Welcome to LinuxCNC.org

I have encountered a problem that defies logic in my mind.

I have been running EMC 2.4 on Lucid 10.4 for about a year now. Recently I noticed that the Z axis was not consistently lowering to the proper depth. It was noticeable when I would pocket something, then do a finishing pass in the pocket, you could see on the floor of the pocket that the finish pass was not exactly flush with the rest of the pocket. My Z axis is the least mechanically accurate axis, with a stepper motor driving pulleys via toothed belt, driving a 90 degree spur gear, driving a gear rack with teeth cut into the quill, so I just passed the poor depth control off as a mechanical problem. Now, maybe this problem has been there since day one of running Linux/EMC, but it was not noticeable.

A few weeks ago I was machining some parts, and I noticed that the problem was becoming more obvious, to the tune of maybe .030". Then last week I was doing some work and the bear really bit me. On multiple depth cuts of .100" the quill had actually lowered .54" when it was supposed to be lowered .40". I noticed it as the mill cutter was taking a pass thru the jaws of my vise.

When I started troubleshooting, I noticed that the problem was occurring only when the DC mill motor was running. At full speed I could feel and see the Z stepper turning. I have also seen and felt some movement in the X axis, but so far have not noticed it in the Y axis.

I have the control board for the DC motor mounted in an old computer case, along with the drivers for the stepper motors. Thinking that I was getting EMF interference from the DC drive, I moved the DC motor driver board out of the case and into a separate steel case, no difference.

I tried re-routing the cables for the Z stepper - no difference.

Thinking it might be a feedback problem, I tried different combinations of plugging the DC motor, stepper power supply, and computer into different house circuits. No- success.

It does it regardless of whether EMC is actually running or not. As long as the computer and Ubuntu is running, the problem is there.

I grabbed my old 733 P3 loaded with Win98 and Turbocnc that the mill was using before and hooked that up. SUCCESS! No problems at all. So the problem has to be the newer computer, right?

So, I stuck another hard drive in the old 733 and installed Ubunto 8.4 and EMC 2.3. ( I wasn't sure that such an old computer would work well with Lucid/EMC2.4) Problem returned!

This is where the problem defies logic in my mind.

1. The fact that the problem has worsened over time would point to a hardware (computer, driver, breakout board) problem.
2. The fact that the problem went away with the 733 computer and Turbocnc leads me to think the original computer, or Linux/EMC is the problem, and that the rest of the hardware is fine.
3. The fact that problem comes back with the 733 computer and a different version of Linux/EMC points me back to either Linux/EMC, or another hardware component, but the hardware is fine with Win 98/TCNC.
4. I tried switching the leads for the X and Z axis at the driver boards, but the problem was still worst on the Z-axis.

Makes my head hurt. I might try yet another computer that I will load with Mach3.

Any thoughts would be appreciated.

Rick Denoon

Thats pretty weird
I wonder if its related to the EPP VS SPP mode parallel port output drive difference

It does sound like marginal levels perhaps at the parallel port/breakout interface

Debugging things like this a lot easler if you have a 'scope. Lacking that
I would try a 1K pullup (to 5V), 1 K pulldown (to ground) and 1000 pF capacitor to ground
(one at a time) on the parallel port pin that drives Z Step and see if they effect the amount of creeping

Also do you have the breakout specs?

Perfect example of why I asked, you are thinking of ideas that I don't even understand.

I do know a guy that I used to work with that does drive systems in robots for a living, so I may be able to get him over with some equipment to help with diagnostics.

I have attached the breakout board specs. This attachment is hidden for guests. Please log in or register to see it.

are the power supplys common at all , or a matter of using heavier power cables because of current draw ,
or just the simple fact the stepper drive is not large enough for the job in hand , or the power supply current availablity is dropping when the dc motor is on full .
causing a drop in the power availabilty for the stepper , are you running screened dc cable to the motor , if it's using a vfd ? then use sheilded cables

Also the DB1V2.0 breakout board does not isolate or buffer the step/dir signals
so if you have step/dir signals with a 5V common and a 3.3V parallel port you will
have marginal signals (3.3V high will not fully turn off a drive OPTO connected to 5V)

If this is the case I would insert a diode or 2 in the 5V line to the step drives common +
to drop the common voltage

Since the DB1V2.0 is not buffered, the pullups, pulldowns and capacitors can be
connected directly at the terminal blocks which should make it easier to try

Also just thought about it and the idle polarity of the step signal may make a big difference in suscepability to noise and this may be different between software systems

To aid in your troubleshooting you might want to start by checking your parallel port voltage, when you check a direction pin and you change directions it should toggle from 5 volt to 0.
Are you using the motherboard's parallel port or a pci card?
Do you have another a pci parallel port card to try?
I avoid having the machine turned on if EMC is not running as you do not know what state the parallel port pins are in and what may happen, however that sounds like a noise problem which should be addressed as posted above.

Rick G

Great news! When I set up my basement shop I had a dedicated 30A line run to the mill area, and I had the DC motor and the Stepper power supplies plugged into a common switch powered by this line. That way, when I was done working, I flipped one switch and everything powered down.

I had tried plugging the DC motor into a different circuit, and it did little to improve the problem. But, looking at CNCBASHER' s post, I tried another combination, leaving the DC motor on the original circuit, and moving the Stepper power supplies to a different circuit. That seemed to solve the problem. I will continue to test before announcing the gremlins fully beaten.


It is still a mystery to my why I have run with this setup for 5 years without issue, and why it was getting progressively worst just in the last few weeks.


PCW - I was looking around on ebay and there are some nice breakout boards available now for a pretty reasonable price. I might just pick one up for testing. Maybe something on my existing board is on it's way south.

Rick G - I was using the original motherboard's parallel port. Again, maybe I'll pick up a pci card for testing/backup.

Thanks for all your help.

Rick Denoon

Do you have a good solid ground to your panel? Is your neutral bonded to the ground?

John

40 year old house and wiring. I was thinking about running a dedicated ground line from the house ground (water main) and making sure that everything that could be grounded was tied to that ground. Might still be worth doing.

Thanks
Rick

I would consider driving a 10' ground rod next to the panel if possible...

John