HP-21 battery leakage


I acquired an HP-21 which had a serious battery leakage problem. The battery spill went down the positive contact and formed a pool up against the large chip near the bottom of the main circuit board. The leak does not appear to have reached the keyboard. After cleaning off the corrosion I am unable to find any circuit breaks but the unit does not function properly. On battery power it is completely dead. On adapter power it displays an "8" in the far right digit and randomly flashes the other digits. Since there appears to be continuity in all the contacts is it possible that the large chip at the bottom of the board (where the pool of corrosion formed) went bad? Any ideas?


I'm not that expert in repairing the Woodstock Series, but if you don't have a good battery contact, I wouldn't risk driving the unit from the mains transformer.

The batteries act as a filter (as really huge capacitors) and as a stabilization for the input voltage. If there's the possibility, they have no contact to the rest of the calculator, you could fry your calculator by driving it directly from the mains circuit. You should check the PCB traces (calc - battery contacts) first.

I don't know, if the chips get bad from battery leakage, but would guess, that this is not the big problems. You should check the PCB traces and when the batteries work again (and only then) you should try connecting a mains adapter.

Remember, there could be also defunct traces under some ICs.

I wish you best of luck in repairing the 21.



First of all, in my experience (acquired the hard way) driving a Woodstock from the adapter alone with no battery pack present is almost guaranteed to kill the calculator due to overvoltage.

Second, if the machine fails to work from batteries alone, it could be a result of the most trivial of reasons: a dead battery pack! (If a dead battery pack is inserted, the calculator won't even work from the charger.)

Third, corrosion can often be present UNDER the chips, so even if the chips aren't dead, the calculator may malfunction. You need to desolder the chips (don't try it without the proper equipment/expertise) and clean the corrosion from underneath.



As said before, NEVER RUN A WOODSTOCK SERIES CALCULATOR FROM AN AC ADAPTER IF RECHARGEABLE BATTERIES ARE NOT WORKING. The batteries act both as a filter to smooth the half-wave pulsating voltage, and also as a regulator to ensure no more than some 2.6 V are applied to internal circuits. Capacitors may work as filters but NOT as regulators: replacing the batteries with capacitors will result in excess voltage applied to internal parts and, almost sure, irrecoverable damage.

To eliminate the battery pack as a source of problems, you may try to power the calculator with two standard alkaline batteries connected in series (BE VERY CAUTIOUS ABOUT THE POLARITY), in a manner similar to the original battery pack. DO NOT USE AN AC ADAPTER WITH ALKALINE BATTERIES.

Please take this as a suggestion only, it should be done carefully. I must said that it worked for me many years ago with an HP 25, but I can not assure success on your case.


2 alkaline batteries would be at least 3 volts or more (if new). Is 2.6 volts an absolute maximum?


In my experience, 2 alkalines (under load) will give about 2.8 Volts; 2 NiCd when charging may reach about 2.6 Volts. It will be within a 10% tolerance, a reasonable value.

Many years ago, I used my HP-25 with alkalines without any problem, there was people at this forum telling similar experiences. It is very important NOT to use the AC adapter with alkaline batteries, since a charging current will be applied to the batteries, what may cause overheat and (eventually) make them explode.

If capacitors are used in place of batteries, the applied voltage will be around 8 Volts, absolutely out of range!

On the HP 21, 22, 25, 27, the batteries voltage will be applied to a switching power supply circuit, which will furnish appropriate voltages to internal components (+5 Volts, -12 Volts, I guess). On the Continuous Memory models, batteries voltage will be applied more or less directly to ICs, in order to keep memory contents while the calculator is off. So the 25C and 29C are more sensitive to overvoltages and polarity errors. I think (no warranties) that the alkalines are acceptable even on those models.


I'd like to add that because the memory is constantly powered in the 25C/29C, these machines tend to die even if they're turned off, and the battery momentarily loses contact (or has a bad or corroded contact) during charging. Another reason why I charge my Woodstock batteries in a broken HP-25!



Have you thought of fitting a zener internally?

I have done this to HP-41 nicad paks for a similar reason. Some will deliver 20V to the calculator if the batteries fail.


It is a good idea to put some extra protection in these circuits, not to only rely on the batteries...

The zener should have a series resistor, otherwise the current thru it may be excessive and blow it up. Or are you taking advantage of the resistance of the charging circuit already present?

(I never sympathized with zeners, even on my EE college days; I would prefer some IC regulator, such as LM340 or 78L05, etc. I also disliked the LM723...)


No series resistor. The charger is already current limited by virtue of the fact that nicads must be charged with a current limited (or constant current) supply.

If the charge current is 150mA and the nominal voltage is 2.4V then I would place a 3.3V (lowest voltage zener?) in the circuit. At 150mA a 1W zener would be suitable.

(perhaps I'd place a couple of normal diodes in series to get a lower voltage? - but 3.3V is likely to be safe)

In the HP41 nicad pack, a lower rating is required, 400mW is fine. The charge current is far lower.

I too dislike zeners as a means of regulating voltage sources. A shunt regulator always seemed so wasteful to me. I do have a soft spot for the good ol' 723 though.

These days I find myself using LM317s a whole lot more than anything else.

But, if you've got an existing power supply that's already current limited, then a shunt regulator (expecially when used simply to prevent overvoltage conditions) is so much easier to impliment.

Oh, and be warned that in calculators like the HP21, running off the charger only with a shunt regulator is still problematic because the charger is not filtered. A zener on its own is just there to prevent the premature death of the calculator, but not make it work without batteries.

On the HP41 you _can_ operate the calculator from the nicad pack without nicads because the charger circuit (at least two of them - there are several designs- how many?) rectifies and filters the AC from the charger. Neither of these versions will provide enough current to power the card reader. To do this, one version _requires_ a zener, the other doesn't.


No, haven't tried a Zener yet, good idea. I always considered this design somewhat stupid, where even the momentary removal of batteries can kill vital parts of a calculator.

Didn't know about the danger with the HP-41 NiCd packs, thanks for the warning!



I have an HP-25, which had similar battery leakage damage, that I brought back from the dead; basically the bottom third of the PCB was damaged by corrosion. I carefully immersed the damaged portion of the PCB in a baking-soda solution. After the baking soda had done its work for a few minutes, I scrubbed at the corroded area with an old tooth brush, particularly trying the work the bristles under the chips where additional corrosion was visible. I repeated this process a couple of times, using a fresh baking soda solution, until no evidence of corrosion was visible. Last, I carefully rinsed and dried the PCB.

The battery contacts required additional work. I cleaned them with a fiberglass brush from Radio Shack (Cat. No. 64-1986). This brush reaches into pits and cracks you can't get at otherwise, doing a much better job than scraping, and not taking away as much of the underlying good material as sandpaper.

I hope you didn't kill your HP-21 by trying to run it without batteries. Good Luck!


I wanted to thank you all for the advice regarding my HP-21. Since I've had the unit I know it was never run off of the charger without batteries in it, but it was definitely run off of the charger with the badly corroded batteries in it. From what you all have said, it sounds like that situation could have fried the 21 as well, correct?

Thanks again, Brian


Yes, it is a (regrettably) firm possibility. I suggest again to try with two alkaline, standard AA batteries. VERIFY POLARITY and work without the battery charger. If the calculator does not turn on, we should assume internal damage... You may still open it up, clean corrosion and contacts, etc., and try again. At that moment, the chances will be low and final.

Please, again, take these ideas as suggestions. I cannot guarantee any good or bad result.

Possibly Related Threads...
Thread Author Replies Views Last Post
  HP 21 - How to troubleshoot? Adam Vaughn 13 990 10-02-2013, 09:40 AM
Last Post: Adam Vaughn
  aOT--Battery/battery contact care advice needed Matt Agajanian 0 325 04-24-2012, 07:35 PM
Last Post: Matt Agajanian
  HP-35A/21 D.MS<-->Decimal degrees routines Matt Agajanian 9 873 03-29-2012, 11:03 AM
Last Post: Matt Agajanian
  21, 45, 32E--which should be next? Matt Agajanian 14 1,180 03-25-2012, 02:48 AM
Last Post: Thomas Radtke
  If the 21 came after the 45...? Matt Agajanian 5 539 03-24-2012, 05:45 PM
Last Post: Matt Agajanian
  HP-21 displays all zeroes Barry Smith 5 542 01-27-2012, 01:22 PM
Last Post: Steve Simpkin
  Cell leakage in my 48S Palmer O. Hanson, Jr. 2 368 05-14-2011, 04:11 PM
Last Post: Katie Wasserman
  New -21 and -25 apps for iPhone Eric Smith 14 1,044 03-12-2011, 05:39 PM
Last Post: Eddie W. Shore
  HP 21 calculator Chris Stam 0 234 01-30-2011, 07:58 AM
Last Post: Chris Stam
  HP 21 keyboard problem Yvan Van Hoof 2 355 11-23-2010, 12:21 PM
Last Post: Jeff O.

Forum Jump: