Z moving negative????

Things started getting very wierd this weekend.

I usually print with MC, and I've been using it about a year now, so I don't think I am doing anything wrong there. I can drive the printhead manually in all three axis, and it does as expected. However....

During a print, the printhead will stop increasing in Z and sometimes go NEGATIVE. I've caught it with the printhead down on the glass, and the Z nuts about to disengage! It never happens when I'm watching of course., and simple short test prints work fine.

Having discovered a failure, I can abort the print and go into manual control, and everything works as I expect again.

Thinking I might have somehow hosed my settings in MC, I installed Cura. Same problem.

Given that I can drive it manually 10mm, 100mm, and in any direction, what could possibly be going wrong here?

 

The Z end stops look like they've always looked.
The printer will home and level the bed just fine.

Don't know about the comm log, this is the same printer/laptop/cable that's been working for the last year. Windows updates of course.

The laptop never sleeps or hibernates.

Where would I find the comm log from MC?

 

Two different object files, on two different pieces of software are having the same problems.
Both object files have printed successfully in the past on MC.
The Cura install is new, and specifically in response to this problem to see if I somehow screwed something up in MC.

I can look at the Gcode window in MC, but I HIGHLY doubt that MC is suddenly telling the printer to move negative in Z during print other than the normal rocking back and forth as the printer transits X and Y. And I see that happening exactly as expected.

I think something is wrong in the printer, but I'm at a loss as to what it might be since I can can control all three axis manually without issue.

Also, FWIW, the Z motors turn freely by hand when the printer is powered down, there's no binding.
The carriage is level to the printbed.

I'm trying a print of the cal cube in MC now, simply because I don't know what else to try. No odd sounds, everything smooth so far.

Cal cube printed ok. Terminal output showed no surprises, Z steadily increasing.

Trying one of my parts now.

Watching the Gcode, so far no problems, Z keeps increasing by 0.1mm.. I don't imagine that MC and Cura both decided to send the printer into negative Z territory.

After 8 layers, the printhead is scrubbing plastic. It has apparently not moved totally negative, but it's nowhere near the 0.8mm above the bed that it should be at this point. I've seen worse, where the Z motors have gone negative enough that they are backing out of the carriage, and the head is scrubbing the glass. The Z numbers in the terminal window though are what i'd expect at this point in the print. Through the whole time, the Z motors are rocking back and forth a few degrees as the X and Y motors move, as normal. No odd noises, nothing out of the ordinary except that the head is shoving plastic out of its way again.

So I tried another cal cube print. Somewhere around the last mm of the print, the Z stopped working, and the head is now shoving plastic around. The Z numbers in the terminal window are as expected, 10.499 at the moment and still going up by 0.1mm per layer. So it's not in the Gcode.
The cal cube now has a nasty look at the top, with plastic shoved out well past it's proper sidewalls.
Meanwhile, the Z motors are still rocking back and forth as I print, exactly as normal.

So, the printer hasn't lost control of the motors, and the Gcode should produce proper output.
The last half mm of plastic ended up shoved out of the way by the printhead, off to the sides. Ugly.


Tried another cal cube print just for laughs.
The Z value is now 3.099 in the gcode window, and the head is scrubbing the glass.
It has shoved the material that was printed out of place, and is now climbing in Z again.
So, it went to about 1mm properly, then started either not ascending, or descending some then ascending again.

I've printed three cal cubes tonight. One without issues, one failed very late in the print, one very early.

 

The R1.

 

Well, thanks Mark! I was hoping that wasn't the problem, but it looks like my ramps fan died, and it damaged the Z stepper board. That board is kinda hidden under wires and apparently never did get good ventilation.

I bought the last one in town last night, and replaced the dead one, also I added extra cooling under the case in the form of two fans blowing air across the ramps board. After calibrating the current on the new board (a bit of a PITA I wasn't expecting) it's all back together and is running its first test print. The Z board draws a lot of current when I put it on my 5V power supply, so it's definitely cooked and I'm surprised that it was actually still stepping at all. Allegro does make some amazing chips.

 

Well... It worked for a while. Now having the same problem.

I am hoping it's due to the absolutely bottom of the barrel excerable miserable cheap pots used on those boards.
Pots are a bad idea to start with, but those pots in particular I would never use in a design.

Is there a doc somewhere that tells what the motor current should be? I'm inclined to replace the pot with a fixed resistor.
I have seen a number of 0.462V for the REF setting on the Z motor, but I'm not sure that's accurate. I could set it up to run full rated current, but I would hope that was overkill.

 

I bought the last one that Micro Center had. So they may be inexpensive, but they aren't much on immediate availability.

 

I ordered those on friday.

I would have liked to get hired by Lulzbot. They are local to me, and I could seriously help them with EMI issues (FCC part 15 and CE) and the general poor design quality of the arduino world. Arduinos are fine for hobby hacking but they are in no way a full production ready design. I've been designing with those processors since they first came out. I applied there recently, but didn't hear back, and I just took the position of lead hardware engineer at Backcountry Access, making avalanche beacons.

I've used two pots in designs in over 30 years of professional engineering. One was a volume control. The other was a 10 turn precision pot. Pots suck.

If it were me, I would have put the pot (or preferably EPOT) on the RAMPS board, and brought the reference voltage up to the stepper board. That way the calibration is on the RAMPS, and the stepper boards are then just generic boards that need no calibration. That's just one issue, I see many in the system, but that's what I get paid for. For example, the mechanicals of those boards violate a rule I have been designing to for decades, "If it fits, it works, or it does no harm". It's possible to plug them in backwards, and then all kinds of magic smoke gets let loose.

Anyway, I just went under the printer and like I figured the pot had drifted from where I adjusted it. It went from 0.46 to 0.42.
I was working from this:
http://community.robo3d.com/index.php?threads/stepper-driver-calibration-information.6466/


Those pots are IMHO, worse than worthless.

 

I do like my prime membership!