Strange Z step loss drives me crazy, pls help!

Hello,

I have a problem with random step loss and cant solve it. If i move Z axis, i can hear the motor knock sometimes, and z-value gets corrupted measured with probe. X and Y work fine!

I already tried:

- Different Timings (step, hold, jitter ...)
- ALL combinations of microstep and amp settings on driver (previous TB6560, now FMD2725B)
- inverting step, direction
- inverting stepper cables
- changing pins from breakout board
- 5 to 5.75V breakout board supply
- 12 to 30V stepper driver supply
- Normal, Bidirectional, ECP+EPP port setting
- Thicker cables (3m) from stepper to driver, also stepper with 15cm cable directly on driver
- Drive stepper with/without load
- Curse, pray, sacrifice animals, ...


Then i found something out:
NO STEP LOSS IF I GO TO BIOS AND CHANGE PORT MODE - After this the stepper runs perfectly smooth.
BUT: This only works for stepconf wizard!!!
As soon i start axis, the stepper gets crazy again. After closing axis also step loss in stepconf again.

So i think stepconf uses a different method then axis to drive the stepper. Axis seems to change the parallel port or something else, which resets when i go to bios.

Has anybody an idea? Really thinking about getting rid of damn EMC and use GRBL instead...

My versions: Linux 8.4 - Axis 2.3

PS: I cannot install a newer version, dont have a cd-burner and EMC machine has no usb-boot-support.

Thanks in advance!

Don't know if LinuxCNC 2.3 has the feature but you might try setting the parallel port mode to "EPP" in your hal file
Also it should be possible to upgrade by moving your hard drive to a newer machine, installing LinuxCNC there
and then moving the hard drive to your machine controller (LinuxCNC 2.3 is really pretty ancient ~8 years)

Are you by any chance connecting the Z axis to pin 1 or 14 or 16 or 17?
If yes, try using pins 2 to 9.
If you are using a BOB ( break out board ) try it without it.

Thanks PCW, ill have a look at the hal file as soon my pc works again.

Tommylight, its definitely not the breakout board because:
"Then i found something out:
NO STEP LOSS IF I GO TO BIOS AND CHANGE PORT MODE - After this the stepper runs perfectly smooth.
BUT: This only works for stepconf wizard!!!
As soon i start axis, the stepper gets crazy again. After closing axis also step loss in stepconf again."


For now i have another problem: Tried a Ubuntu 12.04 LiveCD, and after initialising the CDRom, the PC shut down. Tried to restart - No reaction, PC is completely dead!!!

Checked the PSU, works, but as soon i connect it to the mainboard, it shuts down. Made a CMOS reset, no reaction. Then i disassembled the mainboard, made a longer CMOS reset, took the CMOS battery out for a while, then put all together again. Now the PSU does not reset anymore, CPU fan runs, but PC does still not boot


Im done with Linuxcnc, installed an Arduino Uno with GRBL1.1 and it works perfect!

Then i found "UGS Platform", the new Universal Gcode Sender with drag and drop customizable menu, custom buttons, and continuous jog control like in Linuxcnc. I can totally recomend it for all who have problems with Linuxcnc.

Maybe i change back to Linuxcnc one day, but for now im happy with GRBL and UGS Platform.

Solutions to my problem are still welcome!

Thanks!

No backlash comp in GRBL, unless they have made some changes.

From what you described with the PSU\MB that's not a linuxcnc issue....that's a hardware issue.

It was a 12.04 LiveCD, not LinuxCNC. There are a few reports on the web where linux killed hardware! I guess linux made some changes in bios, cpu or power management which caused death...

UPDATE:

New PC, LinuxCNC 2.7 wheezy, SAME PROBLEM!
X Y work, Z goes crazy. All the work for nothing

So i'll have to work with GRBL, but i'd prefere Linux because it supports parameters, subroutines and loops...hope this feature will come soon for UGS.

PS: My Machine has no backlash, its a lightweight belt driven cnc with medium accuracy (+/- 0.015mm) but high speed and acceleration (300mm/s XY = 18.000mm/min, 2000mm/s² acc)



Thanks so far, topic is closed.

Kolbi wrote: It was a 12.04 LiveCD, not LinuxCNC. There are a few reports on the web where linux killed hardware! I guess linux made some changes in bios, cpu or power management which caused death...

Thanks so far, topic is closed.

Geez I'm been using Linux since about 96 or 97 and never ever had Linux killed any hardware, I'd like to see these reports, just out of interest.......I know there was reports of ubuntu 17.10 bricking Lenovo, Acer and Dell laptops but that was an Intel spi driver issue.

ozzyrob wrote:

Kolbi wrote: It was a 12.04 LiveCD, not LinuxCNC. There are a few reports on the web where linux killed hardware! I guess linux made some changes in bios, cpu or power management which caused death...

Thanks so far, topic is closed.

Geez I'm been using Linux since about 96 or 97 and never ever had Linux killed any hardware, I'd like to see these reports, just out of interest.......I know there was reports of ubuntu 17.10 bricking Lenovo, Acer and Dell laptops but that was an Intel spi driver issue.

Same here, must have used over 100 versions of Linux over the time on at least 200 types of hardware and never ever killed anything !!! Hell i had computer shops, worked as an IT admin for the largest organisations on planet earth ( they still use Windows, but i used Linux on all their hardware, their servers and firewalls are on Linux, etc ).
Such statements are really not helpful at all.

Sorry for my childishness....

But I hear these are great for long term windows users