RAM Parity Error and Other CPU Error, Fanuc 15mb

So I'll attempt to make this relatively short but I feel this will be a bit long winded, hopefully someone can shine some light on this.

A couple days ago, our Kitamura h400 with 15mb control gave a "SYSTEM EMERGENCY - OTHER-CPU" error that required a restart of the machine. I'd never seen this error before. I restarted it and it gave no more issues for the next 2 days.

This morning, I have gotten 3 times a "RAM PARITY ERROR - CNC-MAIN FILE-RAM" error. Twice while running, one on startup just before it pulled the MCC.


I am wondering, are these alarms related?

I did notice yesterday, the machine started behaving weirdly. The air blast for the pallet changer was delayed by a couple seconds, and there was a time when the air blast for the tool changer stopped working. I noticed pretty quickly, and "fixed" the tool change air blast by commanding the air blast M-code a few times. The pallet change air blast remained delayed the rest of the day, but it is just fine this morning.

My limited knowledge of the errors I see on the screen lead me to believe that the first error, "OTHER CPU" is related to the PMC, which leads me to believe that the PMC board might be failing (which might also cause the air blast delay?).

But the RAM parity alarm seems to relate to the main CPU board, because that is where the FILE RAM is. Could it be the board that is failing, or the RAM card itself?

Here are pictures of the CPU error:







And here is a picture of the RAM PARITY ERROR:




Edit: one thing I just noticed, on both errors, the ACT TASK was "FFFD". And on the second image of the CPU error, it gives hardware information, and where it lists the PMC CPU board, it gives a value of "FFFD" for it (see "SLOT I/O" line under PMC CPU Board)

Does that indicate that the PMC is calling up the memory? Or is the task value of FFFD just circumstantial?

If I had these type errors and symptoms, the first thing I'd do would be put the oscilloscope on the power supply and check for noise. Particularly on the +5V line but all need to be checked. If you find more than ~50mV then replace all the electrolytic capacitors in the power supply. Then re-check. Usually that's all it takes to quiet it down.

Even before that, power it down. open it up and see if the memory is plug in simms, dimms etc. If it is, then unplug them and plug them back in. If the board is dirty, blow it off with some 'canned air'.

FIRST: do a SRAM backup (see BOOT section in the maintenance manual)

Vancbiker said:

If I had these type errors and symptoms, the first thing I'd do would be put the oscilloscope on the power supply and check for noise. Particularly on the +5V line but all need to be checked. If you find more than ~50mV then replace all the electrolytic capacitors in the power supply. Then re-check. Usually that's all it takes to quiet it down.

Click to expand...

Just out of curiosity what kind of filters are you applying and what bandwidth scope? I normally see 150+ mV after new caps, although I haven't done that many. The ripple smooths out but I've never been able to get rid of the noise

I use the 50mV as a guideline that I remember from a Fanuc 10, 11, 12 series maintenance manual. I use an old Fluke 50Mhz digital scope. It has a default "smoothing" factor. I don't recall the details, but never change from the default setting.

memoryman said:

FIRST: do a SRAM backup (see BOOT section in the maintenance manual)

Click to expand...

Do some versions of 15M have the BOOT function similar to the 16/18 series? I've worked on a few 15 controls, but do not recall seeing one with that.

Thanks for the replies guys.

I don't have access to a oscilloscope today, but I did have another power supply that I ended up swapping it with this morning.

It did not give me any hard alarms, but the air blast for the tool changer was delayed again this morning, even after a reboot. I swapped power supplies and that went away, but I'm not sure if that was really the problem?

I checked voltages, they were very good just with a multimeter. But I didn't check 5v on the power supply, because it appears that there is a testing pin on the board, but I'd have to run a jumper wire to it after plugging it into the bus? I did not want to turn on the machine with the power supply removed, because the memory gets battery power through the bus.

I did also check voltages on the fanuc I/O boards, they were all good, 5v included.

SRAM is fully backed up, and I re-backed up programs/offsets/parameters again this morning.

Is BOOT the same as IPL?

If you have a higher end multimeter like a Fluke 87 you can put it on AC and sometimes get an idea of how much noise/ripple is present. Still not even close to a scope tho.

dandrummerman21 said:

.......Is BOOT the same as IPL?

Click to expand...

BOOT is a 16/18 series screen that allows one to handle system and builder files via SRAM PCMCIA cards on all versions and Flash PCMCIA cards on later versions.

Will be interesting to see if the parity or CPU errors persist.