as I have become addicted to using this thing, I wish I could just start over with a better designed machine, instead of polishing this one. That's not in the budget right now, so it is what it is...

[Jim Dawson]

Is it possible that you have Exact Stop set in Mach3, and the G1 moves are broken into segments?

[CNCMAN172]

I think you might be dealing with the gantry not staying square as it moves down the rails. As soon as it gets out of square just enough that causes a momentary bind. Then it moves again. You probably need to check on the stiffness of the gantry and make sure it is not flexing. Also check the motors on each side and ensure they are working in sync. That would be my guess based on the video.

Russ

[dameronw]

Jim Dawson - just double checked, and I do have it in CV mode. I read through the Mach help file on CV and various settings, and as far as I can tell, I have it set up correctly, and the settings are all the same as they have been for a long time. Also searched through some Gcode files, and don't appear to have any g61s, or g09s or anything odd like that in there. Just standard gcode from VcarvePro, with the 'exact corners' feature turned off in Vcarve.

CNCman172 - I think it may be getting out of square on one side or the other, but can't for the life of me figure out why that would be. The gantry is made from a 6" x 2" steel beam, with a 2" x 2" beam stitch welded onto the bottom of it. The beams are welded to 1/4" steel end plates, which go down to the bearings - 2 bearings per side, 4 total that my Y, or long, axis runs on. The beam has zero flex - the 1/4" end plates I don't think are the best design, but have been working so far, and with the motors disengage, I can't make the gantry bind if I try - it slides super smooth with finger pressure from either end, or anywhere in between.

I just wire brushed out the rails again, and set the backlash pretty loose, and it's still doing it. It does it in the middle of long straight runs, no weird speed changes, changes in direction, or anything fancy at all. Ugh - I'm wondering if one of the drivers, or the breakout board, isn't damaged somehow, and every now and then one of the motors is stopping or changing speeds for a second. Has anyone ever heard of something like that?

Oh yeah, Ger21 - I'm going to go see if I can find anything the PC is doing in the background right now. It's a simple PC, which is used for nothing but the CNC, but we did network it a couple of months ago so we can load files on it without having to use a flash drive. Worked great like that for a while, but maybe at some point it got infected with something??? Dunno, grasping at straws here.

[dameronw]

Winner winner chicken dinner!... I think...

I believe Ger21 nailed it. I looked at the processes running and CPU usage, and nothing jumped out at me, but I don't know exactly what I'm looking for. Nothing appeared to be spyware or trying to update itself or anything. But, next I disconnected the internet/network. I tried running through some programs, and didn't get a single jerk. Next I tried using Windows Media Player to play a video while running the router, and boom, jerks all over the place. Turned off the video, and it seems to work fine again.

I'm leaving this computer off of the network, and just using flash drives from now on. I also may look at getting a more powerful computer down the road. I looked through my settings, and my 'lookahead' value is only at 20. I'm afraid to set it higher, as that might bring problems. I've got it running through some Vcarving now, and (knock on wood) no problems so far...

Ger21, if I could, I would buy you a beer

[CNCMAN172]

Yeah, sounds like your PC is probably on the lower end of the specs required. I recently picked up a used i5 motherboard on ebay for $80 with processor and memory and put it in a case with new copy of windows and hard disk and it works great.

[dameronw]

Yeah I'll have to get something better. But, like I said before, I wish I could just scrounge up enough to get a better CNC router, and not polish this one. The more I learn about these machines, the more I want to get a much nicer one. One day...

[ger21]

I run Mach3 on a 1Ghz PIII, and have it on my home network without any issues.
One issue that used to come up in the old days with Mach3.
Check your network card settings, and make sure the speed is not set to Auto. Set it to a fixed speed, and it might help.