HP-41C flakiness



#23

I recently acquired an HP-41C (serial number indicates 1980 manufacture) that is exhibiting some decidedly flaky behavior. First of all, it freezes up on any non-arithmetical mathematical keypress: the screen goes blank, and the calculator is completely unresponsive until the batteries are removed and replaced. At this point the screen remains blank but a subsequent numerical keypress will seemingly bring the calculator back to life, and will be displayed even without the calculator first being powered on.

More strangely, I can induce the calculator to freeze by keying in an arithmetic expression that evaluates to 10 or greater, although expressions that evaluate to less than 10 are processed just fine. In particular, if I key in an arithmetic expression involving addition, subtraction, and multiplication -- and evaluating to less than 10 -- then the calculator behaves normally. But as soon as I evaluate an expression that reaches double digits, the calculator goes into its zombie mode.

Finally, keying in a division leads to something else entirely: the calculator displays "out of range" regardless of the dividend and divisor. The calculator is still responsive after providing this message.

Any ideas what might be going on here, and whether it might be correctable? Any thoughts appreciated.


#24

Sounds like it needs a good clean (inside) and/or has cracked or broken screw posts. The latter is documented well in the this forum. Also cleaning involves disassembling the calculator, and using a good circuit board cleaner to clean the contacts between IO block and keyboard, and CPU board and keyboard including connector strips.

Cheers,
John


#25

Thanks! I'll give that a try.

#26

Hi John, what's about the battery contacts inside the calculator have you seen any trace od corrosion or leakage?
About the suspect of a "broken posts" issue, I'm agree with John, but I suggest, before to open the calculator (and removing the four screws, with the risk to create a new problem if not "the problem"), simply to tight a little more the screw, starting with the two on the bottom, closer to the zebra connector contacts and looking to see if something changes in that flaky behavior.


#27

Good call on tightening the screws, definitely worth a shot !

#28

The battery contacts look absolutely clean, and there is no residue anywhere in the battery compartment. Thanks for the suggestion of tightening the screws; I will try that before I disassemble anything.

Thanks to both of you for these suggestions! This is my first HP-41 so I wanted to ask the experts before trying anything. I will search the forum archives and read up on the posts, and then proceed.

#29

The same thing could be checked just pressing it together without tightening the screws, right? (so no harm would be done by over-tightening for example).


#30

I agree with Garth. I have taken apart many HP-41s, and not a single one had loose screws. If the case was loose, it was always crack/broken posts or the screw wells in the back case were broken out. If there the screws are not loose, attempting to tighten them further could create new problems.


#31

Quote:
I agree with Garth. I have taken apart many HP-41s, and not a single one had loose screws. If the case was loose, it was always crack/broken posts or the screw wells in the back case were broken out. If there the screws are not loose, attempting to tighten them further could create new problems.

Believe me Dan, I never said that you must force while attempting, you break through a opened door!
I have not a big experience with the repair of this kind of issues, but I tried differents methods: the Geoff's one and others.... always with no luck, so that in my mind it's better to keep away from any cracked post =:)
I got months ago a problem of unexpected auto shut-down and (following the experience of other members) I solved it, at the beginning, just tightening a little the screws on the bottom.
Then when the problem came back, I opened the calculator looking fore the zebra contacts and then started the real problem!
At that time all the four posts were still safe but, opening today, tomorrow and the day after..... was the best way to get the broken posts!

Pressing the bottom, can show (if John is lucky) the way to follow and tightening a little could be the first attempt.

Cheers from Italy


Edited: 9 Apr 2013, 4:55 p.m.

#32

Update: I tried applying pressure to different parts of the calculator, with no change in behavior. So, I carefully disassembled it. The screw posts are all intact, and the inside looked pretty clean. I gave all contact points a good scrubbing with alcohol and carefully reassembled. The only thing that has changed is that now when I press the division key, I get a freeze instead of an "out of range" message.

It seems like this problem is more deep-seated than an intermittent or incomplete contact...I'll take it apart again and scrutinize things even more carefully, but I'm not very hopeful that there is going to be an easy fix.


#33

In the C, it is most likely a bad rom :-(

#34

Replacing old capacitors and re-soldering contacts near CPU at main PCB can fix this. Try diode testing too - it's simple. You can try to re-solder connectors between LCD block and keypad PCB (I mean contacts on the keypad PCB) This can help too (as this is the second major weak point of the fullnut 41's) It's better to re-soldering ALL LCD pin contacts at keyboard PCB, as visually they looks good, but sometimes has large resistance inside. And (more simple) sometimes thin dielectric washer over right PCB post can help too (as it correct PCB pressure to the right direction).


#35

Wow, thanks for the detailed recommendation. Maybe I'll give this a try. I haven't soldered anything in a long time, but this would be good practice for the WP 34s I am going to build....


#36

What a coincidence! I ordered HP30b for building 34s too )) As for 41's





here I made a photo with washer as a reference.


#37

I see you are one to not follow your own advice to replace capacitors... as the original HP hermetic tantalums are clearly shown in the photo... as well as all other capacitors being original.

Further, I suspect your washer "fix" is really just compensating for stripped posts caused by the short screws used by HP in the lower holes in the original C and CV machines. Early 41's had 0.25 inch long screws which all too easily stripped the posts. They were later replaced with 0.375 inch screws which solved the problem.

The other possible problem is the zebra's themselves. If your machine has the early "pink" color elastomers, they are the source of the problem as they always developed high resistance connections once the board was removed.

Oh, I see the diodes are original too ;-)

Edited: 15 Apr 2013, 5:54 p.m.


#38

This is from other machine. Washer is the simplest fix, so I'm apply this, usually, first :)

#39

Quote:
Replacing old capacitors and re-soldering contacts near CPU at main PCB can fix this. Try diode testing too - it's simple.
No, it will not. Did you read the original post? Based on the reported problem, that's a complete waste of time and will more likely just introduce more problems.

Quote:
You can try to re-solder connectors between LCD block and keypad PCB (I mean contacts on the keypad PCB) This can help too (as this is the second major weak point of the fullnut 41's)

Agree in principle, but just test them first by pulling upward on the center area of each connection with a small screwdriver or pointed tweezers. If they lift up, they're fractured. Resolder those. Leave the others alone.

Quote:
It's better to re-soldering ALL LCD pin contacts at keyboard PCB, as visually they looks good, but sometimes has large resistance inside.

Nonsense. If they are mechanically sound, they are electrically sound. Resoldering everything leaves flux behind that should be removed as it may cause leakage current, especially water soluble fluxes. Done improperly and you'll contaminate the switch domes of the ON USER PRGM ALPHA keys which will then be intermittent or completely dead.

Quote:
And (more simple) sometimes thin dielectric washer over right PCB post can help too (as it correct PCB pressure to the right direction).

If it did not have washers, it does not need washers. There were four generations of HP41 logic boards. Only the last board needed the washers. Original C and CV boards were 0.063 (1/16) inches thick. The last 41 logic board design was used on all variants (C/CV/CX) and they were only 0.050 inches thick. It was *only* these boards that HP installed 0.015 inch thick teflon washers to increase the board thickness to meet the design dimensions of the case front and back so it would connect properly.


#40

Sometimes old capacitors has notable DC leakage. In that case, in simple words, they "stolen" power from the main circuit. Side effect of this failure is relatively fast battery discharge and CPU glitches. But anyway - if electrolytic capacitors are more, than 30 years old, all of them needs replacing.

Quote:
but just test them first by pulling upward on the center area of each connection with a small screwdriver or pointed tweezers.

Sometimes this do not indicate failure. Multimeter, generally, better :) From my real experience - "When I press "ALPHA" key - display goes blank from the right side for a tiny period" All pins visually was ok, but inside there was some numbers of so-called "cold soldering". Of course, you must do this with caution! Re-soldering contacts on LCD PCB, for example, are a little more tricky, as you must constantly control temperature and soldering time, use compatible 63-37 Sn-Pb solder and non-agressive flux paste + professional instrument...

Quote:
if they are mechanically sound, they are electrically sound

No, no, no... 10 times. See above :)

[quote]
If it did not have washers, it does not need washers.
[/quore]
Obviously! New units do not needed that. But zebra contacts become to be unreliable after some time of constant pressure. Washer (only one, at the right!) sometimes needed for compensating this.

That all based on my real experience, of course.


#41

Quote:
Sometimes old capacitors has notable DC leakage. In that case, in simple words, they "stolen" power from the main circuit. Side effect of this failure is relatively fast battery discharge and CPU glitches. But anyway - if electrolytic capacitors are more, than 30 years old, all of them needs replacing.

While electrolytic capacitors do dry out over time, especially when subjected to high ripple currents, this is not the case with the capacitors used in the HP41. Further, the capacitors used in the original 41C and CV's designs were hermetically sealed tantalum devices. In my years of personally repairing many HP41's, I have yet to see a leaky or outright defective hermetic tantalum.

Quote:
"When I press "ALPHA" key - display goes blank from the right side for a tiny period"

That is not a soldering problem, it is one of mechanical deformation of the LCD. If the display is mounted too low (and this is easy to do), pressing of the PRGM or ALPHA key will deform the pc board which then places pressure on the display driver and LCD directly causing the display anomaly. The proper mounting procedure is mentioned in the service manual which requires the placement of the display as high as possible to avoid the mechanical interference. The same thing can occur as well on the left side with the ON and USER keys.

This problem was mis-understood early in the 41's development and it was thought it was just keyboard deflection. Early 41C and CV case frames had extra heat stakes along the top row of keys to prevent this anomaly. HP later decided that is was not a problem, it did not affect usability and the extra stakes impacted assembly and soldering of the display connections so the stakes were eliminated. HP also later changed the design of the LCD/driver clamps which helped with this issue.

Quote:
That all based on my real experience, of course.
Mine as well. I've personally repaired thousands of HP calculators so my postings are real world experiences, not a one off, got lucky repair or second hand something I read on the web.

IMO, wholesale replacement of components as a troubleshooting method is not a method at all, it will typically introduce more trouble, especially when done by non-experienced persons with less than ideal tools, components and supplies.


#42

Ordinary electrolytic capacitors tend to loose capacitance. But tantalum's sometimes acts strange... I have a couple of those capacitors (from HP65 and TI59), having leakage under load (one even has notable heating). Tantalum C's are much more reliable, so I do not replace them, if they show right parameters after testing. But testing, generally, mean un-soldering.
If you do not have the right equipment - it's much cheaper to replace some parts for sure, as 41's do not allow many tries for assembling-reassembling. It's better to do all prophylactic works at a single try. Or give it to professionals with a right equipment :)

Quote:
That is not a soldering problem



May be you are right, as for typical cases. But re-soldering helps at my situation a lot - now display act flawlessly.

Quote:
IMO, wholesale replacement of components as a troubleshooting method is not a method at all, it will typically introduce more trouble, especially when done by non-experienced persons with less than ideal tools, components and supplies.



Yes, I think so too... But from other side what I, for example, would lose, if I replace one-two cheap capacitors additionally? All other components (CPU, RAM, LCD etc...) are too rare and costs much more for "guessing". Those "hard cases" repairing are too expensive for most 41's users, so it's better (and cheaper at most cases) to find working 41 at eBay(TM) and left non-working machine for spare parts. Or build custom 41CL :)

Edited: 15 Apr 2013, 8:37 p.m.

#43

Quote:
Resoldering everything leaves flux behind that should be removed as it may cause leakage current, especially water-soluble fluxes.
Water-soluble flux should never be used-- for anything! Our experience with it in our company has been absolutely terrible every time it was tried, no matter how careful someone was to try to clean it off. It practically made boards unsalvageable. On our sensitive analog boards, clean as they may look, the water-soluble flux caused leakage that upset bias points and also caused a frying-bacon sound in the audio. Rosin flux OTOH never causes leakage. It is basically a perfect insulator. I understand there's a high layer of atmosphere that will react with rosin-core solder and corrode things if you go into space, but our products stay below 40,000 feet, and rosin flux has never been a problem. The flux is of course cleaned off after automated soldering, but sometimes there are wires or non-washable parts to attach later by hand using rosin-core solder, and the flux does not get removed after that. It's no problem.
Quote:
All pins visually was ok, but inside there was some numbers of so-called "cold soldering".
If the solder joint was good from the factory, it cannot somehow turn "cold" over the years. The only problem would be if it cracks from stress, which, along with peeling off foils, is why we do not use SMT connectors on our products that are subjected to vibration or in some cases even dropping. The solder joints on thru-hole connectors in plated-thru holes are pretty much immune to this cracking, and have proven to be able to continually handle a lot of force.

#44

Quote:
Water-soluble flux should never be used-- for anything!

Definitely! In general, I've try to clean even non-agressive paste flux for sure too.

Quote:
If the solder joint was good from the factory, it cannot somehow turn "cold" over the years

Yes, usually, this is factory fault. Those pins can oxidize inside, if they do not prepared for soldering properly.

Forum Jump: