HP-55 transistor info needed



#6

I'm trying to replace a damaged transistor on an HP-55 logic board. The transistor's legs were crushed, and its emitter no longer appears connected. Unfortunately, it is a non-standard type and although I tried to use some reasonable replacements, they didn't appear to work.

The transistor appears to be an NPN bipolar type, with a B-C diode forward bias voltage of 660mV. It has the following markings:

M641 4-668 L

The M appears to be similar to the Motorola M, except that it is not circled.

Can anybody tell me what would be a suitable replacement for this transistor, something commonly available?

Thanks in advance,

Viktor


#7

With only a single PN junction, how have you determined the possible type of transistor (circuit tracing, marks on board, etc)? (just curious).

Do you have any idea what function this transistor performs?

What have you tried replacing it with? (I would have tried a BC547 or something like that for my first try)


#8

Steve,

Thanks for your reply and your trying to help.

The transistor in question is part of the unit's DC-DC converter. When I received the calculator, this transistor's legs were crushed (it was located right underneath a screw post) and shorted; the calculator itself was dead, with only a few segments from a single digit lighting up faintly.

I tried several replacements (basically, obvious choices that I had in my toolbox), for instance, a 2N2222. With this replacement, the proper voltages seem to appear. Unfortunately, I still only have those few segments lit on the display, although the calculator now responds to keystrokes (the segments flicker.) Moreover, as I key in digits, the segments become brighter; I can even write a program consisting of ten 8's and watch it run, brightening up those few segments, so the calculator's logic board appears to function.

Before you say that I have a broken display driver chip, I don't. With great difficulty, I managed to actually replace both driver chips with ones from a really dead unit, and nothing changed. I'm not going to desolder one of those chips again .

The unit is otherwise clean with no signs of damage or corrosion. This is why, before proceeding any further (and possibly doing more harm than good through another misdiagnosis) I want to make sure that I am using a suitable transistor here.

Thanks again for your help.

Viktor


#9

Since it's part of the DC-DC converter, I can't see it being particularly critical. If, as you say, the voltages from the DC-DC converter are correct, then I'd suspect that the problem is elsewhere.

I can't see that undervoltage is likely to damage the calculator, and I also can't see that crushing the transistor could have damaged anything other than the battery.

The segments that display... Are they the same segment in each digit? I would probably try manually turning on the segments by using a probe with a 1K resistor in series with it to try to get other segments to light.

I'm assuming (from your description) that the 55 has cpu + drivers + display.

Depending on the way that the display is strobed, you should be able to light up all of the like segments in the display. This would be a good check of the display. Obviously you've got to know (or be able to figure out) where the appropriate conections are, but it _shouldn't_ damage anything. That way you can check the display.

If the display is OK and the drivers are OK, then it only leaves one thing...

I've noticed with older devices that soldered joints can fail in a way that is almost invisible.

You may be well advised to go over the entire board, gently levering all connections to the board with (say) the lead of a resistor. I do this while watching the connection carefully with a loupe. Alternatively you could resolder every connection...

But you've probably tried all this...

good luck.


#10

For what it's worth (in case anybody still follows this thread)...

The actual cause of the problem turned out to be a failed diode. I presume this diode died as a consequence of the crushed transistor in the DC-DC converter, which caused invalid voltages to appear all over the place. I'm glad it was a diode that gave up its life protecting the trickier bits because a plain old 1N914 works fine in its place. Identifying the cause and locating the faulty bit was a dreadful exercise but in the end, I was rewarded with a working HP-55 calculator!

Viktor


#11

Congratulations on fixing your calculator! It's been a long time since I got down to the level of soldering individual diodes and transistors. These days, "repairing" electronic devices usually is a matter of swapping IC chips or entire circuit boards. I'm glad there are still people like you, with the skill to diagnose and fix problems at the component level.

#12

Thanks for the update. I'm glad it was something simple.

I know how satisfying it is to fix something like this.

#13

Congretulations on fixing it. I'd think it would be of interest anyway and if somebody knows the identity of some unknown discrete parts it would always be appreciated to hear anything about that. If I have one of the calcs open, I might desolder and measure the Transistors.

In a switched mode power supply, the speed should be more of interest than beta, because modern small signal or switching transistors have better overall ratings (UCEmax , ICmax, ...) anyway.

SMPS drive the transistor into saturation, so that the amplification shouldn't be really important.

Saturation voltage should also be of interest for finding a modern replacement.


#14

That's interesting, because I've found some quite old transistors I'd pulled out of a circuit have much lower (and more consistent with changing I(c)) VCE(sat).

But this is probably no use to anyone :-)


#15

Did you mesure the Vbe? Could be Germanium. Or very old technology with a smaller rce.


#16

Yes, Vbe indicates silicon.

They are from an era long past. Fitted to the board with long (individually spaghetti-ed) leads.

They came from a logic board of some sort, and I asumed the low Vce(sat) was there to give the widest voltage swings and thus highest noise immunity.

#17

Which diode was bad? I've got a dead 55 too. Maybe I'll pick it up and do some probing.


#18

Hmmm, I'm not exactly sure how I'd identify the faulty diode without opening my HP-55 again, which I have no desire of doing anytime soon! Anyway, if you just test all the diodes in-circuit with a DMM, you should be able to find any that are shorted or open. Also important were the symptoms: a few segments were always faintly lit on the calculator's display, they responded to keystrokes with a flicker, brightened when I entered all 8's, and did the expected flicker when I keyed in a program and ran it. This is what told me that the calculator itself was functioning, and the problem was with the display driver.

Of course, before I got to this point, I had to replace that bad transistor in the DC-DC converter. Without that replacement, the calculator was completely non-functional; a few segments did appear lit on the display, but there was no response to keystrokes or anything.

Viktor


Possibly Related Threads…
Thread Author Replies Views Last Post
  HP 50g info Jan 7 2,685 10-21-2013, 09:09 PM
Last Post: Mike Morrow
  Prime 'CONTENT' section not retaining info text Geoff Quickfall 2 1,381 10-17-2013, 05:24 PM
Last Post: Geoff Quickfall
  HP 55 Olivier De Smet 8 2,742 08-26-2013, 04:54 PM
Last Post: Matt Agajanian
  [OT?]Orthography corrector needed. Diego Diaz 8 2,701 06-22-2013, 04:31 PM
Last Post: Andrew Nikitin
  Prime Info DavidShenk 0 945 06-14-2013, 11:39 AM
Last Post: davidshenk
  Update of Emu48 to v1.55 Christoph Giesselink 4 1,972 05-26-2013, 05:28 PM
Last Post: Andrew Nikitin
  HP-50g How to store many equations - help needed Timo Labrenz 5 2,122 03-03-2013, 12:22 PM
Last Post: Timo Labrenz
  Disaster: help for a couple of HP97 needed! Jose Gonzalez Divasson 10 3,062 01-21-2013, 04:45 PM
Last Post: aurelio
  HP41 emulator for HP48sx - keyboard overlay needed Doug (NYC) 3 1,658 12-31-2012, 03:11 PM
Last Post: Allen
  restoring HP29C - help needed Alberto Fenini 9 2,596 11-25-2012, 07:56 AM
Last Post: Alberto Fenini

Forum Jump: