Membrane Keyboard Quit Working

[10k]

My membrane keyboard on my M400S quit working. I've tried several things to get it working again without any success:

- Replaced with new membrane keyboard ($650 - ouch)
- Pulled daughter board from motherboard, cleaned connectors on daughter board and right angle adapter with scotchbrite. Examined PCI connector on motherboard. Pulled cable from daughter board to membrane keyboard and reset.
- Changed out SD hard drive with working version from a few years back
- Dug through set up parameters within Centroid. Didn't find anything to change except for ability to turn on/off membrane keyboard. Went ahead and changed, then rebooted. Still does not work.
- Noted that when booting program, lights on membrane keyboard flash, so I don't think it's a power supply problem.
- Unplugged pendant in case it had a short.
- Checked cable from daughter board to membrane keyboard with voltmeter for breaks or shorts.

Mill had been running for years, and nothing was changed on Centroid. It just stopped working one day.

I'm not sure what to do next. It could be a problem with the daughter board, but since all other functions it performs seem to work OK, it's probably unlikely. I have been able to:
- home system
- run programs
- move the table using G-code using MDI

I get a message about the PCI when I boot Centroid. Here it is:

PCI driver not loaded, checking for legacy card
Dual port base address = d0000
Return code 66

When I look at the computer boot log, I see this right near the end of the log:

PCI: Found IRQ 11 for device00:07.2
PCI: Sharing IRQ 11 with 00:0b.2
PLC driver v1.20
Pci9050: ERROR no supported devices found

I don't know if this is really an error. It may find a legacy card, and I don't remember if I've seen this message before when it was working. In any case, Most CNC functions still work, so it must be talking to the card.

I'm running Centroid version 2.7 on Red Hat Linux.

Report.zip attached

Any ideas?


(Note that this is a repost of a question originally asked on IndustryArena with no replies in a month!)

[dpascucci_support]

Your report msg_log.txt does show quite a few intermittent "422 Check Jog Panel Cable" messages. This makes me think that there's an issue with the cable causing the trouble communicating with the software.

[cncsnw]

I get a message about the PCI when I boot Centroid. Here it is:

PCI driver not loaded, checking for legacy card
Dual port base address = d0000
Return code 66

When I look at the computer boot log, I see this right near the end of the log:

PCI: Found IRQ 11 for device00:07.2
PCI: Sharing IRQ 11 with 00:0b.2
PLC driver v1.20
Pci9050: ERROR no supported devices found

Those messages are normal and expected when running CNC10 with an ISA-slot CPU board, such as your CPU7P5S.

I infer that you are inquiring about your jog panel keypad, and not the PC keyboard. Some Centroid controls use a membrane keyboard for the PC keyboard; so it is important to distinguish. The jog panel connects to the CPU7 board, via a ribbon cable and rectangular connector. The keyboard connects to the computer motherboard, via a PS/2 connector.

[10k]

dpascucci_support-

I checked the cable with a voltmeter and saw continuity on all the lines. I also checked for shorts between the lines. I'll go back and recheck it tomorrow, but I think it's ok.

cncsnw-

Thanks for clarifying about the messages being normal. I suspected they were, but I could not confirm that.

You are correct, when I say membrane keyboard, I am referring to the jog panel.

[10k]

I carefully rechecked the cable. I checked pin to pin for continuity, pin to other pins for shorts, flexed cable during checks to see if there was a break, examined connectors for tightness. All looks OK. I was hoping to find something, because that's an easy fix!

[cncsnw]

I assume, in the course of this testing, that you have verified you have the cable plugged into the "JOG PANEL" header (H1) of the CPU7P5S board. It will physically fit in the H9 "expansion" header, but it will not work there.

Also, many controls had jog panel ribbon cables with no orientation key, so it was possible to plug the cable in backwards. As a general rule, the lettering on the black plastic connector housing needs to face in towards the middle of the board (away from the edge of the board).

http://www.cncsnw.com/CPUs.htm#CPU7P5S

[10k]

I checked and verified it's in the right plug, and that it's oriented correctly. It's fairly obvious, since somewhere along the way someone used hot glue on the edge of the connectors to make sure they didn't vibrate out. It's still on both halves.

I talked to the operator, and got some additional info on the failure. The jog panel was starting to get "hard of hearing", and sometimes he had to press the buttons more than once to get them working. It got steadily worse until it totally failed. That's why I thought that the problem might be with the jog panel itself, and replaced it.

[rjtechserv]

I'm wondering if the cable got crushed or flexed enough it is shorting with high resistance or open.