Wanted working HP-67 board

[YoungerThanThat]

Does anybody have a working HP-67 board that they are willing to part with?
I am more than willing to pay for it. I have been trying to get my HP-67 to work for a year and I am still stuck.

Thanks

Michel

[andylithia]

Ha ha,I’ve been looking for a HP67 without the cpu board in any condition. Please let me know if you find an appropriate listing. I’m more than willing to take it apart and separate the CPU board for you.

[ThomasF]

(04-21-2022 12:32 AM)YoungerThanThat Wrote:  Does anybody have a working HP-67 board that they are willing to part with?
I am more than willing to pay for it. I have been trying to get my HP-67 to work for a year and I am still stuck.

Thanks

Michel

Hi Michel!

Have you seen Tony's page? https://www.teenix.org/
He has complete CPU-board replacements (which even improves the calculator with e.g. connectivity and continuous memory (a HP-67CX you could say ... )).

Cheers,
Thomas

[YoungerThanThat]

In fact I have Teenix board in my HP-67. It did not work the first time and the calculator did a very lengthy (timewise) trip to Australia and back to be updated and it still has issues. I am going to try to do an update myself when I get an FTDI board but I thought I could also explore the simpler option of finding an original board that is still functional. My HP-67 is truly cursed. It has done exactly 5 trips for various repairs and it still does not work.

[teenix]

Yes, it has been a terrible time for Michel.

The original board went faulty on returning from repairs from someone and when I got a chance to inspect it, a ROM had gone faulty for some reason. Why, I don't know - age, faulty board handling (static) ??.

My board had a software fault with the key scanning that I wasn't aware of and I repaired this at the same time and sent the board and calculator back after testing. Unfortunately not long after posting it back I received a report that the new code had a different problem with the key scanning.

I then removed all the key scan code and re-wrote it from scratch. It now scans the keyboard and if any change is detected in the state of the keys, a debounce counter is started. If the keys bounce the counter is continually reset. The keys state will not be passed to the calculator code until this counter finally decrements to zero and therefore everything should be stable. This seems to work ok now. I had the keyscan code send the keys to a PC where I could see every value returned and they appeared correct after many many key presses.

In some situations, there is no way to correct for a faulty keyboard, as some old key contacts might be dirty or worn so much that they cause problems like double keying etc. Trying to null this out would require excessive debounce delays and would be noticeable to the user.

Just for info, each key column is tested in a single HP instruction time (320uS) window which is due to the way each digit in the LED display is updated from the Cathode Driver. There are 8 LED segments to refresh during the 320uS and during this time, a row of up to 5 keys can be tested. The CPU board tests the keys towards the end of this time when the cathode driver output should be stable. Only 8 of the 14 cathode driver outputs are used for the keyboard, the others are ignored. When a keypress is detected, further testing is ignored until the full display has had a refresh which takes about 4.5mS. (320uS x 14)

Some astute reader might recall that there are 15 digits in the display, so why only 14 outputs on the cathode driver. The reason is that the Mantissa and Exponent signs use two LED segments from a single digit. Seg E is the Mantissa sign and Seg G is the exponent sign. The rest of the segments in each of these digits are not used at all for normal operation. The segments are there though because the display is made up of common 5 digit modules. If you look at the single 97 display module, you will see only 'g' segments in these digits, the rest are not physically there. The Spice display module also only has a single 'g' segment for the Mantissa sign.

cheers

Tony

[Etienne Victoria]

PM sent.

Etienne

[YoungerThanThat]

I want to profusely thank Tony (Teenix) for all the hard work he did on my calculator to try to get it right. There might still be bugs in his code (or a faulty keyboard on my HP-67) but I have been designing computers (hardware) for a living so I perfectly know what it takes to get things right. I want to also commend Tony for having engaged in such a huge project of developing a new board and the software from scratch. This is pretty impressive and I am lucky to be able to benefit from it.

Last year I found my old TI-58 at my parents' place in France and surprisingly it still works perfectly. I remembered that when I was in college only the rich kids had HP calculators and the rest of us had TI (30 or 58) or 4 operations basic ones (we had to use slide rules and log tables during exams at that time in France) so, I decided it was time for me to get into HP calculators. I first got an HP-21 to learn RPN and then an HP-25. The HP25 is so quirky to program that it is fun. Then I moved to an HP19C which I use every day. It is very easy to use as I can find bugs immediately by printing the program.

I have been dying to move to the HP67 and I recently bought a couple of PACs and started to write some programs. However, the one I bought seems to be cursed. I will try to reflash this coming week if I can find the time.

On a final note, even If I do not post, I thoroughly enjoy this forum and all the regular contributors. Many thanks for all the insights!