GE 90-30, CPU 363, Logic equal failure... firmware? (Pictures)

Continuing saga from an earlier service call... My customer has a machine with a 90-30 PLC. We couldn't get logic equal by uploading the program, so I recommended that they get a spare CPU, and we use that one to try to get logic equal for troubleshooting.

The customer found a reconditioned one online and bought it. That one had worked for getting logic equal, but there was an issue with the flash memory, and the program was never saved during a power cycle. So he returned it for a replacement.

The replacement came, and I went over to install it today. After uploading, I got a little red "F" over the "logic equal" green icon (see picture). I checked the diagnostics and found a configuration error (see picture). I realized the problem was probably caused by a mis-match of the firmware. The CPU labels confirmed that the replacement CPU had a newer version of the firmware (see picture).

I need to ask... Is this an indication of a firmware issue? Do I re-flash the firmware back, or change a setting in the configuration and re-download? If I flash the firmware, how do I get the file, and do I use Profacy ME, or another application?

I realized the problem was probably caused by a mis-match of the firmware.

Click to expand...

No! Configuration mismatch indicates that the module actually present in a slot is not the same as the hardware configuration file. In your case the mismatch is in slots 2, 3, and 4. Double check the modules installed against the hardware configuration.

OK... I see how the slots are addressed in the diagnostics. But that makes no sense to me. I uploaded the file from the old CPU, then downloaded it to the new CPU. It's the exact same file. Is there some quirk I'm missing?

Is the new CPU mounted in a different rack with a different population of I/O modules? When you uploaded from the old CPU did you include the hardware configuration in the upload?
Configuration mismatch means the CPU has detected a discrepancy between the modules actually present in the rack and the modules it expects to see based on the hardware configuration. To the best of my knowledge, configuration mismatch does not include firmware level difference. Also, the location parameter in the fault table points to the location of the mismatch. If the mismatch were in the CPU module the location would be rack zero, slot 1.

I do not believe the 90-30's had a way like the newer Rx3i's to flash upgrade via USB. I believe upgrades were done via "kits" that you had to order from a distributor and I believe it was actually a chip swap out...not actually done one though and I'm not certain you can backwards flash them, you would have to speak with a distributor if these kits are even still supported.

Either way, I second w/ Steve that the firmware should not cause you a mismatch. Either the hardware order is out of order or you have some actual hardware faults. Make sure you copied the hardware configuration entirely and I usually do logic and initial/forced values as well. Another key thing to remember too is the 90-30's are not hot swappable like Rx3i, need to power down and do the swap, then power up, otherwise the CPU faults.

Your checksum picture states hardware and logic are Not Equal, so I would guess that the download was never successful to the new CPU unit you're trying to use. Make sure no passwords are enabled or that the front key lock is not in a position of blocking your program download (on/off position key slot that stops downloading). I know that's general knowledge, but even I forget sometimes after years and go "wth is going on here!?" haha.

Supported quite a few of these in the mid 90’s. Firmware upgrades were done by replacing EPROMs.

For CPU models 350 and up, firmware upgrades are done by reflashing. For models prior to that you had to change the EPROM.

I need to go back to the customer's Monday afternoon and work on this some more. I was looking through the drawing and stumbled upon this. I haven't found how it works yet, but I'm wondering if GE PLCs have the ability to add a system-type bit that prevents program changes to the PLC. I figure it would be more likely to target the recipe setup in the PC/HMI, but I can't leave any stone unturned at this point.

You would have to look at the ladder logic to see what the original programmer was doing with input %I0021. The CPU363 module has a keyswitch that can be configured to prevent changes to the program. That and the password levels are the only ways I know to prevent changes to the ladder logic.

Edit:
I took a look at %I0021 in the project file you sent me last month. That input turns on internal bit %M1021 which is used in a lot of places including several one-shot contacts when it is turned on and off. Several of the instances prevent calling he program block "IRTCO" which moves a lot of data into a bank of registers starting at %R6181. That block of registers is zeroed along with many other registers on rung 65 of the block "IRT".

You're right to question to assumed procedure details, but I've been getting more familiar with the GE PLCs, and I believe I'm covering all of the bases. I turn the key off, upload the program in program mode, and I'm including all of the check boxes when I upload or download. I power down the PLC, switch the CPUs, power it back up, connect to target, go into program mode, then download everything- including the "write to flash," turn the key back on. And this is where I'm getting that "F" on the icon.

But just now I thought of something.... how important is that detail I remember hearing once... that when I load Profacy ME, I right-click and "run as administrator". Seems like a long-shot, but I should ask.

I don't know the answer to your question about running as an administrator. On all of the computers I use to run Proficy I am logged on as an administrator so it has never been an issue for me.
The picture you appears to match the module population in the hardware configuration in the project file you sent me. But was the diagnostic picture you posted in the first post in this thread taken when you downloaded to the spare CPU363 the customer purchased? And if it was, did you take out the original CPU and install the spare in its place? Or did you mount the CPU in a spare rack with a different I/O module population?

Fair question... but, no, I put it right back where the old CPU was.

I think about the previous CPU that we returned... I got logic equal, it just wouldn't save the program to flash.

Steve Bailey said:

Edit:
I took a look at %I0021 in the project file you sent me last month. That input turns on internal bit %M1021 which is used in a lot of places including several one-shot contacts when it is turned on and off. Several of the instances prevent calling he program block "IRTCO" which moves a lot of data into a bank of registers starting at %R6181. That block of registers is zeroed along with many other registers on rung 65 of the block "IRT".

Click to expand...

From the sound of that, it seems like it is the recipe it is writing... cycle times, etc, which is what I suspected. That would make sense because this machine had to be "certified" to run the parts it's running, and not altering the recipe seems like a logical procedure.

I've met some really great people in this profession, and Steve you are definitely one of them! We are really fortunate to have your knowledge and experience as a resource on this board.

I achieved logic equal with the new CPU with Steve's help offline. It appears I may have had some issues with the serial port on the old CPU. I didn't know there was another serial port behind the door on the power supply. I also downloaded to the flash separately. I'm still not positive what happened before, but we seem to have everything figured out now... knock on wood. We are just going to watch the ladder and figure out where the problem is.

AutomationTechBrian said:

I've met some really great people in this profession, and Steve you are definitely one of them! We are really fortunate to have your knowledge and experience as a resource on this board..

Click to expand...

+1 He has saved my a$$ more than once