HP97 memory fault?

[Hans-Peter]

My 97er works in run mode without any trouble. The printer works also.
But the cardreader says on every card "Error". I fixed the transport spindle with the silicon tube but no luck. Then I wrote a program with a endless loop just for fun. After about 1-2 minutes the calc stops the program. Then when I switched to pgm mode the display shows the step 994. Back to run mode the calc works fine but the pgm mode was stuck. Pgm clear no chance. Power off, wait for a minute and the pgm mode was fine. But only for a short while.
A optical check of the CPU Board with the roms shows a slide black film over the silver legs of the roms, like my silver can in my livingroom. No corrosion like a battery drain.

Is there any chance to fix it without change the complete Mainboard?

   

Hans-Peter

[Paul Berger (Canada)]

That does sound like it is potentially a memory issue, and unfortunately the memory is in custom RAM/ROM chips so the only source is likely another 97. One thing I would check is to make sure all of the voltages in the calculator are correct. There is a service manual for the 97 on the MoHPC memory stick that include schematics. The card reader problem may or may not be related. The cards are indeed read into program memory, however since you say your program loops runs for a few minutes before crashing, if it was a memory fault I would expect you would be able to read a card when the calculator is first turned on. Aside from the drive wheel there are a number of other things that can cause cards to not read such as, wrong speed, wheel too high or too low as adjusted by the eccentric, broken or incorrect connections to read head, switch timing, bad connection between card reader and the main board, or a fault in the card reader chip. Incorrect voltages to the chips can also cause card reader issues.

Paul.

[teenix]

(09-16-2017 05:25 PM)Hans-Peter Wrote:  A optical check of the CPU Board with the roms shows a slide black film over the silver legs of the roms, like my silver can in my livingroom. No corrosion like a battery drain.

Is there any chance to fix it without change the complete Mainboard?

Hans-Peter

The "black" is probably age related corrosion or discoloration on the surface of the IC pins and may have affected the soldering.

If you have a magnifying glass, you should see shiny silver looking solder joints on all IC pins. If there are some that look a dull grey colour and perhaps pitted where I joins the actual pin, it might indicate a dry joint causing the problem.

With care, these can be resoldered.

It is possible that a memory "bit" after all this time is beginning to fail either in the ROM or in RAM. It may work ok for awhile, then fail on a specific read or write causing the crash. If it was a totally failed "bit" then the problem would most likely show up all the time and programs would not run.

If the calculator functions ok with all normal operations for a long period of time, including storing and recalling to/from memory, then I would suspect a ROM issue where the PRGM/RUN microcode resides, or in the program memory.

These types of problems are hard to solve without parts substitution or specialist test gear.

If it is a program memory problem, you might be able to sequentially test it by building up a simple program. That might cause a fail after a certain amount of instructions are added.

LBL A
GTO A
------------
LBL A
+
GTO A
------------
LBL A
+
+
GTO A
-----------
etc

Can you reliably on multiple times enter a program and verify by SST in PRGM mode, wait 5 mins and verify again.

cheers

Tony

[PANAMATIK]

(09-16-2017 05:25 PM)Hans-Peter Wrote:  Is there any chance to fix it without change the complete Mainboard?

According to your description I suspect a ROM chip or ACT chip failure.

For verifying a defective ROM code, it would be possible to readout the ROM code with the "new ACT" and compare it to the known HP-97 code variants. The procedure is to unsolder the original ACT, insert new ACT, read ROM code, and test the original ACT in another HP-97 or test another ACT in your machine, then reinsert the original ACT if it is OK. If the ACT is the cause, I would have an original HP-67/97 ACT as spare part for you. If the PIK or CRC chip is defective I have no solution, you need another HP-97 for spare parts.

If you consider to send your calculator to me, I would check the above.

Bernhard

[Hans-Peter]

Thanks for your sympathy.

Now I have tested all normal operations and all registers many times.
In the run mode, the calc works perfect.

The measurements of voltages is my next step.
The change of the ACT is perhaps a deal for next year at the Allschwil Meeting with Matthias. Then I would prepare a socket for the "new" ACT.

Hans-Peter