HP 8112A repair

[flyglas]

Hello,
I have bought a HP 8112A pulse generator with error code E01 (RAM pattern test failed).
With two new RAM chips the self-test shows no error code, but the error LED on the frontpanel is on and the WID indicator is blinking.
No reaction to all buttons execept of button LCL. After pressing LCL the display changes to A(blinking dot)00.

The service manual relies on error codes, but I do not have one.
All voltages are within their limits.

Perhaps somebody knows this behavior and can help me?

[flyglas]

A(blinking dot)00 seems to be the GPIB address. I can change 00  to 01...31 with the DIP switch on the back of the unit.

[MarkL]

I have an 8116A, which is similar to the 8112A.  Here are some ideas that might help.

Although the manual doesn't explicitly say, I'm going to guess the blinking parameter LED, "WID" in your case, is an indication that various settings conflict with each other.  For example, setting a width of 100us when the period is 10us.  In your power-up photo, a period of 0.00ns I would say qualifies.  (The 8116A uses the ERROR indicator in this way but does not blink the parameter LED.)

The settings are stored in battery backed RAM when the unit is off.  Because you replaced the RAM and maybe the battery too, the contents of the RAM are now garbage and appear to be creating a nonsense setting.  The manual says the firmware is supposed to restore the unit to factory default settings when RAM corruption is detected, but maybe this is not happening and/or the firmware is totally confused.

You can try performing the factory setup manually by pressing STO/RCL and then use the right hand vernier to select "0" if not already displayed.  Then press the middle bottom vernier to recall.  However...

You also mentioned the front panel buttons are not working except LCL.  There's nothing special about LCL.  It's in the keyboard scan like all the rest of the buttons, so it's strange it's the only one working.  Perhaps there's something mechanically wrong with the keyboard (maybe a cracked PCB or open traces).

What I would try next is to see if the unit can detect button presses at boot.  At power on, the processor does a scan to see if any buttons are stuck.  You can use this to verify the function of all the buttons.  Start with the LCL button.  Power off, hold the LCL button, and power on.  A "o" should immediately appear in the display.  As soon as you release it, the unit should complete the boot sequence.

Now try the same thing on another button, like STO/RCL.  In the 8116A, and probably in the 8112A too, as long as you press the next button you want to test before releasing the previous one, the unit will stay in the "o" mode.  You can use this to quickly verify all the buttons on the front panel.  This should at least help point to either a firmware or a hardware issue for further troubleshooting.

Oh, and if you haven't done so already, make sure all the power rails are in spec.

[flyglas]

Thank you very much for your answer and tips.
My understanding of the manual is also that the blinking WID LED and ERROR LED indicates an invalid setting
and no hardware fault.

I have tried to perform the factory setup by pressing STO/RCL, but nothing happens. No response at all.
Every single button shows "o" in the display when pressed at power on. Therefore, the complete front panel (indicators and buttons) is working properly.
The voltage rails on the main board and on the microcontroller board are within spec.

The symptom "only LCL is working" matches with the following description in the manual:
"When in remote control mode all the front panel keys, except (LCL),
are ignored by the microprocessor. If the local lockout command has
been received on the HP-IB then the key is also ignored."

The only source for remote control is the GPIB controller. I have removed the GPIB controller chip and rebooted the unit. But this does not change anything.

At the moment I suspect the RAM. How trustworthy is the RAM check at power on?
Perhaps the check does not find an error, but the data read by the controller is garbage?
The original RAM UPD444C-2 was replaced with US224D20. The US224D20 should match all parameters of the UPD444C-2.

I will try to hook up the logic analyzer to the RAM. Perhaps this will reveal some information.

[MarkL]

I don't know how thorough the RAM test is on these units.  It could be determined by disassembling the ROM image, but I don't see any images for the 8112A.

It sounds like it's stuck in local lockout mode.  In local lockout the LCL button is ignored  (service manual page 10.7-1).

That was a good idea to try removing the GPIB controller.  Do you have a GPIB interface for your PC?

Going on the theory that something in RAM might be confusing the firmware, one thing you could try is to put different garbage in the RAM.  Turn off the unit, remove the battery, and temporarily short Vcc and GND on the memory chip.  You could try this a couple of times if unsuccessful at first.  (It's possible you could remove W2 instead of removing the battery.  The schematic is unclear.)


EDIT:  If you're going to put a logic analyzer on the RAM, you could look for a string of reads and writes at boot time that sequences through (most likely) all the addresses.  That should tell you what it's doing to verify the RAM.  This is best done with a logic analyzer that has enough bits to watch the RAM control, data bus, and address bus simultaneously.