HP-67 weird behavior after reading a card

[albertofenini]

Hi everyone

we are looking for advice ...
this unit works well, though we haven't tried to input manually a program
reading a card, in this example SD-04A, works fine too
when trying to input some data, in this case we wanted to find out the weekday for a certain date,
it behaves erratically as shown in the picture
other CPU boards coupled with this card reader and this chassis have no problems at all ...
what could it possible be wrong ?

pictures can be found here : HP-67 acting weird after reading a card

thanks in advance and take care !!

[teenix]

Maybe it is a RAM issue. This small program will span the 2 RAM pages for program space. When run by pressing [A], the display should show 8.00

Code:


f LBL A
7
ENTER
8
X
GTO B

GTO .112

f LBL B
7
/
h RTN

If it works, you would have to try filling the program space with any simple program and see for a proper result. Something like 1234567 +, repeated until step 217.

Press [g] [RTN], CLx and R/S, the result is: 33333309.00

If this works, then maybe a CRC or ROM issue.

cheers

Tony

[albertofenini]

Hi Tony, hi everyone,

will definitely try, and let you know
it's a little less clear what we should do to fill the program space,
just typing 1234567+ until the program line is number 217 ? (shame on us if it was obviuos)

BTW, we also have another CPU board that when reading cards keeps the
motor running a little after the card has gone thru and show Error on the
display, no matter what card we try to read

Card reader known to be good and so the chassis, what could be ? the CRC ?

Thanks for help, take care !

[teenix]

It was just to fill program memory with steps and when run give a known answer, in an attempt to prove the RAM is ok.

The card motor running briefly after reading a card is normal operation. The HP ROM code does this when it encounters a checksum error.

Same possible issues, CRC, ROM, RAM, pin connections, corrosion, 60uF power supply capacitor.

cheers

Tony

[albertofenini]

Hi Tony, hi everyone

We tried to key in the program on this CPU board which has serial 1706A0091.
BTW it has a small circular label with a X on the 1820-1751, don't know if it helps

Anyway, the first 6 steps were entered ok, but when we addressed the 112 step
with the GTO .112 instruction, the calculator positioned itself to line 98 which
contained 84 as value

Pressing the SST key once, moved the calculator to line 105 which contained
31 24 and going forward we found instruction 31 25 11 at step 114
Pressing SST more time take the calculator to line 995, 996 .....

what do you think ?

we also tried this program on the CPU board 1805S02334,
the one that let slip the cards when reading them and it worked ok !

it also worked fine on other CPU boards installed in this chassis with this
card reader

thanks again and take care !

[teenix]

The X on 1820-1751 might indicate the CRC is faulty, but hard to say. Does it look like it has been replaced, ie new looking solder.

The programming fault seems like a RAM issue. RAM location $3D nibble [2].
If the calculator was just turned on, all program steps should be R/S (84).

You might be able to switch the calculator on, switch the PRGM mode and SST all the way to step 111 and see 84 in the display. If not there is a problem.

If that passed, SST to step 112. If you now start to see issues, it seems the steps 112 to 224 might have RAM problems.

Can you change the display settings, DSP 3, DSP 9, FIX, SCI, ENG? If not, it seems like a RAM issue.

cheers

Tony

[albertofenini]

Thanks Tony
Will try tomorrow
Eventually, which chips are the RAM’s one ?
Take care and thanks !

[teenix]

The four 8 pin IC's above the CRC.

[albertofenini]

Thanks Tony
Is there any way to find out which are the faulty ones ?
Thanks a lot for your support and take care !

[brouhaha]

(11-15-2022 11:20 PM)teenix Wrote:  The four 8 pin IC's above the CRC.

It should be noted that those are combined RAM/ROM chips. They are mask-programmed with both their ROM code, and their RAM address selection. They can only be replaced with the same part number (1818-nnnn) from another 67 or 97. (Some but not all RAM/ROM chips are shared between the two models.) There is no source for compatible parts other than other HP calculators; these were chips custom-made for HP.

I have information on which ROM is in which chip, but I don't have information on which RAM addresses are in which chip. Teenix may have that infrormation.

There are two different HP-97 PCB layouts, which position the RAM/ROM chips differently, so they are best identified by the HP part numbers.

The RAM/ROM chips that are unique to the 97 are 1818-0229 and 1818-0230.

The RAM/ROM chips that are common between the 67 and 97 are the pairs 1818-0226 and 1818-0228, or 1818-0550 and 1818-0551. These must be paired, i.e., you have to have -0226 together with -0228, or -0550 together with -0551, but not a mix.

There is one other 8-pin ROM chip in the 97, 1818-0233, which does not contain any RAM, and it has a different pinout than the ROM/RAM chips.

The 97 also has ROM, but not RAM, in the ROM0/anode driver, but that one is not an 8-pin. It's either 1818-0225 or 1818-0267.

[albertofenini]

Thanks for clarification
We are going to source those from another HP-67 that had a problem with the ROM0
Hopefully the RAM chips could be good
Thanks again !!!

[brouhaha]

(11-16-2022 07:01 AM)albertofenini Wrote:  Thanks for clarification

Sorry, I was confused and wrote mostly about the 97 rather than the 67. It's much the same, though without the PICK (aka PIK) chip. The part numbers are:

ROM0/Anode: 1818-0227 old (not compatible with 1820-1812 ACT), 1818-0268 new (compatible with 1820-1812, 1820-1596, 1820-2530 ACT, possibly others)

ROM1/RAM: 1818-0228 old, 1818-0550 new (must be matched pair with ROM2)
ROM2/RAM: 1818-0226 old, 1818-0551 new (must be matched pair with ROM1)
ROM3/RAM: 1818-0232
ROM5/RAM: 1818-0231

[albertofenini]

Hi everyone

This the chipset on the CPU board not working :

1818-0550
1818-0551
1818-0231
1818-0232
1818-0268
1820-1751
1820-1596

This the chipset we have as a donor

1818-0226
1818-0228
1818-0231
1818-0232
1818-0268
1820-1751
1820-1596

What do you think ?
Thanks for any advice and take care !

[teenix]

One of these should be the culprit.

1818-0231
1818-0232

If you change them, after removing both of them, can you solder in just the 1818-0231.

[0. 00] may appear on the display, or the normal [0.00], when you turn the calculator on.

If you can do this, please let me know so I can clarify some RAM addressing and update my records. I'll post the result.

You can then replace the 1818-0232 and hopefully everything will work again.

cheers

Tony

[albertofenini]

Hi Tony, Hi everyone

we just tried,

first we changed the 1818-0231, but then when we powered on the calculator
the display was like -0000000000000 and the keyboard was stuck

then we changed the 1818-0232 and when we powered on the calculator
the display went back to 0.00

however, when we tried to key-in the test program the calculator behaved like before,
and even reading the SD-04 card was producing the same weird behavior than before the ICs swap

Any idea on what we can try ?

thanks again and take care !!!

[teenix]

Hi Alberto,

When pressing GTO.112, just after pressing the [2], these ROMs execute...

ROM 00 - 1818-0268
ROM 02 - 1818-0268
ROM 03 - 1818-0268
ROM 04 - 1818-0550
ROM 05 - 1818-0550
ROM 07 - 1818-0550

as well as 1818-0231 and 1818-0232 which you already replaced

These RAM locations are also accessed...
32
48
61

... but you already replaced the ICs for this RAM range, 1818-0231 and 1818-0232

ROM0, 1818-0268, is executing common key press code, so maybe this is ok if you are sure the normal key functions work.

Most of the processing for error checking of the entered value (112) and then jumping to the entered step seems to be executed in ROM 04, 1818-0550, so maybe this is faulty.

As you have 1818-0226 in the donor board, you would have to replace the 1818-0550 with 1818-0226, and 1818-0551 with 1818-0228.

It is hard to say where the fault lies without analyzing the Is and DATA buses.


Edit:
Just looking at 97 service manual and it mentions inconsistent program stepping faults point to ROM 1818-0550 also.


cheers

Tony

[albertofenini]

Hi Tony, Hi everyone

we are happy to say that now it works !!
yesterday we replaced the couple 1818-0231 and 1818-0232
with the donor chips but the behavior was the same

today following Tony's suggestion we replaced the couple 1818-0550 and 1818-0551
with the donor couple 1818-0226 and 1818-0228 and now the board works
fine, can read and execute programs stored even on two sided cards

since we didn't test it after replacing the first of the two (1818-0226) we cannot
say which one of this pair was faulty, but it may be safe to say that
couple of chips we replaced yesterday were good

again thank you very much to everyone for the suggestions received

will try now to fix the other unit that let card slip out of the card reader with a consistent error message ...

take care !

[albertofenini]

Just wanted to add that also the unit that was constantly giving an error message
when reading cards has been fixed

After replacing the CRC, ACT, ROM0 with no success
we fixed it, as suggested by Tony, replacing the 47microF
capacitor in the power supply

We have now only more unit to fix and that’s the one that
doesn’t turn on immediately… we’ll keep posting

Thanks again !!!

[teenix]

Nice work Alberto.

It looks like one or more bits flipped in the ROM causing invalid program reads. I guess this is not surprising considering the age of the devices.

cheers

Tony

[albertofenini]

Thanks to you and everyone’s help we did a good work
Now will try to fix the last one
BTW trying to fix it the display showed the same -000000000000
the we had last night replacing the first couple or ram chips so
a good candidate could be the ram chips
We still have a good couple of the 0231 e 0232 so tomorrow will try
Thanks again !!!!!