Classic repair - ghosting digits - What's your experiences?



#2

Hi
Lately I've bought 3 classic 2x 35 and 1x 80. Including the one in my junk box all exhibit some kind of ghosting.

The junk box one: Slight enlightening of the middle segment of all digits. Not severe but made difference between 0 and 8 not as clear as it could be. Corrected it by "shortening" one of the coils with a resistor. Found a value which caused no visible fade but also made sure it was dark when supposed to. Shortening with a resistor of the driver coil of the corresponding segment doesn't seem to change power draw any significant value.

The 80. Ghosting of one specific digit.
A 3rd edition 35. One segment dead this is also pretty worn and UV exposed so perfect for spares.
A 4th edition 35 (the one with molded key caps). Heavy ghosting on one digit.

I read somewhere some one else "corrected" such issue by loading the cathode of a bleeding digit. While that works for the 80 and with a low value (high current drain) also for the 35 4th edition that raise power consumption between 10 and 30-40 mA depending on needed value.

Since I already have swapped around the cathode driver of the 4th edition with another from one of the worn ones still no change i also tried swapping the anode driver. Surprise!! Still no change. Kept the original anode driver as I just relied on pressure to try the swap out. That leaves it up to the displays themselves. Took a module from the one with the dead segment. Presto! The 4th edition works perfectly - without any pulling resistors. Yeah the 4th had I marked (luminosity) and the modules in the one with dead segment was H marked. No visible difference though so I'm happy. Really like the 4th edition.

So far 2 working. The junk box one and the 4th edition. Guess I should try taking the last intact module from the worn 35 and put in the 80 and see it that solves the ghosting here. Can be pulled with a resistor to perfect visible result but raises power draw around 10-15 mA depending on how many digits visible.

So what's your experiences with the classics?

Is it really that hard to come by perfect samples?

Seems it's not the driver circuits to blame as I read on some page where one also corrected his 35 with a resistor pulling cathode on a digit. Seems like it's pretty much the displays themselves. Which makes me wonder - did the classics suffer from ghosting digits even back then?

Or is it the module themselves that can't cope with the reverse voltages they see as a result of the complex multiplexing or simply bad quality so some go bad with time?

Normally calculators with molded LED displays are very robust. I've _never_ seen a display module fault on neither TI nor Sharp or any other less known using molded or window chip like encapsulated LED modules. Sure the ones which are a PCB with plastic lenses heat seated or glued on top it happens from time to time they have faulty segments. Of course here the bonding wires are exposed to air and a bit of moist (incorrect storage) and the weak fastened ones let go on either the PCB or the LED chip - here I've both HP, TI and others.

-----

If only someone would display replacements for the classic series. Yeah the modules are long out of production but guess small discrete SMD diodes on a PCB and a mask could do the trick... If not small discrete ones aren't available to just put enough on a PCB.

Edited: 10 Mar 2012, 3:49 p.m.


#3

Short answer from my experience,

Replacing the three digit module which contained the ghosting digit corrected the problem on three classics that I have.

Cheers, Geoff


#4

Thank you for your answer. Pretty much the conclusion I've come to myself after hours of carefully (de)soldering.

Guess I shall keep my eye out for cheap non working / barely working / trashed / heavy worn ones then ;)

But still wonders what's the cause and whether the modules have been defective from start and tolerances for build errors where higher back then or from what reason they've become defective since. Except for the heavy one the displays are readable but of course very annoying and you need to read it carefully.

Quite "funny" on it's the last module that has been defective on all mine although not the same digit.

I presume you mean 5 digit module. Else it's not the same calculators we're talking about ;)

#5

I have no such experience with the clasics or any LED calculators for that matter, but on VFD calculators it is often solved by replacing the electrolytic capacitors in the power supply.

See the results of todays repair of a 1975 Prinztronic Scientific here.

Has anybody tried this with a classic? Did it make any difference?


#6

While I've not experienced ghosting digits on VFD myself they generally do benefit from new capacitors in the power supply. Very often a dark lazy display will enlighten significantly.
Normally I don't replace them with electrolytes but with polyester as the typical values of 3.3 µF can be had in polyester (Wima etc.). Although I usually use 2.2 µF as they're more than enough for the job (lower ESR than electrolyte) and doesn't need replacement ever again. If choosing electrolytes you can't take it for granted just to put in some batteries to power it up again in another 50 or 100 years and expect it spring to life and work like the day it left the factory eons ago. With polyester types the odds are very high.

The cause of ghosting in LED and VFD calculators are totally different causes is I see them. VFD can ghost for almost nothing because of the higher voltages and very small currents needed to drive the display. Just touching the keyboard PCB can make them ghost (VFD is typically driven with some of the same pins that the keyboard is scanned with.

Oh. And one or two lazy digits on VFD are common too. Caused by having them powered on for long time on mains which wears the most used digits - nothing can be done about that.

Edit: Btw. Nice calc there! You don't see them too often. And good you where able to save it instead of some one who doesn't know about a soldering iron throwing it out.


Edited: 11 Mar 2012, 2:30 p.m.


#7

Thank you for your answer. I have not had ghosting with any of the LED calculators I have repaired. I only have one HP classic, a 45, that works fine. I have repaired a few Commodore LED and a TI-58. No ghosting, but erratic behaviour or missing segments. Mostly corroded tracks and broken wires. A PR-100 display had a missing segment on one digit and I found a replacement display from a non working P-50.

Modern polyester capacitors are a good replacement for 30+ year old electrolytics and should live many times longer than even modern electrolytics. And no need to power them up regularly to re-form.


#8

Most, if not all the caps in HP classic switch mode power supplies are hermetic tantalums. IMO, it is a waste of time and money to replace these capacitors because they are "old".

Edited: 11 Mar 2012, 5:17 p.m.


#9

I agree, replacing capacitors is only useful if they were Al-electrolytics.


Edit:

Actually, I rescind that. I only replace the Al-electrlytics if the calculator shows signs that it may be necessary (as in the case of the Prinztronic above) - otherwise don't fix what isn't broken. One may damage a delicate case for no good reason at all.

Edited: 11 Mar 2012, 5:27 p.m.


#10

Quote:
...otherwise don't fix what isn't broken. One may damage a delicate case for no good reason at all.

I agree and think this is the correct methodology. Innocent until proven guilty is my approach to component replacement. It just seems silly to think that blindly replacing parts will have a positive result.


#11

Quote:


I agree and think this is the correct methodology. Innocent until proven guilty is my approach to component replacement. It just seems silly to think that blindly replacing parts will have a positive result.


I'd say it depends. Very often when I some by an LED or VFD calculator it's "always" in a dirty condition (I'm cheap not bidding on the mint ones on eBay??). I've only seen very few mint _and_ clean ones. Believe my SR-40 is the only one and a TI programmer coming close - none of those contains capacitors - funny. What I say is that I often find calculators in a state needing a good thorough clean. You can clean it on the outside but then pushing buttons and moving switches you can just hear and feel the dirt and grime grinding and chewing on your calculator. Also it's usually much easier to clean when taken apart as most things can then be cleaned with water, soap, a sponge and/or brush. You can get most calculators wonderfully clean, perhaps a new sponge in the keyboard (Typically TI's) and they're such a joy to use after wards with silk smooth wonderfully reacting switches and buttons and they doesn't smell like a basement either.

Having them taken apart anyway I replace electrolytes in the power supply (secondary side) if I feel the display is a bit on the dark side (VFD). More often if the calculator is difficult to take apart like difficult snap lugs.

Tantalums I usually don't replace either. That said one should be aware that tantalums when they fail usually short while electrolytes go open circuit. That's why many times when I replace electrolytes I only take the secondary ones. Those sitting more of less in parallel with the battery I don't bother. It'll work fine whether they're there or not.

So no clear picture but hope you get the point. And that is that I'm deciding on a case to case basis.

Oh and mostly that's VFD calcs that needs electrolytes replaced. For some odd reason LED calculators mostly use tantalums while VFD mostly use electrolytes. Perhaps because VFD typically where cheaper back then than LED. Typically it's also LED that has nice high quality fiber glass PCB while VFD uses phenol paper. Yes you can find exceptions to just about any point but it's the typical picture ;)


#12

I should add. That's why I have a bit of love/hate relation to many of HP's models. While they're wonderful as new they can be almost impossible to clean when found in a dirty but otherwise good condition. Like Woodstocks and Voyagers with all their heat stakes. Spice is however relatively easy when you first know the trick. Before assembly I trim the snaplock so it's much easier next time without being too loose. But also no need to by so hard.

NON LCD TI on the other hand except of the Datamaths are very easy (a few heat stakes in the Datamath keyboard but done right can be converted to screws). Many Japaneese sort of in between. While easy when case is apart they can be tricky to get apart in the first place. Generally the earlier the more service friendly without snap locks and heat stakes.

#13

Bart,

Thanks, I have two Commodore n60's.

One with ghosting digits an some minor ghosting decimals. The other has a very dim display. There are two tiny electolytics on both these at the display.

I was going to replace them, and now will.

Hi, gene, will email you about your 65's.

Cheers, Geoff


Possibly Related Threads...
Thread Author Replies Views Last Post
  A Wonderful Classic HP Calculator BShoring 2 499 11-22-2013, 05:25 PM
Last Post: BShoring
  Classic battery polarity Kiyoshi Akima 2 575 11-08-2013, 01:31 AM
Last Post: Kiyoshi Akima
  CLASSIC/SPICE KEY LEGENDS Mike T. 2 486 08-21-2013, 04:53 PM
Last Post: Mike T.
  Woodstock, Classic & Spices for iPhone Matt Agajanian 8 981 07-10-2013, 06:10 PM
Last Post: BShoring
  Classic calculator display chip type number? Juergen Hofstaedter 3 519 01-22-2013, 12:47 PM
Last Post: Eric Smith
  Classic Series Enter Key Dan Lewis 15 1,365 12-19-2012, 09:10 AM
Last Post: Thomas Radtke
  Obtaining More Decimal Digits (50g) Eddie W. Shore 3 518 09-13-2012, 05:38 PM
Last Post: Gilles Carpentier
  Classic (HP-67) charger issues revisited Matt Agajanian 6 729 05-20-2012, 10:50 PM
Last Post: Matt Agajanian
  OT: The first 4,000,000 digits of Pi, visualized in a single image Jeff Kearns 36 3,009 04-29-2012, 11:06 PM
Last Post: Jeff Kearns
  Classic battery contacts james summers 2 424 04-08-2012, 01:45 PM
Last Post: james summers

Forum Jump: