ACT available

[Dreato]

ACT 25E/IR printing problems:

Recently I updated two HP25 calculatiors to the 25E ACT. No problems here - everything worked great! After this, I added the IR circuit to both ACT boards, and installed the IR LEDs in these 2 calculators. One prints perfectly (see photo), and one gives occasional errors. I swapped the ACTs between these 2 calculators, and the problem followed the ACTs.

   

I though that I may have made a mistake soldering the SMD parts on the ACT, possible shorting the 33 ohm resistor? So I removed them and soldered new parts on again, but still have the problem. I also reflashed the 25E v 1.14 firmware, to see of that helped. It did not - same problem. Any suggestions from the helpful readers of this forum?

Thanks and Best Regards,

Dave

[PANAMATIK]

Did you try to replace the NPN transistor? Perhaps you have chosen a unsuitable type? I recommend BC847 or BC817 with high current gain. The firmware follows the HP82240 timing very accurately from the early versions and is not responsible for the problem.

Bernhard

[Dreato]

Yes - I used the BC847 transistor, the Vishay TSAL4400 LED, and the 2.2k ohm and 33 ohm SMD resistors. It's very puzzling! Identical components in both working and non working IR circuits. All other functions are fine. I flashed the ACT to HP29 firmware, which also worked perfectly, but which still showed the random printing error. Scope traces looked OK.

Scratching my head!?

[Dreato]

I've got a follow up on my IR printing problem. I hooked my scope up to the IR LED, and compared the waveforms on the working and non working calculators. The waveforms were very similar. Since the printing errors were reduced when the calculator was farther from the printer, or when the calc was off axis, I thought maybe the IR receiver in the printer was set to a very high sensitivity. I opened up the printer, but there was no sensitivity adjustment. However, when I covered the IR receiver photo transistor with my finger, allowing almost no IR light in, the printing was perfect.

The problem is in my printer. The IR receiver is VERY sensitive, and there's no adjustment that I can see. I added some attenuating material behind the printer receiver window, and all is well now.

Regards,

Dave

[Dreato]

Final update on my printer issues. I went down the rabbit hole: After narrowing my printing issues to the printer itself, I decided to look at the schematic (attached - from Kees Vandersanden's site). I traced the IR signal through the U5A flip flop. From there, it goes to the M80C50 processor where decoding occurs. This processor pinout shows a crystal oscillator at pins 2 and 3. However, this printer uses a cheaper LC oscillator. I decided to measure this frequency with my old HP frequency counter. When I attached the probe, the printer would print the output from the suspect ACT, error free. The frequency counter presents a load of 1 Meg ohm and 10 pf. I soldered a 10 pf capacitor across L3, and the frequency changed enough to assure error free operation with both my 25E ACT calculators.


  HP82240B_Schematic.pdf (Size: 40.15 KB / Downloads: 38)

[PANAMATIK]

Bravo for tracking down the problem to the high sensitivity of the printer receiver. You could replace the 33 Ohm resistor of the ACT circuit by 100 Ohm, this would attentuate the IR stream and would also solve the issue.

Bernhard

[Dreato]

RE: 100 ohm resistor. I actually tried that, with no effect! This is what led me to open up the printer. Looking at the signals at flip flop U5A- they all looked the same from both my HP25E ACTs - attenuation made no difference in the appearance of the signals. I was actually attenuating the signals by placing my finger over the IR receiver diode. This led me to look for bad solder joints, and the M80C50 chip which decodes the digital signals to driver the printer module. The spec for the printer specifies the pulses at 32,768 Hz rate (you know this from the HP82240 interface guide), but that decoder chip had a simple LC oscillator. My experience working on the elusive HP45 hidden stopwatch feature (see my website https://davidreaton.com/hp-calculators/hp-45/ ) told me that LC oscillator frequency can vary widely depending on component values. All the 'Classic" series use a simple LC oscillator to multiplex the display LEDs. I played with swapping capacitors into this circuit to adjust the frequency, with limited success, until I found the MEMS oscillator, which gave the required 784 kHz frequency with only 20ppm drift.

Therefore I measured the frequency at the M80C50 oscillator pins. When I clipped the frequency meter on, the printer began printing flawlessly. I assumed that the RC load from the frequency counter input changed the LC oscillator resonant frequency enough to bring it into spec. I added a 10 pF ceramic cap to this circuit, and now it works well. BTW, the schematics shown for the HP82240A and B models (from Kees Vandersanden's site) show C values of 47pF and 390pF, respectively. I couldn't see the caps values in my printer, so I don't know which one is correct.

To be perfectly honest, I don't know if the LC oscillator modification is the true, definitive answer or not. I know the IR attenuation observation I made, DID WORK, but I don't know why!?? Maybe putting my finger over the IR receiver diode added some stray capacitance? I'm a chemist, not an electrical engineer!

Clipping the frequency meter on made the difference. I guess it's an example of the Heisenberg uncertainty principle. The act of measurement changes the very thing you're trying to measure!

Sorry for the long winded post. It was a long, strange trip, indeed! Thanks for all feedback, help and suggestions.

Regards,

Dave

[calc-calcs]

Hello all,

I am a new member of this forum and have been in contact with Bernhard for a while due to the search of an HP25 in need of a new ACT to try his amazing creation. I actually found a 25 in very nice condition complete with battery holder and very cheap, and after cleaning up a few corroded spots it turned out to only have a problem related to the RAM chip which seems to be non working, even though by just briefly heating up one or more pins of the chip with a soldering iron temporarily brings it back to completely normal operation.

When I first opened the calculator I found that besides the battery contacts, there was corrosion on all the pins of the RAM chip that face the battery connector side but it cleaned up quite well with some vinegar and IPA, and to a lesser degree there where smaller corrosion spots on a few other places on the board that also had some of the typical bluish corrosion. I am guessing that maybe the corrosion managed to get into the chip itself and that is why it failed. But I still don't understand why it starts working again when heated up.

I am now in the process of ordering an Extended ACT from Berhard to get this HP25 back to fully working condition and then some with all the nice additional functionality included in the extended kit.

Thank you Bernard for such a nice resource to revive so many otherwise useless HP calculators, and for your always very helpful email replies.

Alex

[mpark]

(05-21-2019 03:16 AM)calc-calcs Wrote:  ...it turned out to only have a problem related to the RAM chip which seems to be non working, even though by just briefly heating up one or more pins of the chip with a soldering iron temporarily brings it back to completely normal operation...

Hi Alex,

While waiting for your ACT to arrive, you might try the 20k resistor fix described here: https://www.hpmuseum.org/cgi-sys/cgiwrap...i?read=267


Michael

[calc-calcs]

Hello Michael,

I did try the suggestion of the 20K resistor between the data pin 11 and ground pin 22 of the ACT but it made no difference. Since the RAM chip is doing no good in there anyway I am going to pull it off the board and see if under it there might still be some residual corrosion between pins that might be causing the problem. If its not the case at the very least I will be sure its all clean under there.

Extended ACT, update kit plus 3 overlays to play with (29C/34C/67) ordered yesterday from Panamakit.

Thanks.
Alex

[calc-calcs]

Well the RAM chip is off the board, and yes there was a bit of remaining corrosion between 3 of the 4 traces that run under the chip. Extraction went quite straight forward although it was a bit tricky to clear all the vias because most IC pins are leaning on the inside wall of the via plating itself, so most of them needed to be centered in the hole first. It was a good experience in preparation for the ACT swap.

The RAM chip seemed clean underneath and between its pins, but I gave it a good cleaning with a soft toothbrush and 91% IPA, as a matter of course.

Without the RAM chip the calculator is still functioning the same way as before. I am tempted to solder in a socket and insert the RAM chip to try again and see if anything changed. But don't have any low profile sockets at hand. Oh well guess I could always unsolder the socket if needed.

I have some pictures of the restoration process since I first opened the calculator for the first time, but guess there nothing most of you have already seen or done yourselves.

If I decide to reinsert the RAM chip with a socket I will report back findings.

Alex

[calc-calcs]

Guys, stop the press!

The HP25 is fully working again, and it has been for the past two and a half hours since I reinserted the RAM chip with a low profile socket. Full power down between tests, 30 or 60 minutes or longer cool down time, and it still works. All is good, registers, statistical E+ operation, programming.

Please visit this other thread bellow for added details, want to respect Panamatik's own thread and avoid posting stuff not directly related to his ACT kit.

Post #11 on this page:
https://www.hpmuseum.org/forum/thread-13...#pid116842

Alex

[PANAMATIK]

More than 200 Woodstock calculators repaired!

After the introduction of the new ACT repair kit the serial numbers have reached 200!

My great thanks to all of you, who have brought a HP Woodstock calculator to life again.

For this great success I will make a special offer: to repair the next 10 calculators for those, who are not familiar with soldering. Just send your calculator to me and I will repair it for the costs of the DIY kit + shipping.

Please send PM if you are interested.

Bernhard

[Gene]

This makes me want to fry an ACT chip on a woodstock to send it in for repair. :-)

[PANAMATIK]

(10-19-2019 03:36 PM)Gene Wrote:  This makes me want to fry an ACT chip on a woodstock to send it in for repair. :-)

Calculators which are intentionally fried, burned, minced or cooked will be explicitly not repaired. :-)

Bernhard

[BobVA]

Very glad to hear that! Your huge investment in ingenuity and time has been a great gift to our community.

[PANAMATIK]

(10-19-2019 01:47 PM)PANAMATIK Wrote:  More than 200 Woodstock calculators repaired!

For this great success I will make a special offer ...

The details:

Repair for any Woodstock calculator HP-21 HP-22 HP-25 HP-25C HP-27 HP-29C possible, not HP-19C, HP-97 and other models

Two options available
1.) HP-25A 50,- Euro no extras
2.) HP-25E 80,- Euro many extras
shipping 13,- Euro within Europe, 20,- Euro US and world

Please send PM, that you want participate, then I will confirm that you are in the queue, then send the calculator to my address (see my website).

9 offers left, no time limit

Bernhard

[Harald]

Wow! 200 is quite a lot. I wouldn't have thought there is such a big market for these.
I think I sold about 40 of the classic upgrade boards. But the classics usually don't have bad main boards. So it really is an upgrade rather than a repair.

Cheers,
Harald

[Ren]

Bernard,
I tried contacting through your website.
Is the ACT replacement still available?

[edit] PS. it is for a HP-22

[Ren]

Folks,
Bernhard responded to my inquiry on his website.
I'm seriously thinking of making the purchase.

Good Night!