Zener diode


Hi, i have a HP-25 fully working with original charger and new NI-CD batteries. Is it normal that the case of the calculator on the right side is getting hot when charging?

I read here that we should add a zener diode in order to protect the electronic power circuit on the 25; what zener diode to use and where to connect it?



Hello, Michel;

in fact there is a somewhat "big" resistor (for a pocket calcualtor) at this very side of the case. It is a 1 watt (maybe 1/2 watt, for the time, I guess) and it is right in series with the "power line". And it gets warm when batteries are charging, it's normal.

BTW, I'm sending you an e-mail right now. Other subjects...

Hope this helps.

Luiz C. Vieira - Brazil


LOOK! Up in the attic! It's a techie, its an engineer,
its a Heathkit assembler,

NO! ITS CAPTAIN ZENER ! Saving HP-25C's and their owners from untimely death, fire and destruction, handing out zeners to those truly in need.

Originally from planet Motorola, Captain Zener stuck his
finger into one lightbulb socket too many, AND SUDDENLY, the voltage transients transformed him into Captain Zener, unable to go back to his originally non-Zener self.

Captain Zener's only weakness is seeing an HP calculator w/o a double-width "ENTER" button, which will cause him to lose all his super-powers until it is taken away.

And the only radioactive element to which he is vulnerable, a greenish material called "Carlynite" which if he is exposed to causes severe hemmorrhaging of profitability, and loss of any calculators worth buying.

UH OH, got so far carried away there nobody is reading.
Captain Zener will get down to business in the following post.



We had a worthwhile discussion going on the 'most correct' zener for an HP-25C. One guy said "6V zener" and I was objecting to the possible excessive amount of heat in the unit, and somebody said 0.5W zener and I said 1.0 Watt zener (also 6V is a less available value than 6.2V) One guy said the charger is 0.8VA but I said it is 1.8VA, and I didnt even have a charger handy to test with.

Therefore the thread ended w/o conclusion.

The three main concerns are

(a) making sure the zener doesn't leak current thereby slowly draining the battery, and

(b) making sure that when the zener diode is conducting heavily (somebody pulls the battery pack but leaves in the charger) that the zener can absorb the current w/o overheating.

(c) making sure that it has a predictable and rapid turnover in conduction, not conducting below a threshold, but conducting heavily above that threshold.

I finally ran a test out here, and the 1N4735A passes all the tests. It is, technically, a 6.2 Volt 1.0 Watt +-5% glass-jacketed zener with solder-leads. Motorola used to make the best ones, now it is "On Semiconductor". The doping at 6.2V gives a sharp transition not true of the lower voltage models like 5.1V etc. Therefore it doesn't leak current from the battery.

I tested this today:

For overheat, I tested it using the 1.8VA charger. It was a 3-component series test. I put two of them "back to back" (AC symmetry) and added a current-limiting resistor in series. Recalling that the resistor in the calculator is around 8 ohms, I went with 3.3 ohms just to make it a
"torture test".

The charger is indeed 10VAC unloaded, meaning 14VAC peaks (enough to fry CMOS chips with a little luck). These parts
get hot, but not unacceptably hot. The peak current is 0.4amps, and acceptably brief in the torture test. The
charger remains quite cool BTW.

Where to solder in ? Across the battery terminals, at a mechanical place of your choosing. Deliberately leave enough lead length to encourage some heat dissipation.
Presumably internal so the calculator is not cosmetically degraded. With the black stripe towards the (+) terminal of the battery.

What if you want to run your calculator with no batteries at all? Well, I doubt this gives you the license to run the calculator w/o a battery present. It only prevents high-voltage damage to the CMOS chips. The thing is, you need a bulk storage capacitor, and if memory serves, a value would be in excess of 1,000 uF and impractical to add as a rework.

I do recall running my HP-34C continuously from wall power,
around 1985, using 2,000 uF instead of batteries, and a series clamp of 5 1N4003 diodes that regulated the voltage sitting on the bulk capacitor. That's a specialized enough idea that I wouldn't expect anybody to get interested in that. Considering the classic value of these calculators, a
fully regulated DC supply would be more appropriate if you want to leave it on 24/7 . Set it carefully to something like 2.8 volts (not too hard on the LED's).

Keep reading! ----->>

Where to get a 1N4735A zener: Its the sort of thing you can get from a Digi-Key. If somebody feels they cannot locate this ten cent part readily, email me at analogee@gte.net and ** I will simply mail you one for free ** for the first few users who agree to test out and report their satisfaction in HP-25C, but dont bug me about that unless you cant readily get it from other sources (wouldn't want you to pay $4.00 UPS shipping on a ten cent zener).

- Captain Zener


Somebody plz explain whether the HP-34C is equally
vulnerable in the same way that HP-25C is.

Like if you pull the battery but you plug in the charger, is it a possible way of frying the chips on the HP-34C.


The 34C has about the same battery and charger circuit and also needs to have the battery in place at all times to work properly, but the Woodstocks - and apparently the 25 and 25C in particular - seem to be particularly sensitive to the higher voltage. The RAM chips that provides memory for the storage registers and program memory in the 25 and 25C fail a lot. I have several 25's that work except for the memory and only one that has a working memory.


OK, thanks for that information.

So if a feller was really fussy, he could add that same 1N4735A zener diode to the HP-34C .

However, sounds like it isn't necessary since there was not an obvious track record of damaging the units by overvoltage.

WELL, what's going on here is the progression of semiconductors, CMOS computing / memory chips in particular.

We're trying to run the chip normally on 2V - 3V and all of a sudden comes a 14V overvoltage. The question is which generation of chips was most rugged.

In the very earliest days, the chips were fragile due to the technology being embryonic and not fully thought out (the 1970's and early 80's).

In the intermediate time (the 80's) the chips were reasonably tough as the process matured, and they could probably take a lot of overvoltage.

But in the 90's and up to the present day, the chips are fragile once again, because they utilized the mature process techniques to make the geometry super-tiny so that they will run at 100MHz or a gigahertz.

Super tiny geometries once again, are frail.

So what you're telling me is that historically, the chips in the HP-34C could withstand the 14V overvoltage because they were from the right vintage of CMOS manufacture.

And therefore it's none too necessary to add a zener to the 34C.


The Museum article on Series 20 technology (http://www.hpmuseum.org/tech20.htm says that most of the chips in the Woodstocks are PMOS and only the 25C and 29C had CMOS memory chips. The Series 30 articles don't say what IC technolgy is used and only 3 models (33C, 34C and 38C) have continuous memory.

In the earlier zener diode discussion, I was thinking of asking you what you thought of the "programmable zener" IC's as regards sharpness of knee and leakage current. Then I remembered that the one of these that I have seen used most is (without any external voltage divider) a 2.5V device! TL431, so it might be a candidate for a "battery replacement" for AC operation. I've only seen it in TO-92 and 8 pin surface mount packages which are rated for less than 1 watt dissipation but according to a TI databook it was also available in a surface mount transistor package with a heat sink tab (SOT-89) rated for up to 3 watts. However, I don't see this package listed in the DigiKey or Mouser catalogs. So I'm thinking of taking a TO-92 part and epoxying it to a piece if aluminum the size of the inside of the battery holder. I've seen small lamp dimmers that heatsink their Triac that way, which will dissipate a few watts into and through the plastic shell of the assembly. The databook says the 431 only needs 1 mA to meet its voltage specs. I'll need to measure the calculator current drain with just one "1" and with all "8"s displayed to get a good idea of minimum and maximum current requirement - better see what it is with a program running too. One other thing that I noticed in a PPC Journal article is that the HP25 power switch closes two circuits when the unit is on - it might short out the current limiting resistor and just depend on the transformer winding. How do you suppose a device like a 431 would behave with a large capacitor across it?


"We're trying to run the chip normally on 2V - 3V and all of a sudden comes a 14V overvoltage."

In the PMOS Woodstocks like the 25, a DC-DC convertor generates +6V and -12V for the IC's from the 2.5V supply. If the convertor provides the same step-up at 14V input, the IC's would see 84V across their power pins!

This suggests another possible way to protect the calculator: add voltage limiting to the DC-DC convertor, preferably on the primary side of its transformer.

The RAM chip in the 25C must run directly from the battery. I don't have a 25C but I have heard that its memory chip is even more sensitive than the 25.

BTW, I understand there is another supply, +4V nominal, which feeds the displays in the Woodstock and Classic machines.

Also BTW, the Classics have the DC-DC convertor and use PMOS IC's like the Woodstocks, but don't suffer from the overvoltage problem because the recharger provides two separate power supplies: a constant voltage supply to run the calculator during charging, and a constant current supply to charge the battery. When you remove the recharger plug from the calculator, a spring loaded contact connects the battery to the calculator power input. The 65 and 67 have their card reader circuit powered directly from the battery connection so it is said that their card reader circuit can be damaged by running from the recharger without a battery, since the constant current supply voltage will rise to something like 12V when the battery is missing.

I wonder what kind of power supply the Spice models have inside - have you noticed a toroidal transformer?


I dont have a data sheet for the TL431. I think
it would be OK to have the large capacitor across it
because it will shunt the current BEFORE the capacitor
charges up higher than the threshold.

MEANWHILE what will bite you on the active zener thing
is that it always draws some power. I dont have
a spec sheet. Say that it is 50 microamps (its
probably 5 or 10 times more) well, that will drain
the battery slowly and reduce the time that your
calculator can sit before the battery is drained.

The 1N4735A has now been carefully chosen, and
the thing is, it wont draw 10 uA, it wont even draw 5uA,
when the battery is just sitting there (i.e. when it is not zenering to do protection).

So the 1N4735A meets the requirements of the application. You could try to find something else, but why, if a 10 cent part will satisfy the requirement.

Its quite puzzling that HP didn't put the part in
at the original time. I mean, the part was available then.
Obviously the designers of the time didn't appreciate the problem of integrated circuits failing from momentary over-voltage transients. It was learned after the HP-25 was being manufactured.

- Norm


Its quite puzzling that HP didn't put the part in at the
original time. I mean, the part was available then. Obviously the
designers of the time didn't appreciate the problem of integrated
circuits failing from momentary over-voltage transients. It was learned
after the HP-25 was being manufactured.

Hmm. I expect that they were well aware of how easily ICs could be
destroyed, but manufacturers often tend to be "penny-wise and pound
foolish", especially since a few pennies per unit can add up to a lot of
pounds when a product is being mass-produced. So perhaps they settled
for a stern warning not to connect the charger / external power supply
without having batteries installed? But they may well have been guilty
of not thinking through the possibilities of failed batteries and poor
battery contacts.



I was primarily considering the 431 as a sort of battery eliminator for AC only use, similar to the way you describe using a string of forward biased diodes. It's only $0.32 in the Mouser catalog ($0.63 in DigiKey) or you can pull one from many old VGA boards where it was used as a reference for the RAMDAC. It is very sharp - 0.2 ohm dynamic impedance - and only requires 1 mA to meet its voltage specs.

The only reason to use it as a protective zener would be to get a desired voltage lower than 6.2V with a sharp knee. To do so, it would be necessary to add a resistor voltage divider to put 2.5 volts on the control pin at the desired zener voltage. The control pin itself draws 4 microamps max (2 typical). Using a "rule of thumb", I guess you'd want to swamp that 10 times in the voltage divider, for 40 microamp total - but that would be at the zener voltage, at 2.5 V battery potential it would be proportionally less. It has a diode across the shunt transistor to conduct when forward biased and its reverse leakage is specified as 1 microamp max.

Of course, at the higher voltage it would be dissipating more when it was "working". Part of my pipe dream is that if I can use some of the space where the NiCads go to put a sheet of aluminum to use as a heat sink (with the flat of the TO-92 epoxied to it) to pass the heat into the plastic of the battery pack, I might be able to dissipate as much as the SOT-89, which is 3125 mW at 25 C case temperature (2000 mW at 70 C and, strangely, 2625 mW at 85 C - one of those must be a typo, or else they are swapped). The TO-92 itself is only spec'd at 775 mW at 25 C ambient (496 mW at 70 C, 403 mW at 85 C).

It could do both jobs (battery eliminator, protective zener) with a switch to short out the top resistor in the voltage divider when the cells are out. I've got a TI2500, their first calculator, that can run on NiCad or alkaline AA cells. There were several versions of the 2500 that used different numbers of cells - mine uses either 3 NiCads or 4 alkalines - sounds backwards but that's what it is. It uses the same cell holders for either kind of cells. When you install the fourth alkaline, you displace a metal clip which I assume is operating a switch inside the calculator. I guess it either disconnects the recharger jack or conditions the unit for the different total battery voltage, or both.


Hello Ellis,
just replying to your post.

You think much like me, about having an "AC desktop" solution for an HP calculator. You know, to leave it on 24/7 on the desktop. You can't do that with batteries in it, because its a charging circuit that will overcharge and outgas the batteries.

I was doing that back around 1985 - 1990 with my HP-34C.

I took out the batteries, and put in two aluminum electrolytic caps. I used the largest pair that fit into the "AA" battery spaces. You know, pick a voltage rating on the cap of only 6.3V, but then look for lots of capacitance.

I think two caps at 1000uF was enough. But apologies
I dont remember for sure. Maybe you needed even bigger
like 2200uF apiece.

Then came the string of 5 diodes to limit the maximum
voltage value.

BACK TO THIS 431. Well, you could apply it, but I'm not gung-ho. 5 diodes times 0.6V per diode gives you a
predictable 3.0 volts. Want the RED LED's a little dimmer? Maybe to extend their life in a 24/7 situation? Make it 4 diodes and you'll be at about 2.4 volts not 3.0 volts.

What could be wrong with that? Simple as putting
a wheel onto a go-cart. If you hook that TL431 up wrong
(or it fries) then you lost your regulation.

ALSO, you mentioned that maybe it would do both jobs
(i.e. protect the calculator ). Well it can't do the job
of protecting the calculator if you are running the calculator on batteries.

Didnt' you say that the active regulator circuit required 1 milliampere of current to operate in-spec ? That's too much, it will drain-down the battery and noticeably shorten the charge storage time.

Don't get me wrong. I like fancy stuff. I like circuits from the linear databook, and beyond. I like active zener circuits and voltage regulation that is good to 5 decimal places. BUT I like to use the right hammer for the job. If all you need is 5 diodes, why go beyond. They are very reliable and dont need a heatsink.

I'd take a line from Scotty "the more they overdo the plumbing, the easier it is to stop up the drain".

Suggest that if you just use 5 diodes, your modification
will be finished same-day but if you use the 431, it will be a lot of foolin' around and delay you from a satisfied result.

Good luck.


"Didnt' you say that the active regulator circuit required 1 milliampere of current to operate in-spec ? That's too much, it will drain-down the battery and noticeably shorten the charge storage time."

Well, in the protective version of the circuit, you have a resistor voltage divider to program the device to be a ~6V zener. Until the control pin reaches 2.5V, the shunt element is open circuit and only has 1 microamp max. leakage. So only that and the voltage divider would be drawing current.

What got me started on this is the possibility that something less than 6.2V would provide better protection, or to state it another way, 6.2V might be too high to provide the protection desired. Considering the DC-DC convertor in the HP25, and assuming its output increases linearly with the input voltage, 6.2V in could be putting 44v across the PMOS chips! I know the 6.2V zener has the best dynamic impedance - BTW, why is that? Is it one of God's secrets?

So I started thinking about using a programmable zener to get a lower zener voltage with a sharp knee. Then I remembered that the one part I was familiar with, the TL431, operated at 2.5V - and the idea of the battery eliminator application jumped out at me! Again, the advantage over a string of forward biased diodes would be the better dynamic impedance.

When the idea of doing both jobs with the same part - with a voltage divider for the protective circuit, then shorting out the top resistor to make it a 2.5V zener for the battery eliminator circuit, accomplished by a switch that acts when the cells are installed as in the TI2500 - at that point I guess mission creep had set in!

I'm pretty sure before I'm finished I'll have a little 1" square fan in the battery compartment to cool the device!


It's true that 6.2V is none too low, BUT, if you
put in the 14V peaks to a 25C, they dont immediately
fail. Some of them fail some don't. So if you drop
back to 6.2V (from 14) it would imply that, effectively,
none of them would fail.

What still holds up the active zener, to me,
is whether it leaks too much when non-conducting.
I didn't see a link to a data sheet, or I could
try to review it with you.

If you set the active zener to about 5V,
then you drop the voltage to 3.0V
then we need to measure leakage of the whole

The 1N4735A would be around 1 to 5 microamps.
Up to probably 100 microamps is not too worrisome.

Somehow I thought 1 milliampere was implied on
the active device. If it is only 1 microamp
then that's well below 100 microamps.

However input leakage at one terminal does not
determine the leakage of the entire circuit.

It still appears to me that the 1N4735A
meets all the criteria for a protective device,
and the 25C is far better off with that item
than without it.

You ask why the 6.2V zener has such a low
dynamic impedance. Well, its just a P-N junction
and they manipulate the doping of the crystals to
deliberately set a low breakdown voltage. A zener
breaks down much lower than rectifier diodes, of course.

When they change the doping around, they get inherently
good results at 6.2V , quite lousy results at 5V and below, and reasonable results at higher voltages (10V 22V 33V etc).

Yes 6.2V seems to be the best. No, I dont know why, but one could go back to the semiconductor math class for that and get a feel for it. I'm satisfied just to see that it is so.

Motorola used to make those parts and they had very high standards (kind of like Mr. Hewlett and Mr. Packard).
Motorola is now selling only bigger ticket items (cell phones, microprocessors) and the zener factory went to "ON SEMICONDUCTER" (www.onsemi.com) you can get a 1N4735A data sheet there.

"Diodes Inc" makes diodes, and they arent' hardly going to stop with a name like that. Their zeners were of poorer electrical qualities than Motorola's. Poorer transitions in the conducting I/V curve. I also bought Diodes from them once that had been packaged in a low-temperature meltable plastic, barely able to make it thru a soldering process.
Of course, I presume they fixed their error, but makes you wonder who's in charge.


Although I have not traced out the power supply of the Spice machines (never had a bad one), I suspect that it is somewhat better than the Woodstocks (hard to be much worse). The chip failure rate of the Spice machines is MUCH less than the Woodstocks. I have never seen a bad RAM chip in a Spice machine. A couple of bad ROM chips and a couple of bad CPU chips is all.


More points in the plus column for Spice! and the 34C


Hello, Norm; good to read your posts again.

> One guy said the charger is 0.8VA but I said it is 1.8VA, and I didnt even have a charger handy to test with.

I was the one who mentioned the 0.8W, and I explained I was using a Spice's charger, rated 0.8W, to charge the batteries in my HP25 for more than ten years. When you mentioned the 1.8VA (power factor to be disregarded) Woodstock AC adapter in another post, I posted a message about this all.

I would like to carry the final workbench tests that you suggested, but I have no precision "meter", just one multimeter that I repaired once and did not have the chance to calibrate and a capacitance meter. And you are completely right: the thread ended without conclusion.

About the "final" Captain Zener solution: if not available, what risks one will face if using a 6.2V, 1W zener diode that is not the 1N4735A?

Two things: if batteries are O.K. and in place, the zener diode is transparent to the circuit because battery voltage will not be over (nominal voltage plus 15 to 20%), right? If voltage across the zener diode does not reach the test-voltage value, then it will not conduct. I never considered the fact that the zener diode would leak battery current. Is it a wrong analysis?

The other thing: isn't the charger based on a half-wave rectifier? I compared the HP21's P.S. diagram with the HP25 circuit and I found the first components to be the same: one resistor and one diode. Should the peak voltage be considered 14V or 7V?

I once thought about driving a transistor to form a "higher-current" zener. As it will only act when batteries are not O.K. or absent, the AC adapter would only supply the maximum available current, that would be partialy limited by the (~)8-ohms resistor.

I myself had a few doubts (not about your posts, about my own concerns) because I did not have the necessary equipment in my workbench, and I thought I should research a lot more and offer facts instead of questions. You now offered a lot of facts and I keep asking questions.

Thank you.

Best regards.

Luiz C. Vieira - Brazil


Luiz Vieira wrote:

> I would like to carry the final workbench tests that you suggested, but I have no precision "meter", just one multimeter that I repaired once ...

Hello Luiz,
Well, to get a clear picture you want to use a Tektronix oscilloscope, a mainstay back at the secret underground laboratory of Captain Zener. It helps to see exactly what's cooking with the waveforms. The charger has got a bit of 3rd harmonic distortion for example, and also, to exactly peg the peak current in my 'torture test' that made sure the zener wouldn't melt



Luiz Vieira wrote:

> About the "final" Captain Zener solution: if not available, what risks one will face if using a 6.2V, 1W zener diode that is not the 1N4735A?

What matters here is the 1 Watt most especially. After that, something around 6.2V is good. If you drop down in voltage (like a 5.6V or a 5.1V) at some point your going to get some leakage even at the idle battery voltage (2.8V).
I'm not sure which other zener you have in mind, because there's not that many zener families in circulation, one of the most common alternative families is a 1/2 watt series... dont go there.

Again, again, again, let me dwell on something about the family of zener diodes that comes from a semiconductor factory: They have a very sharp transition at 6.2V, and don't leak any current if they are at or below that optimal 6.2V threshold.
But if the diode is adjusted by dopants so that it clamps lower, for its manufacture, specification, and sale,
like the ones specified at 5.1V or heaven forbid 3.3V (those are ghastly) then the thing is so "moooshy" that it will leak current almost no matter what voltage you put onto it.
That's why we are, in the end, reaching for the 6.2V units, to exploit the sharpness of the threshold. I would probably breathe easy with a 5.6V but expect trouble even with a 5.1V (it will tend to draw down the battery gradually due to leakage in such a zener).

Luiz Vieira wrote:

> if batteries are O.K. and in place, the zener diode is transparent to the circuit because battery voltage will not be over (nominal voltage
plus 15 to 20%), right? If voltage across the zener diode does not reach the test-voltage value, then it will not conduct. I never considered the fact that the zener diode would leak battery current. Is it a wrong analysis?

Hello Luiz,

Yeah, you are drifting over to the 'dark side' here
(sorry, old Star Wars laserdiscs tonight). Your comment is true for the 6.2V unit because it transitions so sharp and below 6.2V does not leak hardly even a few microamps. At 3.3V no problemo.

But if you go and grab for example, a 3.3V zener and figure it wont conduct at 3.2V, you are in for heap-plenty disappointment. A 3.3V zener will still be conducting crazily at 2 1/2 V. And that unacceptable leakage is present even on a 3.9V or a 5.1V zener, because of the qualities of the semiconductor doping.

Basically a zener diode "likes to be around 6V" they can be built to run very well there with a sharp threshold. When they fill the purchase-orders for diodes with lower voltages, the excessive dopants create a mooshy and gradual transition, which isn't really what people want. They want abrupt conductivity at the specified zener voltage, and no conductivity below it.

BTW, for purposes of navigation, zeners specified HIGHER THAN 6V (8.2V, 10V etc) are also quite reasonably sharp. But the 6V zeners are sharpest of all. The data sheet chooses values spaced logarithmically so 6.2V is the typical number called out (like for 1N4735A).

Luiz Vieira wrote

> The other thing: isn't the charger based on a half-wave rectifier?

Yes it is. For the torture test, I happened to choose 2 zeners in "back-to-back series" so that the transformer has a more symmetrical loading. Asymmetric loading (one diode) is a bit distasteful. For purposes of charging the batteries, the single diode, obviously, was enough.
For purposes of observing heat at the zener, the torture test chosen was fair enough, and worse than the actual situation for which we are applying it .

Luiz Vieira wrote:

> Should the peak voltage be considered 14V or 7V?

Not 7V. It is 14V. You must always have one lead of the transformer as a ground reference, and there is a peak at +14V and another at -14V, and nothing else. It is true that due to the single diode, HP only uses the positive half-peak at +14V.

Luiz Vieira wrote:

> I once thought about driving a transistor to form a "higher-current" zener. As it will only act when batteries are not O.K. or absent

It is common to utilize transistors as pass elements when a zener is not available at a high-enough wattage. Such techniques can also 'synthesize' a multi-component zener that behaves better with low thresholds (that 3.3V specified zener problem). But now you are into all sorts of tedium to actually specify and design the transistor and its circuit. You would only do that if necessary; here its not so use a 1N4735A.

Luiz Vieira wrote:

> I had a few doubts (not about your posts, about my own concerns) because I did not have the necessary equipment

"Have no fear, Captain Zener is here."


- Norm
Seattle area


Hi, Norm;

Thank you again! I must go back to my Analog Footnotes and read them all over again. Some information is lost among the years and it's necessary a "refresh" (dynamic memory afairs) form time to time.

Hey, can I (we) count on you when this sort of subject comes to bother? I see that you have gone deeper enough to even consider construction techniques and their consequence in components performance.

What secret sign should we use to call Captain Zener? (a particualr nice touch of yours...) What is the color of his back cover so we do not confuse him for other flying entities?

Thank you again; mostly to keep your posts in such an easy, funny, educative (no pun intended at all) speech.

Luiz C. Vieira - Brazil


Hi Luiz, glad it helped.
A lot of analog engineering work is knowing
approximately what is the big picture,
and only going into micro-details when necessary.

For example, if I need to pass 30 watts thru a
transistor and its only rated for 5 watts, then it
would be a waste of time to run a bunch of equations
for what bias resistors to use, since it will go up
in smoke anyway. The big picture says go back to the transistor catalog and pick another transistor, BUT, a lot of the design engineers, amazingly, will be busy running formulas. As an aside, obviously, making no constructive progress either.

Similarly, with the zener family, you have to see the big picture that although they provide an entire family of parts, doped to any voltage, that the 6.2V unit is "the best one available" and the lower voltage ones are really lousy.

It therefore works out nicely that the 6.2V unit (1N4735A) also fills the need for the HP-25C problem.

(HP-34C has the same problem ????)

As to captain Zener's secret identity, it can be readily
found if truly needed, 2-3 posts back was my e-mail address and an offer to send out free zeners to those truly in need and willing to report back.

Thanks for having helped to hash this out, Luiz.

- Norm
(Seattle area)


You can always spot Captain Zener when he's flying: He's the one with the knees bent at almost a 90 degree angle (unlike those other super guys, who fly with their knees stretched straight behind them!).


His knees are sharply bent.


Hi, folks;

Excuse-me digging here again, but I do not like keeping error records, instead I prefer understanding what was wrong.

I was connecting to the www right now, trying to figure out where did I get the 7V peak voltage instead of 14V peak (correct answer, for sure) and I found where my brain stuck down: symmetrical power supply and bridge rectifier.

I believe I am writing this post to justify myself for my own guilty feeling.

Thank you, Norm. Send my regards to Cp. Zener

Luiz C. Vieira - Brazil


Any relation to Clarence Zener?


I think maybe Captain Zener is in breakdown!


Maybe he's caught in an avalanche?


Is tunneling out an option?


... its just a quiescent matter.


Hello Luiz Cláudio:

You wrote me and I answered you....

have you received my answer?.
Please ... write me and send me your e-mail address to write you


Daniel (www.teclas.org)


So what is your conclusion : I'm not gonna use my calculator with the adapter connect and i'll never remove the batteries so is their still danger for my 25 in these conditions?

Can i make a power supplies fully regulated for my 25, 41 41 printer, 67, and 97 so i can discard the original dnagerous AC wall adapter that come with all these claculators?



The danger lies in the chance that the battery may appear to be OK or inserted and connected properly, but the woodstock series battery contacts (and HP-10A and HP19C) are rather well known for, well, not making contact. Often times one will insert a good battery and the machine will not come on. You then take out the battery and snap it back in and, presto, you have joy. One thing that can help is to put a small wad of aluminum foil on top of each battery post before inserting the pack.


Here is your conclusion:

You are a 25C user, and so, it depends on what you plan
to do with it. If you dont plan to use it much and you want it 100% stock and you dont like modifying something, and you don't want to take the back off, then don't do anything !!

Just be careful about this issue with not removing the battery while the charger is on, and leave it at that. We can't MAKE YOU put a modification into your calculator.

On the other hand, maybe you plan to use it quite a lot, the battery terminals are a little tired-looking, and you fear you might be forgetful about this 'remove the battery pack while the charger is on' (or maybe you got a 12 year old and he might do it). And you have already got the back off, and you dont mind modifying something, WELL, IN THAT CASE GO AHEAD AND ADD A 1N4735A and your HP-25C is that much more bulletproof. If you dont have such a zener but you plan to install it, email analogee@gte.net and provide a street address and I'll just mail you one for free.

SO it all depends on you. If memory serves, somebody asked about the zener rework because they wanted to put the rework in. Maybe that was you, or it wasnt, I dont know. It all depends on your personal electronic preferences and the amount of take-apart your happy with, and whether you want to risk blowing chips in a classic calculator over a bit of dirt on a battery terminal.

Let us know what you decide !! Good luck !

- Captain Zener


One other thing....

You mentioned an interest in "replacing the dangerous wall adapter". (Wall adapter occasionally referred to as "wallbug" btw).

I'll provide this add'l post to try to de-rail that train of thought.

The thing is, the wallbug works fine, and its an authentic original device, looks right, and really isn't the essence of the problem. The lack of voltage protection in the calculator is more the real problem. If you replace the wallbug with something else, all of a sudden you have to start worrying again... "what about this, what about that".

There is an internal resistor in the calculator and the wallbug matches to that. If the zener is added you have a complete, working, problem-free solution. If you replace the wallbug with another, then I gotta say 'which wallbug' and I dont want to go there, and we DONT have a complete solution, and it wont even look cosmetically correct and the cord wont have the right plug on the end.

ALSO ABOUT BATTERY CHARGING..... the reason batteries
charge up is because of periodic pulsing of the 60 Hz
(or 120Hz if full-wave rectified). If you put a smooth DC current into a battery, surprise, it wont charge up properly. It will develop a surface charge and quit taking current. The pulses help because they sort of 'rattle the electricity into position' and help load the cargo into the container. If you didnt know that dont feel bad, I'm 38 and only learned it recently.

THEREFORE you talk about replacing the wall adapter with a steady DC power source and if you do that it wont charge up and so you've permanently attached yourself to a wall socket.

Maybe you want to put your calculator on a desk and leave it on 24/7 with household power. In that case, you might seriously consider a fixed steady DC power source because its well regulated, but even then you have to sacrifice a classic wall adapter to steal the cord off of it. Why do that if the wall adapter can power the thing anyway. Besides most people arent going to want a classic calculator sitting powered for 24/7. And the battery wont charge correctly, in fact, if you did that, you would take the battery out because the battery has its own ideas about what the incoming voltage ought to be and thats a conflict.

If you want to run the calculator 24/7 on a desk, I already discussed something carefully... creating a battery pack substitute for an HP-25C or an HP-34C. You need 2000 uF of capacitance (approx?) and 5 diodes. You put that into it
instead of the battery pack you plug in the charger and you leave it on forever (saves wear on the slide switch!) I was doing that around 1985 with my HP-34C. But really you still dont truly need a dedicated DC power supply, you can STILL use just the original wall adapter charger that came with the unit.

Whats keeping you from leaving a unit on 24/7 is actually the batteries. The circuit HP uses doesn't stop charging. Therefore, it appears you could overcharge the batteries, so HP set you into a cycle of charging it, then discharging it. Most of us are accepting of that, but maybe you want to just power it from a wall socket permanently.

There is nothing bad or dangerous about the charger, rather, its the lack of regulation/protection to the IC's in the calculator, and to a lesser extent, that you can overcharge the batteries if you dont unplug the charger after awhile.

- Captain Zener

HP didn't really set you up for contin


Ok, thanks forr all these advice, i didn't know about rechargeable batteries needs.

Well, my HP-25 is working like new BUT is very ugly; full of scratches, no more gold label on the faceplate (i put a DYMO sticker on all the shift keys on the faceplate!), the plastic LED cover has been artistcaly scratches etc... so i don't care about modifying it

Secondly, i use it very often : i have a HP-41CX on my desk connect to a printer at school (i teach physic) but since i prefer LED and the form factor of the 25, my newly installed batteries drained in about 2 working days. So i'll make the modification! My students ask me why i carry this ugly calculator in my pocket all day long insted of the one i teach them to use : the TI-83 ;-); i don't really have an answer yet; don't you have one for me?

Just a last question : will the zener draw some current when the calculator if off? I mean enouf to make my batteries drain noticely faster?

Thanks again!



Don't laught ;-)



Michel asked if the 1N4735A zener is going to
drain-down the battery. NO, it will not leak current
at the battery voltage of about 2.6 - 3.0V .

MEANWHILE, sounds like the plan here is to leave
the calculator plugged in continuously, turned on ?

Is that OK? If you leave the charger plugged in forever, then does the battery pack overheat or outgas ?


No, just plug it for changing and never use it with the adaptor. Any problem with discharging in this situation?

thanks again


OK, just to confirm your intentions, your plan for usage of the HP-25C looks A-O-K and typical.

You describe you'd be using the HP-25C exactly as the manufacturer intended. Plug it in to charge it (maybe use it simultaneously) and then, unplug it and run the battery down, and repeat that process.

Adding the zener enhances the ruggedness of the calculator, by giving protection against damage if the battery pack is removed with the charger on (or gets an intermittent connection at the battery terminal while the charger is on).

As per your last question, I'll confirm, one more time, NOOO, there wont be any unwanted discharge of the battery by adding a 1N4735A . None whatsoever. Because it only conducts heavily at 6.2V, and does not conduct at all for voltages of up to 5V.

If it proves useful, here is a link to a data sheet, for a 1N4735A zener diode (it is part of a numbered family starting at 1N4728).


If you have an electronics store around, they would have them. It's kind of like going to a hardware store and asking for a #10-24 screw. But even at that, not everybody has them. In bulk, they are a 7 cent part. I've got a few around here.

I'm starting a post to discuss the 48G+ graphing calculator (I dont like it) if you have any input feel free to help get that dialogue going.

- Norm


Before Captain Zener departs the scene and returns to his secret underground laboratory,

plz be sure to consider this one last thought:
don't solder the zener in backwards !! The black band
MUST go to the (+) terminal of the battery contacts.

The unbanded end goes to the (-) battery contact.

If you were to install it backwards, you would get a
short circuit with very high current, overheating of
the battery, and smoking and destruction of the zener with possible harm to the calculator also.

It would be almost as damaging and destructive as putting a chunk of Carlynite in front of Captain Zener.


Ok, thanks for all your advices; i'll tell you what happend after i install one of these zener.


Well actually the zener WILL have a small amount of current leakage. Mayby even enough to double the apparent self-discharge rate of the cells... but if you actually use the machine any at all the increased leakage current will never be noticed.


It is simply a more elegant solution.

RPN KICKS Algebraic BUTT!!!
End of discussion.

LED Looks better than LCD. You have a pocket programmable, and the Ti-83 doesn't fit in a normal pocket.

You like the one extra critical angles mode that the Ti-83 lacks (gradiants, this is obviously not true, but you may get them to investigate a little history, or look into CE).

Any of the above could be used as an excuse to use an Hp25 over the Ti-83. Whether any are practical reasons, doesn't really matter I suspect.

We could deveolp a top ten list as I am sure there are numerous reasons why WE would use an Hp25 (or any older Hp LED) over a Ti-83. Many older Hp's used to have units conversions, a Ti-83 has none built in.

Possibly Related Threads...
Thread Author Replies Views Last Post
  [WP-34s] Updated: How to install crystal and IR diode Alexander Oestert 2 868 06-17-2012, 12:19 PM
Last Post: Geoff Quickfall
  [WP-34s] How to install crystal and IR diode.pdf Alexander Oestert 27 3,774 05-04-2012, 06:21 AM
Last Post: Dominic Richens
  HP-48SX Repair - Diode aj04062 3 791 04-21-2010, 05:26 PM
Last Post: aj04062
  HP-41C/CV/CX Battery Pack Circuit (burnt diode) Al Snyder 4 933 01-27-2008, 04:51 PM
Last Post: Al Snyder
  Schottky diode in 48g caribet 0 433 09-18-2006, 05:43 AM
Last Post: caribet
  Calling Captain Zener! james (UK) 0 388 10-17-2003, 06:24 PM
Last Post: james (UK)
  A Call for Captain Zener ( and others) for HELP! Vieira, Luiz C. (Brazil) 17 2,406 04-28-2003, 11:08 AM
Last Post: Ellis Easley

Forum Jump: