[VaderSpade]

Are you using an external step up transformer to go from 120VAC to 240VAC? No

[VaderSpade]

Latest report;

I spent the entire day yesterday running to town for parts and re-wiring the mill into the house electricity. This is a separate drop with its own separate transformer. I bought a brand new surge protector with a “Good Ground” indicator light, and I eliminated the GFCI for the test.

No luck! The mill let out a series of thumps right at the end of a run and failed to return to zero but read zero on the X & Y and 0.200 on the Z which was the proper step over.

Tormach and a few of the good people here have sent me VERY long lists of things to try, at this rate it could be weeks if not longer to track down this problem, and my lapidary business is suffering. I am returning $4,800.00 to eight customers today because I can’t finish their projects. Most of that money has already been spent on parts.

I did get a better video of the problem this time and I will try to post it. The last time I posted a video it was from a DVD, and I’m not sure I have all the cables and software it takes to get the video from my camera to YouTube but I’ll do my best, and hope I don’t lose my entire Sunday on this project.

Sorry for the negative tone but this is not what I want to be doing with my time. I do very much appreciate all the time the good people here have spent trying to help me out.

[VaderSpade]

So here’s a video. I have had several different problems but this one has repeated itself 5 times in a row, although it only happened 2 out of 11 times before this latest streak.

I set up the zero ½ inch off of the project in the bottom left hand corner so I would not break yet another bit as the mill fails.

You can see and hear right at the end as things get tight a series of small thumps. When you hear a thump you’ll see the mill stalls or hesitates, and then stops far short of zero while reading like everything is just fine.

Tormach 1100 Fail - YouTube

[BobWarfield]

I see why your customers are interested in the work--looks very cool. Making cameos from stone?

Not much to add from me for the video--my ear is tuned to servos not steppers. Hopefully someone will comment. That there is a problem is certainly obvious. Make sure the Tormach people have the video, though if they're not following this thread they need to.

It is a 3D job. I have always found those to be more challenging for Mach3 than 2 1/2D, though I recently had a discussion with one of Mach insiders who assures me it isn't.

I wish you had a friend nearby with a Smoothstepper. I don't honestly think that's the problem, but it would be so quick to install I'd try it if it was me just because I believe in them. Just to reiterate, I don't think this is your problem, but it would eliminate a lot of worries about Mach's 3D performance. Still, you're on a tested machine configuration recommended by Tormach and lots of people do 3D with Mach3, so again, not likely to be the problem.

Best,

BW

[shred]

Those things in the video that sound like somebody tapping on the bottom of a bucket? Are they about that loud in real life?

The rest of it sounds ok to me.

[VaderSpade]

Shred,
Yes those are the thumps; they are louder when you’re standing next to the mill. You can also see the hesitation or stalling although it is subtle.

[shred]

I'm by no means an expert in this area, but I'm guessing the thump is the stepper being asked to do something unexpected, so I'm thinking that would point to something in the electrical or computer system (not a big narrowing there )

Some things that may help narrow it down further--

Have you re-zeroed after the thumps and determined how much distance is lost in X, Y and Z? Do they always all lose distance, or sometimes is it just one axis? If the distance-lost is different between thumps, are they multiples of some other number?

Another thought-- have you tried just moving one axis over and over again until it thumps? Do all the axes show the problem?

Does it always take a while to happen (like, say a screensaver is trying to kick in, or some wait-for-the-user-to-go-away task is running?)

[LeeWay]

Oh you are loosing steps alright, but not news at this point. I have run Mach 3 with a laptop for more than 8 hours on one program in 3D with no trouble at all. Then change tools and run several more hours. Mach 3 has no trouble with 3D.
I don't think it is the machine itself. Too intermittent for that. I suspect it is in the computer to the drives. Sounds like something slowing down the signals rather than something wrong with the machine.
You say it is happening towards the ends of the gcode?
If that is the case, it may be hot drives. Low power output to the drives etc. Just something else you may not have time to test.
Thats about all I can come up with after hearing the machine though. I hope you get er going soon.
I know how frustrating this can be when they go down and you are loosing money because of it. Been there a few times myself.

[zephyr9900]

Thank you for posting the video, Randy. The thumps seem to occur when the move is short enough that it is still accelerating when it starts decelerating again (i.e. it doesn't get up to constant speed). And at 0:46 to 0:48 there are definitely lost steps as the axis (Y?) just hums instead of moving on the last 3 cycles before the spindle shuts off. I would suspect the backlash setting--I remember reading that Mach can have problems when backlash compensation is enabled, because the software tries to move the axis instantaneously to take up backlash. Tormach by default has backlash compensation turned off--maybe the latest XML has it accidentally turned on?

Randy

[VaderSpade]

This has continued with a computer change, I can’t see anything being accidentally turned on twice.
Although short fast moves seem to bring the problem out I have had the mill gain steps going beyond where it was supposed to make the next pass and plunge into the part. In that case on the X axis. I’ll post a picture and description of that part ASAP.

[Scott_M]

I have had the mill gain steps going beyond where it was supposed to make the next pass and plunge into the part.

Just being technical here. You may have gained in X or Z but it was due to lost steps. It depends on which way the axis was going when the steps were lost on how it will affect your true position + or -.

Have you replaced the Printer cable (DB-25 parallel port cable ) between the machine and computer ?

Scott

[zephyr9900]

It would be a setting in the XML file (the configuration file that tells Mach3 the characteristics of the machine it's running) in the PCNC software installation, so wouldn't depend on the computer it's loaded into.

A stepper motor won't gain steps unless there is electrical noise on the step line, but then it would be random and constant. What is more likely is that the immediately previous move lost steps. Edit: +1, Scott. We seem to have been typing at the same time...

Randy

[VaderSpade]

In this case I was running a simple pocketing operation. The upper part in the first picture is what was supposed to be cut, but on the lower part while simply returning for the ninth of ten passes the X went 1.7 inches beyond where it was supposed to plunging to the depth of the ninth pass acting as if nothing was wrong.

I ran it again several more times without a problem. This program is supposed to be run on aluminum parts that cost between $100.00 and $220.00. I have ruined two of these parts, and have broken a couple of $60.00 bits before I started cutting air and plastic. No one wants to pay for the air or plastic.

[Al_The_Man]

I know I am coming late to the party, but having skipped through the previous posts I sympathize and what I say may be cold comfort, but for future reference, my recommendation is that if one is intending to use a machine for even semi-industrial environment, IOW to make money especially make a living off it, the suggestion I have is to go with a machine fitted with servo's.
I am not saying that you would be absolutely trouble free, but IMO they are easier to trouble shoot if you do have these kind of problems.
Personally the problems you have been experiencing I have found to be often caused by noise issues due to improper grounding practices.
Just my 2¢.
Al.

[zephyr9900]

...the X went 1.7 inches beyond where it was supposed to...

Again, the true situation is that on a previous move, the X stopped 1.7 inches short of where it was supposed to go.

When steppers lose steps, it's not just a single step. Stepper motors are driven by a sequence of pulses, and if the motor loses synchronicity with the pulses, it stalls and just hums in place for the remainder of the steps in that particular sequence. When that axis makes the next move, it is accelerating from 0 speed to running speed so the stepper can keep up with that next move.

I was in the middle of a run during my last couple of posts, but have now confirmed that, at least through Series II, all axes' backlash is set to 0 and backlash compensation is turned off. I can't tell you how to check this on your machine without all sorts of warranty voiding, unfortunately...

Randy

[VaderSpade]

Quote;
Again, the true situation is that on a previous move, the X stopped 1.7 inches short of where it was supposed to go. end quote

I understand what you're saying, but you can see that the eighth pass was completed, so I can’t see where it could have “lost” any steps. But I’m new to all this CNC stuff.

[LeeWay]

What happens when steps are missed is the machine stalls. The computer keeps going. Then on the next pulse stream the computer picks up where it thinks the machine should be and continues the job. The machine is not where it should be and may or may not complete the job error free from that point. It's just the last part it does is obviously in the wrong place because of missed steps.
The direction of travel at the time of the miss and the direction it resumes in tell how far it actually missed and what axis it missed on.

[RTP_Burnsville]

I am sure you have already tried to recompile the G-code using slower feed rates for everything.

Could it be that one of the control signals is inverted (ie the clock to the steppers) in MACH3? That would be an easy check except you are likely running the locked version for warranty. Without an unlocked version or a scope it could be hard to trace such an issue.


The suggestion to swap cables is a good one too.

Robert

[shred]

Some things that may help narrow it down further (copied over from the last page). If we can't get the issue narrowed down, then debugging is going to be throwing jelly at the wall and seeing if any sticks.

Have you re-zeroed after the thumps and determined how much distance is lost in X, Y and Z? Do they always all lose distance, or sometimes is it just one axis?

Another thought-- have you tried just moving one axis over and over again until it thumps? A simple bit of G-code would do that. Do all the axes show the problem the same amount?

Does it always take a while to happen (like, say a screensaver is trying to kick in, or something is overheating or some wait-for-the-user-to-go-away task is running?)

[shred]

Assuming you're cutting a FCG pocket and using a 1/4"-ish mill, it looks like what happened is the mill was in the vicinity of the moving jaw of your caliper and the computer said 'Move X back to the right side of the pocket', and the machine didn't go anywhere at all, as the distance you are wrong is pretty close to the total distance desired. Were you watching at the time?

If the G-codes are being completely ignored, it might point to something different than if they are just not moving the right amount. Do you have any way to verify that?