ACT available

[MeindertKuipers]

Today I repaired my first ever HP calculator, my good old HP21, with Bernhards ACT replacement, and it works great again!

Meindert

[Mark Hardman]

I sadly wasted my Saturday arguing with bewusstseinsstrom.

But today dawns a bright new day. I've opened and repaired one of my damaged HP-25Cs with Bernhard's Extended ACT replacement. This HP-25C is a wonderful example that obviously had its RAM destroyed early in its life. As a result it shows little use; has a clear, unscratched face plate; and a perfectly functioning keyboard.

I was able to gently lift the existing ACT from its pads and set it aside. I used your suggestion for aligning the socket strips with the chip inserted. As a result I avoided the minor contact issue that Frido reported. I've put the HP-25E through its paces and even played a game of battleship. We done Bernhard and Eric!

I have a second HP-25C in transit that is not as nice cosmetically. I cannot wait to repair this second calculator so that I can compare the results of program level checksums (g GTO 4) between the two units. ;-)

I then tried to put the spare ACT to good use. I have a reasonably good condition HP-29C that exhibits the "all flashing zeroes" symptom. If I read the markings correctly, the spare ACT is P/N 1820-1741 and should be compatible with the HP-29C. Removing the existing ACT from the HP-29C was more difficult than from the HP-25C. In the end, I resorted to clipping a few remaining legs from the bad ACT. The good news is that the HP-29C works well in Run mode. The stack and registers work fine. As soon as I switch to Pgrm mode, the keyboard becomes unresponsive. Switching back to Run doesn't help. Power cycling the calculator restores functionality in Run mode. Still, a partially working HP-29C is better than a completely dead one. If anyone has any suggestions on what might be wrong, I'd appreciate it.

[Mark Hardman]

(02-14-2015 05:00 PM)PANAMATIK Wrote:  Also the HP-27 with damaged ACT/ROM/RAM can be repaired with the new ACT!

Today my engineering mind won over the collectors concerns. I decided to unsolder the NMOS ACT of my beautiful working HP-27.

I will check the HP-27 ROM code and need some time to confirm that every function is working.

I hope that nobody has thrown away a damaged HP-27 in the past. It can be easily repaired now.

Bernhard

Your ACT replacement is even more of a success!

Will you be making Bruce Lee's HP-27 ROM dump available on your downloads page soon? If so, I'll start working on a KLM package for it in Nonpareil.

Warmest regards,

[PANAMATIK]

(02-15-2015 09:12 PM)Mark Hardman Wrote:  I then tried to put the spare ACT to good use. I have a reasonably good condition HP-29C that exhibits the "all flashing zeroes" symptom. If I read the markings correctly, the spare ACT is P/N 1820-1741 and should be compatible with the HP-29C. Removing the existing ACT from the HP-29C was more difficult than from the HP-25C. In the end, I resorted to clipping a few remaining legs from the bad ACT. The good news is that the HP-29C works well in Run mode. The stack and registers work fine. As soon as I switch to Pgrm mode, the keyboard becomes unresponsive. Switching back to Run doesn't help. Power cycling the calculator restores functionality in Run mode. Still, a partially working HP-29C is better than a completely dead one. If anyone has any suggestions on what might be wrong, I'd appreciate it.

Hello Mark

Congratulations! Very good to hear that your HP-25C is working again. And thanks for playing my battleships game!

After your posting I checked an original HP-25C ACT in my HP-29C. But I have the opposite confuguration than you, my HP-25C was working well and the HP-29C has defective RAMS. So I cannot test, whether it is fully functioning with the HP-25 ACT. It is working in RUN mode and you can switch to PRGM mode and it responds to keyboard in both modes, but I cannot do calculations or enter program steps other than R/S 74, because this needs RAM.

It is possible, that the ACTs of HP-25 and HP-29 are not compatible, I dont know and I cannot answer this question until I can prove it in a fully functioning HP-29C. But it is also possible that there is another problem. Try to insert the chip again, perhaps it has only a contact problem.

At least you have two more running calculators now. And possibly one more soon?

Bernhard

[PANAMATIK]

(02-15-2015 09:37 PM)Mark Hardman Wrote:  Your ACT replacement is even more of a success!

Will you be making Bruce Lee's HP-27 ROM dump available on your downloads page soon? If so, I'll start working on a KLM package for it in Nonpareil.

Warmest regards,

Done! It was available for download already five minutes before your request!

Bernhard

[PANAMATIK]

(02-14-2015 05:00 PM)PANAMATIK Wrote:  I will check the HP-27 ROM code and need some time to confirm that every function is working.

The new ACT works completely with the HP-27 ROM code. I checked, that there are no unknown opcodes. I tried all scientific, statistical and financial functions of the manual and all seems fine.

A next step in emulating of HP calculators can be done. I would like to see a photorealistic HP-27 running on MAC, iOS iPhone4,5,6 please with original proportions, and of course the usual Windows and Linux versions. Is this possible?

Bernhard

[Nigel (UK)]

(02-05-2015 09:34 PM)PANAMATIK Wrote:  Hello everybody,

I want to inform you, that today I sent the first ACTs and they can be ordered now officially.

The release is stable and is tested on HP-21, HP-22, HP-25 HP-25C HP-29C , the last not yet working in PRGM mode.

Please send PM or contact via e-mail. You can also wait for some more feedback from other members in one or two weeks.

I hope now, that my design will work on any woodstock calculators of the above models, not just on the ones I tried. I did my best to achieve this, even without available datasheets. I could not test on so many calculators I liked, because they are rare. I could not look easily inside of the original ACT, because this needs even more research and time. I could not make it without having costs, because the small size needed a professional manufacturer. With all the work others have already done before, and expressing here my great respect to the original first "nonpareil" emulator, at least I could have made something that it was thought for. I would be proud, if any of you can also give back a few vintage calculators a new life.

Bernhard

Half-way there! Old ACT chip successfully unsoldered; took much longer than I expected but got there in the end with a solder-sucker and a lot of patience.

Before going any further, I have a question: on page 12 of the guide, the inset on the second picture shows your new ACT with pin 1 labelled. So far as I can see this picture is a mirror image of the actual chip. Can you confirm that when your new ACT is viewed from above with the column of 5 connectors on the left, Pin 1 is the front left-hand pin? I don't want to get it wrong!

Nigel (UK)

[PANAMATIK]

(02-16-2015 10:46 PM)Nigel (UK) Wrote:  Before going any further, I have a question: on page 12 of the guide, the inset on the second picture shows your new ACT with pin 1 labelled. So far as I can see this picture is a mirror image of the actual chip. Can you confirm that when your new ACT is viewed from above with the column of 5 connectors on the left, Pin 1 is the front left-hand pin? I don't want to get it wrong!

Nigel (UK)

You are right, I mirrored the image to show the right orientiation, so treat is as if not mirrored. Insert it as shown in the picture. I will correct the manual to avoid this problem in the future.



Image of ACT inserted correctly into HP-25C. The orientation is the same for all models. Sorry for not having made an image like this before.

Good luck
Bernhard

[Katie Wasserman]

(02-14-2015 05:00 PM)PANAMATIK Wrote:  

(11-20-2014 02:50 AM)Katie Wasserman Wrote:  The HP-27 uses NMOS technology not PMOS like the 21, 22 and 25 67, 97, etc.. Did you design your "chip" to handle both kinds of logic? That would be quite a challenge I think.

Yes, I did! Without knowing, I designed the ACT also for NMOS technology. But now I know it.

Bernhard,

Fantastic! While all my Woodstocks are working fine I'll keep your ACT replacement in mind if that situation changes or I come across one in need of your solution. Very nice work!

[Nigel (UK)]

(02-16-2015 10:54 PM)PANAMATIK Wrote:  

(02-16-2015 10:46 PM)Nigel (UK) Wrote:  Before going any further, I have a question: on page 12 of the guide, the inset on the second picture shows your new ACT with pin 1 labelled. So far as I can see this picture is a mirror image of the actual chip. Can you confirm that when your new ACT is viewed from above with the column of 5 connectors on the left, Pin 1 is the front left-hand pin? I don't want to get it wrong!

Nigel (UK)

You are right, I mirrored the image to show the right orientiation, so treat is as if not mirrored. Insert it as shown in the picture. I will correct the manual to avoid this problem in the future.



Image of ACT inserted correctly into HP-25C. The orientation is the same for all models. Sorry for not having made an image like this before.

Good luck
Bernhard

Thank you for the image. Things aren't working yet, but I still have hope.

Background: my machine is a HP-25 which works perfectly except for anything involving memory - RCL always recalls zero, program steps cannot be stored. When the new ACT is inserted (the socket strips went in very nicely) the display shows all digits lit up with the rightmost digit brighter than the others, cycling through random characters and/or numbers. There is no response to keyboard. The 6V and -12V voltages are correct; there's a 186 kHz clock at pins 16/17, and a rather noisy signal of a few kHz at pin 20.

I doubt that the problem is with the replacement ACT - nothing stressful happened to it during the assembly. Much to my surprise, on re-inserting the original ACT chip the machine is back to how it was before - all operations fine except for memory access. (I am, frankly, astonished that this ACT survived - the final stages of getting it out were rather brutal!)

Any ideas? What sort of fault could leave the original ACT working (to an extent) but cause the replacement to fail in this way? I know it's rather hard to diagnose things like this over the internet but it's worth asking!

Thank you in advance for your help.

Nigel (UK)

[Mark Hardman]

(02-17-2015 06:16 PM)Nigel (UK) Wrote:  Thank you for the image. Things aren't working yet, but I still have hope.

Background: my machine is a HP-25 which works perfectly except for anything involving memory - RCL always recalls zero, program steps cannot be stored. When the new ACT is inserted (the socket strips went in very nicely) the display shows all digits lit up with the rightmost digit brighter than the others, cycling through random characters and/or numbers. There is no response to keyboard. The 6V and -12V voltages are correct; there's a 186 kHz clock at pins 16/17, and a rather noisy signal of a few kHz at pin 20.

I doubt that the problem is with the replacement ACT - nothing stressful happened to it during the assembly. Much to my surprise, on re-inserting the original ACT chip the machine is back to how it was before - all operations fine except for memory access. (I am, frankly, astonished that this ACT survived - the final stages of getting it out were rather brutal!)

Any ideas? What sort of fault could leave the original ACT working (to an extent) but cause the replacement to fail in this way? I know it's rather hard to diagnose things like this over the internet but it's worth asking!

Thank you in advance for your help.

Nigel (UK)

It might be as simple as a contact problem with the new ACT board and the socket strips. See Frido's post #79 above:

(02-14-2015 07:01 AM)Frido Bohn Wrote:  However, as I reassembled the calculator completely, the display either didn't work at all or showed only random signals.

I realized that the new ACT had some play within the socket strip. My workaround was to put some foam plastic on the new ACT, so the circuit board of the keyboard would excert some pressure on the chip when both boards were reassembled.
This strategy works well so far and I am enjoying now the full funcionality of an HP-25E.

To prevent contact issues of the new ACT with the socket I would recommend to install the chip with the socket strips attached onto the board first and solder the socket pins on the keyboard with this kind of in situ technique.

[PANAMATIK]

(02-17-2015 06:16 PM)Nigel (UK) Wrote:  Any ideas? What sort of fault could leave the original ACT working (to an extent) but cause the replacement to fail in this way? I know it's rather hard to diagnose things like this over the internet but it's worth asking!

Thank you in advance for your help.

Nigel (UK)

Hello Nigel,

First I'm very sorry about the problems you have. Each ACT is tested before shipping. I don't know why it is not working in your machine. My offer is, send me your HP-25 calculator, I will analyze and repair it and send it back to you.

But lets try some things before:
1.) Reinserting the ACT into the socket and trying again is easily done and I think you did that already.
2.) Could there be a contact between the ACT Pins and the keyboard?
3.) Could there be some cold solder joints?
4.) Did you try to switch ON without case, just with board and keyboard. Perhaps fixing gives some pressure to the original board.

Im sure we can fix this problem soon. I will write a PM with more details.

Bernhard

[Frido Bohn]

As the solution with some foam pressing the new ACT into its socket was not so satisfying I opened my HP-25 again. I put the new ACT away and desoldered one of the socket strips so I had enough play to place the new ACT easily. With the new ACT in both sockets I soldered the strip on the board again. However, it appeared that the new ACT was still a bit loose in the socket. I checked the conductivity between the pin breakthroughs on the new ACT and the corresponding pins on the other side of the board where the pins of the sockets strips break through. I noticed that there was no conductivity for some pins and so I decided to fix the new ACT in the socket permanently. I applied some solder on the four corners while exerting a bit of pressure on the new ACT. I had to be careful that the corner pins of the new ACT did not move through its board as they were heated. Finally, I managed to have tight joints at all pins with a good conductivity. (I know that not all pins of the new ACT are operational but I cared for them anyway.)
I admit that this procedure is far from professional but I have a working HP-25E now. For the not so skilled electronics geek as I am, the proper placement of the socket strip and the pins of the new ACT resulted in a real challenge. I believe that one has to find the balance of a good fit on the socket strips while preserving some tension to keep the pins in its place and operational. It may be worthwhile to consider if it would be a better solution to solder the new ACT directly on the board.

Cheers
Frido

[PANAMATIK]

(02-18-2015 08:50 AM)Frido Bohn Wrote:  It may be worthwhile to consider if it would be a better solution to solder the new ACT directly on the board.

Hello Frido,

Each socket pin contains a very small spring contact. I assume, that some springs in your socket are now widened, because you reported the socket wasn't properly aligned.
While I didn't had contact problems in five calculators and I removed ACTs and reinserted them many times, and as far as no other contact problems are reported, I very recommend to use the sockets and according to your first proposal to solder the socket strips with ACT inserted for proper aligning.

Bernhard

[Frido Bohn]

I played a bit with my "Hyper HP-25", and as I was checking the program library I noticed some differences in the location of the programs that are due to differences in the German and English versions of the "Applications Programs". Bernhard apparently used the table of contents of the German version as a template to fill the library in the new ACT. In this version the programs are numbered which makes it easier to find the applications while the English version lacks the numbering. However, if you number the programs by hand the order is somewhat different between both language versions. If you are using the English version that is available in the HP Museum Document Set (HP Part No. 00025-90011, Rev. B 7/75) you have to observe this:
"Permutation" and "Combination" are ahead of "Factorial" in the German version and thus the order in the new ACT program library is:
Pr-37 Permutation
Pr-38 Combination
Pr-39 Factorial

For some reason "One Sample Test Statistics for the Mean" was left out in the German version which should have occupied Pr-44 if the English version was used, therefore you will find in the new ACT:
Pr-43 t Statistics for Two Means
Pr-44 Field Angle Traverse

Hope this is useful.
Cheers
Frido

[PANAMATIK]

(02-24-2015 03:30 PM)Frido Bohn Wrote:  Bernhard apparently used the table of contents of the German version as a template to fill the library in the new ACT.

Hello Frido

You are right. And I was aware that only the german book had numbered programs and I strictly used that numbering. But when I compared both books I didn't notice, that the english "Application Programs" has a slightly different order, I just overlooked it. It is easy to add the numbers with a pencil into the printed english edition, which is however more difficult if you only have a .pdf.

Some news: The new ACT will be soon available also for HP-29C. I found an error in the "nonpareil" emulator, which was responsible for an endless loop in GTO LBL GSB LBL instructions. Since today I have a fully programmable 98 steps HP-29C with 30 registers and indirect register addressing, running with the new ACT. I compared about 1 million machine instructions between the original ACTs program flow and the new ACT and they are running synchronuous now. But there is still some work to do before I can release it.

Bernhard

[Mark Hardman]

(02-24-2015 10:18 PM)PANAMATIK Wrote:  Some news: The new ACT will be soon available also for HP-29C. I found an error in the "nonpareil" emulator, which was responsible for an endless loop in GTO LBL GSB LBL instructions. Since today I have a fully programmable 98 steps HP-29C with 30 registers and indirect register addressing, running with the new ACT. I compared about 1 million machine instructions between the original ACTs program flow and the new ACT and they are running synchronuous now. But there is still some work to do before I can release it.

Bernhard

Terrific news!

Would it be possible to post your source patches so that we can create a working HP-29C KLM? Or could you pass along the source changes so that Eric can update the source tar ball?

Thanks!

[N4TVC]

(02-24-2015 10:18 PM)PANAMATIK Wrote:  

(02-24-2015 03:30 PM)Frido Bohn Wrote:  Bernhard apparently used the table of contents of the German version as a template to fill the library in the new ACT.

Hello Frido

You are right. And I was aware that only the german book had numbered programs and I strictly used that numbering. But when I compared both books I didn't notice, that the english "Application Programs" has a slightly different order, I just overlooked it. It is easy to add the numbers with a pencil into the printed english edition, which is however more difficult if you only have a .pdf.

Some news: The new ACT will be soon available also for HP-29C. I found an error in the "nonpareil" emulator, which was responsible for an endless loop in GTO LBL GSB LBL instructions. Since today I have a fully programmable 98 steps HP-29C with 30 registers and indirect register addressing, running with the new ACT. I compared about 1 million machine instructions between the original ACTs program flow and the new ACT and they are running synchronuous now. But there is still some work to do before I can release it.

Bernhard

Thank you Frido and Bernhard. I have numbered my Applications Book with the numbers from the German Book so I am ready when my ACT arrives.

Randy

[PANAMATIK]

(02-25-2015 01:04 AM)Mark Hardman Wrote:  Would it be possible to post your source patches so that we can create a working HP-29C KLM? Or could you pass along the source changes so that Eric can update the source tar ball?

Thanks!

I need some time until the tests are complete, this will be not before next week. Eric knows about my patches.

Bernhard

[brouhaha]

(02-25-2015 04:30 AM)PANAMATIK Wrote:  

(02-25-2015 01:04 AM)Mark Hardman Wrote:  Would it be possible to post your source patches so that we can create a working HP-29C KLM? Or could you pass along the source changes so that Eric can update the source tar ball?

Thanks!

I need some time until the tests are complete, this will be not before next week. Eric knows about my patches.

Bernhard

I have patches that resolve the problem in a different way. I thought they were in the last Nonpareil source release, but perhaps I'm mistaken. I'll have to check.

I don't think either Bernhard's or my patches are a completely accurate model of how the ACT works, but either patch is adequate to solve the label search problem of the 19C/29C/67/97.

The HP engineer that wrote the label search code (likely William Egbert) cleverly took advantage of a boundary condition in the ACT behavior to optimize the loop. I've seen an early ACT specification which documented that there was unexpected behavior when the P register wraps around after an increment or decrement, but didn't give a precise explanation of the actual behavior. The engineer writing the ACT spec probably didn't anticipate that the actual behavior would be useful, and might not have even known exactly what that behavior was. After the behavior was found useful in the label search loop, it was probably documented in a later revision of the ACT spec, since the microcode became dependent upon it.