[Phattie]

Hi guys,

I recently swapped out an axis on my machine to a new axis, leaving everything else the same. Only thing that changed was the frame and lead screw, same motor, same box, same wiring, same PC, same everything. The machine had 2000+ hours of runtime on it, and made maybe 5 mistakes that entire time. Currently I cannot run it for more than 20 minutes without the new axis stalling. I've been troubleshooting for 3 days, here's a little background:

Full size PC, not a laptop, running to a Campbell Designs breakout board, wired to Gecko 202 drivers, then steppers. Power supply is plenty big, no wiring was changed from the original setup, only the mechanical axis.

The only axis that is missing steps is the X axis. On this particular machine, the X axis moves vertically, photo attached. This axis has a 640 oz. stepper on it. The piece that was swapped out was a TKH rail, I replaced it with a K2 unit as pictured. After tuning it and getting the new axis running just right, I was able to jog it at 100, with an acceleration anywhere from 5 to 20. When it came time to actually start carving with the machine, I tuned it down to a feed rate of 60, acceleration of 10. After 20 minutes or so it jumps roughly 1" up or down, then continues carving in the wrong place.

I reduced the feed rate several times, and anywhere from a half hour to an hour into a run the X locks and jumps an inch or so, then keeps going. I finally tried with with acceleration set at 1, feed rate of 20, and it did the same thing after about an hour.

It seems to stall and jump an inch or so no matter the speed or acceleration settings. Neither of the other axis ever miss a step. The code hasn't changed, running the same files I've been running for years.

With everything turned off, the lead screw moves easily by hand, there is no binding through the whole range of motion. Everything is smooth as silk.

I don't believe it could be a lead-screw resonance problem, because even at the slowest speeds possible it still jumped. Currently I think the motor or gecko driver could have a physical problem, but would appreciate some input...I'm fairly stumped and getting a little more behind each day.

[jsheerin]

So the the motor, driver and all electronics are the same as before, and you just changed the frame / rails / leadscrew? If so, I would look for something mechanical first.

Does the axis stop, sort of load up (like maybe the screw deflects a bit from straight) and then it suddenly moves in the direction it was getting pushed in? If so I'd suspect binding due to some cause - maybe the new bearings are not as resistant to debris? But then maybe whatever's binding gets cleared out when it jumps. I'd also check your couplings on the screw and make sure no set screws are slipping. However if it's a random direction of movement after binding, that might suggest a different problem.

Resonance is tricky, but the G202's have midband resonance control so that should not be an issue...

[robe_uk]

my first thoughts were a slipping coupling as you said you have rebuilt this axis so double check that. If not that you can rule out a driver problem by swapping the driver unit with a driver unit from another axis, if the problem stays on your x axis then it not the driver at fault, if the fault follows the driver to the other axis then it would look like a driver problem.
Also double check the wiring and make sure you have no bad connections or dry solder joints in connector plugs, am sure you unplugged things during the rebuild?

Good luck

[epineh]

It does sound like missed steps of some sort, even though you may not have changed the wiring, check any ground connections, especially power supply, it is surprising what a loose connection can do as far as electrical noise is concerned.

Apart from that maybe the axis is heating up after a while and binding, something that wouldn't be picked up from a cold start test, just suggesting here, sounds like you have tried the obvious stuff.

Russell.

[Phattie]

Thanks for the ideas. I'm going to go through the whole setup and check all the connections tomorrow and fire it back up in foam. If I get the random skipping again I'll swap the driver on the X and Z, and then fire it back up again. If the Z starts skipping, that should nail it down to a driver problem.

Thanks again for the ideas-

[bigz1]

Hi guys,

I recently swapped out an axis on my machine to a new axis, leaving everything else the same. Only thing that changed was the frame and lead screw, same motor, same box, same wiring, same PC, same everything. The machine had 2000+ hours of runtime on it, and made maybe 5 mistakes that entire time. Currently I cannot run it for more than 20 minutes without the new axis stalling. I've been troubleshooting for 3 days, here's a little background:

Full size PC, not a laptop, running to a Campbell Designs breakout board, wired to Gecko 202 drivers, then steppers. Power supply is plenty big, no wiring was changed from the original setup, only the mechanical axis.

The only axis that is missing steps is the X axis. On this particular machine, the X axis moves vertically, photo attached. This axis has a 640 oz. stepper on it. The piece that was swapped out was a TKH rail, I replaced it with a K2 unit as pictured. After tuning it and getting the new axis running just right, I was able to jog it at 100, with an acceleration anywhere from 5 to 20. When it came time to actually start carving with the machine, I tuned it down to a feed rate of 60, acceleration of 10. After 20 minutes or so it jumps roughly 1" up or down, then continues carving in the wrong place.

I reduced the feed rate several times, and anywhere from a half hour to an hour into a run the X locks and jumps an inch or so, then keeps going. I finally tried with with acceleration set at 1, feed rate of 20, and it did the same thing after about an hour.

It seems to stall and jump an inch or so no matter the speed or acceleration settings. Neither of the other axis ever miss a step. The code hasn't changed, running the same files I've been running for years.

With everything turned off, the lead screw moves easily by hand, there is no binding through the whole range of motion. Everything is smooth as silk.

I don't believe it could be a lead-screw resonance problem, because even at the slowest speeds possible it still jumped. Currently I think the motor or gecko driver could have a physical problem, but would appreciate some input...I'm fairly stumped and getting a little more behind each day.

I had a problem on my Z axis for a while. After running perfectly for months the stepper would start to miss steps and then just die. I replaced the stepper motor twice. I eventually found the fault with the wiring cracking through continual bending. All three stepper motors now work again.

[Phattie]

Thanks for all the advice guys. I checked all the wiring and fired up the machine this morning, made it four hours before jumping about an inch. I reset it, slowed down the feed rate to 30 and the acceleration to 2, and let it run. Made it about two hours before the axis completely locked up and it just sat there making a horrible buzzing noise. Took off the motor, and the axis moves freely. Tried to turn the motor by hand, and it won't turn. Sucker is stuck, I guess it burned up.

So, this whole speed/axis/driver/etc. problem looks like it stems from the motor going out. The other two motors spin freely with the power off, this one can't be budged, it just wiggles back and forth a very small amount. It is about 6 years old and was part of a shopbot machine before this one, guess I had this coming.

One last question before ordering the new stepper: I had a 640oz motor on there that was doing a great job and never got very warm. I'm planning on just ordering the same motor, but there is a lot of talk about getting a motor just a hair over what you need, so it's running at near it's max output. Any real benefit to this, or should I stick with the semi-overkill motor?

[jsheerin]

Who talks about getting a hair over what you need? Never heard that... I say overkill it.

A very valid point is that some larger motors have crappy torque curves and smaller motors can actually perform better than them, so there is that to watch out for. And of course you actually need a driver and power supply that can let you take advantage of the big motor. But if you go with a marginal motor, as soon as you get a slightly higher load you'll lose steps. I'd rather have power in reserve as long as the rest of the machine is up to it (ie, you're not going to be stripping drive nuts or something because you put a 2kW motor on a 1/4" all thread screw).

[Phattie]

Update, although slightly frustrating:

Replaced the X-axis motor with a brand new one, 940oz with a little more oomph. Went through the entire machine and checked every connection, double-checked the wiring, etc. I actually ditched several of my connecting strips and just soldered things together. Swapped the Geckos out on the x and z, and fired it up.

Still getting the random stalling/binding about once every 4 hours that ruins the piece.

I think it has to be the K2 axis. Even though it moves smoothly on it's own, I don't believe it's up to this task. Everything else has been replaced/rewired/double checked/etc. The axis is the only thing that's changed.

Ordered a TKH off ebay this week, will be swapping that in next week and we'll see what happens!

[Markwind]

Update, although slightly frustrating:

Replaced the X-axis motor with a brand new one, 940oz with a little more oomph. Went through the entire machine and checked every connection, double-checked the wiring, etc. I actually ditched several of my connecting strips and just soldered things together. Swapped the Geckos out on the x and z, and fired it up.

Still getting the random stalling/binding about once every 4 hours that ruins the piece.

I think it has to be the K2 axis. Even though it moves smoothly on it's own, I don't believe it's up to this task. Everything else has been replaced/rewired/double checked/etc. The axis is the only thing that's changed.

Ordered a TKH off ebay this week, will be swapping that in next week and we'll see what happens!

Have you swaped out the parallel port, ?

[Phattie]

Nope...I'm assuming the one on the control box would be the place to start, not the one in the actual computer, right?

Never crossed my mind to swap that out, but given that everything else has been changed, it could be the weak link. Thanks for the suggestion!

[Markwind]

I was thinking of the parallel port in the computer, if it's an integrated port you could add a parallel port card, or try a different PC,

[Phattie]

Gotchya, that sounds easier than swapping out the whole axis. I'm going to give that a go first...will report back. Thanks for the suggestion!

[jsheerin]

Agree with Mark - the parallel port could be a problem. I had that on mine - the port was not supplying enough current to drive the inputs of a G540 and I was getting random, weird behaviour from my z axis. However I'd think you'd be okay with the break out board, plus that wasn't something that changed... But it's still worth trying. Note that some breakout ports still don't have enough current capability on all pins - just more pins with good current capability than an onboard port (at least in my case). The real way to test for this is to get an oscilloscope and look at the step and direction pulses going to your drives. That's how I found my problem.

[Phattie]

The real way to test for this is to get an oscilloscope and look at the step and direction pulses going to your drives. That's how I found my problem.

Any chance there is a thread around here on how to do that? I'd be interested in giving that a go! Could be the crappy old desktop running this thing is another root issue....good times.

[jsheerin]

Go here: http://www.cnczone.com/forums/gecko_...g_problem.html
Start at post 15. I posted some pics of the scope traces I was seeing from my card. Basically, you want to look at the voltage from the step or direction pin to ground. For the step pin, you should see something like a square wave that goes from 0 volts to 5 volts. When you draw current from the pin (like your breakout board does), then that voltage will probably drop and the step will get rounded off like in some of my pics. The board or drives will have a spec for how low the voltage can drop and it will still see it as a pulse. On mine, the voltage dropped down to 1V, and that was not a high enough level. But then it did this weird thing where when I reversed directions, the voltage would float up to around 3V but still just have a 1V step level. I was trying to get 5mA out of my pins for the G540, so I put a resistor across the pins on my parallel port that would draw more than 5mA at 5V. Then when I measured the voltage, I could find the current being supplied by using ohms law: I=V/R.

I did it this way because in my setup it wasn't easy for me to get my scope probe on the pin on the drive itself. If you can, that would be a better way to measure it - see what's actually going on with the drive itself.

[Phattie]

Progress, except not really...Finally got an aftermarket parallel port installed. Opened Mach3, set it to the new port, and nothing works. Says my limit switches are being triggered and won't reset itself, not sure where to go with that...no idea.

Then I installed mach3 on the laptop figuring I'd give the machine a run off of that just to see how it works out. Whoops...no port on the laptop, only USB.

Soooo, without sitting down and figuring out the new pin pattern on the aftermarket card, I think it's time to grab a barebones system, do a fresh install of everything with a good port and a good power supply and go from there. Good times.

[Phife]

random limit switch triggers are usually caused by noise.

For testing purposes Id just disable the switches so you can get mach out of estop mode. If you have noise issues try to clean up and shield the limit switch wiring and if that doesnt help add 500-1000 to the debounce setting in mach3.

[Phattie]

Oops, I should clarify that this machine has never had limit switches installed or wired to the breakout board. That's what's throwing me off about the new card.

[davy182]

I would run it till the issue happens again and then loosen your lead screw and retighten it. Maybe when you changed things over there may not be any slack for lead screw to expand slightly. So after 2-4 hours of running and the screw is warm it may be causing some binding issues.