My Dyslexic Agnostic Insomniac HP-67 is misbehaving again...



My community-fixed daily HP-67 is having problems. It turns on OK, and computes, etc., but after a brief time, the display "goes crazy". No, the nuclear lithium battery pack is STILL full voltage. (Same behavior with bench supply at 4.5V)

This was the unit that went through 2 ACT chips already-- Is this an early failure or is it a more common problem such as another capacitor failure?




This is just speculation -

You mentioned 4.5V, I understand this would be the voltage of 3 alkaline cells - fresh, they might be 4.75V, fresh alkalines seem to measure 1.58V no load. Also, NiCads under charge can get as high as 1.5V each. But the nominal operating voltage is called out as 3.75V, so 4.5V is 20% high, 4.75V is 27% high. The DC-DC inverter has an output of about +6V and -12V. This puts 18 volts across two pins of the ACT and maybe other chips. I don't think those inverter outputs are regulated (NiCads being a pretty well regulated supply) so I think the outputs would increase proportionally to the input, so there might be 21.6V or 22.8V across the ACT etc. with alkalines, at least.

What is the actual voltage of the lithium cells? Factor that into the calculations above.

Now, suppose voltage for these IC's is like blood pressure for the body's organs - necessary, but harmful if too high.

The fact that any number of calculators are working fine on lithium cells doesn't disqualify the argument, because every IC is a little different - how much statistical data was needed before medical science figured out how high "high blood pressure" is? And didn't they announce a few weeks ago that 120/80 is now considered borderline?

Maybe when using lithium (or alkaline) cells, we should wire some diodes in series to drop the voltage to NiCad levels, just to be safe.

It's either that or diet and exercise!



Not a bad idea. Fortunately, none of my other calcs have shown any problems with alkaline or other 1.5v cells. I don't think other forum members have reported any problems either.

But I'll bet you're on the right track... there must be something wrong with the power supply on this one calc. In times like this I miss my oscilloscope... I have to admit: even if it dies, I'll find another ACT chip some day, and I have a couple of other, even better, 67's. But this has been my "'ole faithful" daily machine, and the thought of deceased calculators.....well, I hate to see grown men cry... especially me.

Now this is interesting: As the display acts strange, it'll sometimes still read a card. I guess it could still be a faulty chip, but I'm wondering if it's something else. It always turns on correctly for a couple of seconds before progressively "freaking out."

Ring a bell?




If it runs good for awhile, and then it goes bonkers,
I'd suggest that a bit of heating may be part of
the effect. You turn it on and a semiconductor gets warm.

One fairly common failure in a chip, is that the bond-wire has become cracked where it hooks onto the silicon die. And that's all buried inside the epoxy. It works OK at one temperature, but then the temperature changes and the connection separates.

I'd be looking for stuff like that. Bad news if the problem is INSIDE a chip. Gotta find the bad chip and replace it.

GOOD NEWS is when you get the same effect on a circuit board connection. Maybe its at a solder joint, the solder joint is cracked, and with thermal changes, the joint opens.

*** ***

I'd be in there with three things......... a can of freeze-spray, a shrink-tube heat gun, and, my bare hands.

I'd flex the circuit board a bit to see if something is driving it nuts (an open solder joint) AND, I'd be shooting the chips with freeze spray, to see if one specific chip is thermally sensitive. Then just replace that one....


O.K. I won't trust your ideas.

But I will try them.

Can't do it now (1:30 am here), but I'll try this weekend.



No, I don't have much experience using the 67. But the increased voltage could affect other chips as well as the ACT - in fact, I was reading something just the other day about how the Woodstocks and Spices, running on 2.5V, need to have the inverter generate 4V for the display. I think it was an HP Journal article from the introduction of the Woodstocks and spoke in comparison to the Classics so I guess the Classics powered the display from the battery directly. I wonder what they did on the 67, since it has the Classic battery and Woodstock insides! But regardless, the display driver chip - wherever it gets its power - might be sensitive to the extra voltage. Have you tried the machine on your bench supply at a lower voltage?


No, but I will.

Still, it worked fine for over a year (or so... clocks go so much faster than a few years ago.)

Again, I've not heard of any problems from others on the forum using alkalines in their Woodstocks, etc.



The voltage converter circuit in the classics positively definitely DOES NOT LIKE overvoltage. The DC-DC converter circuit is trying to generate a 6V regulated supply by upconverting the input voltage. This is one machine where the voltage difference between 3 Nicad cells and 3 alkaline cells is enough to give the machine a case of the heebie jeebies.


Actually, the outputs _are_ regulated (and were, right back to the HP35). The +6V output is sensed, and used to control the inverter oscillator (that's what the second transistor is for). The -12V line (and the +8V display line on the 'classics') just tag along.
One feature of this design of PSU is that the +6V output can never be _lower_ than the input voltage (if the oscillator stops, then that output is the input voltage less one diode drop). So never apply more than 6V (or so) to the battery contacts, or you will burn out chips.
Topcats have another circuit to load the input down if it attempts to exceed the +6V line. Which is why if the inverter fails in a Topcat, it will draw a high current from the battery pack (about 1A), trying to bring that voltage down.


A friend of mine used freezing spray and hot air (from a hair dryer) to track down faulty components. In your case, you might use hot air to make the problem appear, and then use freezing air to make it go away. This can help locate what component is failing. This friend used to say that a thermal shock, created by freezing a component, then suddenly heating it, could sometimes fix intermittent components.

Good luck,



Hey Michael Meyer,

When they say 'tracking down a fault' dont trust them, that's actually referring to my pioneering method of using tracks.

I think there is even one other way to look at this problem, in terms of troubleshooting.

Philosophically speaking, you've been given one thing that is to your advantage: it works fine initially, then abruptly goes nuts. That's better than if it never even turned on.

When it goes bonkers, the goal is either to get it to return to normalcy, or, keep it from going there in the first place. Beyond that, determine WHY so that it can be repaired genuinely.

HERE is a further example: Put the darn thing in the freezer!! :o) Then run it for awhile from there, does it NO LONGER go bonkers ???? i.e. that proves its definitely thermal once and for all. Beyond that, you still have to find out what is expanding and contracting to open a connection, make a cap go weak, make an inductor coil open up, etc etc.

And you don't even have to take anything apart to put it in the freezer. SAY that it solves the problem (putting it in the freezer) well, after that, try putting it in the fridge. It just gives you add'l data.

Again, the fact that it works for awhile, philosophically, can be used to help troubleshoot it. I fixed a computer monitor for somebody a couple years ago. Worked fine for 30 minutes, then went nuts. The problem was ultimately traced to a cracked solder joint (it didn't flow right, then it cracked as well). After resoldering that joint, it always worked perfect after that. The 30 minutes was a thermal effect.

If the problem is INSIDE one of the chips, well, then you'll need to find replacements.




A couple of quick suggestions : (1) Have you checked the output of the PSU inverter (at pins 1 (+6.2V) and 2 (-12V) of the ACT chip with respect to ground (-ve battery terminal) when the machine malfunctions?
(2) Check, and maybe replace, all the electrolytic capacitors on the logic PCB. There's at least one across the battery terminals, one in the +6.2V supply, 2 in the -12V supply (capacitive input doubler) and the reset capacitor connected to one pin of the ACT.


Thanks, Tony, I knew you'd know.

I'm glad to hear that 4.5V vs. 3.6 (or so) is still OK. (It's just so darn convenient)

I put the calculator in the freezer, and, Voila (or as I like to say, Viola!) it ran for a couple of minutes without flaw.

I don't know why, but I also suspected one of those pesky capacitors.

I'll work on it tonight.



Hi Tony!

Well, on a bench suppy, pin 1 reads about 5 volts, but pin 2 only reads -8V, and progressively drops to about -5V as the calculator malfunctions.

Does that narrow down a specific capacitor?

Incidentally, lowering the supply voltage to 3.5 dims the display, but results in the same malfunction.




Well, they're both a bit low, but that -ve bias supply is probably the main problem.
Near the top edge of the CPU board there are 2 2.2uF tantalum capacitors (maybe red beads with a green dot on them).Those are the caps for the -ve supply doubler. Check and replace those first.
Nearby there's a larger, 22uF tantalum bead capacitor, which is the smoothing cap for the +6V supply. You might as well check that one too.


Thanks, Tony.

I hadn't heard from you, so I already fixed it, but with your "hint" about "two capacitors...".

The mystery isn't that it's working again. (Tthough I replaced the capacitors with 1uf capacitors. Do I need to change-out for 2.2?)

The real mystery is why it was working at all for the past year or so.

Yours Truly, "The Genious", had replaced those two capacitors long ago, but installed them BACKWARDS! (I think it was me... hard to say, as this board had repairs on it, I think, when I got it...) They must have worked "well enough"!

Anyway, is the value important enough to swap out for the 2.2's? The voltages now read about 6.1 and -12.something.

THE "DYSLEXIC AGNOSTIC INSOMNIAC" CALCULATOR lives again, thanks to everyone here. It's been quite the "survivor" and has come back from the dead several times now. It really now belongs to the Forum community, but I'll hold onto it for now. It does my finances and taxes, has taught my kids math, and gives me something to do while sitting in "the library" (lavatory). And I like just admiring it too! (It's a sickness)




Hi, Doc;

Congrats, again! It's always good to receive good news.

About capacitors: inverting polarity is always a dangerous thing to electrolytic capacitors themselves, mostly if you are working close to 1/5, 1/4 (and upper) of their working voltage (it took me a while when I started working with electronics to figure WV meaning out) because they may become conductive, they will get actually hot and, in some cases, they will leak and even explode. If these capacitors are connected to signal lines, erratic operation may happen.

And using 2.2µF instead of 1µF will allow your calculator electronics to work under design specifications. I do not know in what sort of "point" are these guys connected to, but surely their original value will achieve better operation.

In any circumstances should a polarized capacitor (or any other component) be connected to inverted poles. But it happens, what else to do? The worst cases are semiconductors in general (zeners are easily toasted).

I do not know the HP67 internals so I decided to seat and watch.

Best regards, Doc.

Luiz C. Vieira - Brazil

Edited: 9 June 2003, 12:21 a.m.



Well, I went ahead and swapped-out the capacitors for 2.2uf. I can sleep nights, now, knowing my 67 is happy.

Make sure you check my post on the HP-71.



All these troubles that Michael has been having with his HP67 are something like I'm having with mine.
A couple of months ago I wrote in this forum about my HP 67.
The problem was that the first seven lines of programe wrote the same incorrect imput. Then after the seventh line all was OK.
It was suggested that it was a faulty chip and that there was nothing to be done about it. I put it in the back of my drawer and left it for a month or so.
Anyway I took it out about five days ago and - Lo and behold it worked fine. Then a day or so later - the same seven line fault.
Anyway I left it on battery charge for three days (I have 1700mAh batteries in it) and it works fine again.
Maybe my HP 67 is voltage sensitive. Although when it didn't work properly it didn't make a difference whether it was on the charger or not.
This calculator has a faulty card reader so it isn't important that I get it going as I have another. But it has me intrigued and this thread has given me a clue. I will now try it on alternative full and low powered batteries and see if this is the reason for its mysterious behaviour.
Any suggestions would be appreciated. I don't have access to, or knowlege of, any sophisticated electronic instruments.


The square pad is "+" on HP calculator component holes...


"The square pad is "+" on HP calculator component holes... "

I've been restoring these things for almost 4 years now. Been able by using my head and help here to get most every dead machine to work again. I have them polished, touched-
up, and looking and working like new.

I never figured out the above statement, nor do I remember seeing it posted elsewhere. As I hit my forehead, I can see the square and round holes.

THANK YOU for stating the obvious. I'm not sure I would have ever figured that out.....

Howard: My calc DIDN'T turn out to be a voltage problem, but a "faulty" component problem. (Yep. They don't perform as well when "backwards." Darn Kids!!)

Yours does sound to me like a chip malfunction, but why it's intermittent, is beyond me. If those memory addresses are "fried", I'm not sure how they'd be intermittent.

Again, I defer those more knowlegeable than I.



And the square pad is pin one on chips...


My guess is that the circuit will work fine with 1uF capacitors, but it won't be able to deliver as much current. This may not matter, or it may cause the machine to malfunction when hot, or when cold, or something. Personnally, I'd fit the 2.2uF ones -- they're not expensive or hard to obtain.
One thing I forgot to mention last night is that on most HP PCBs (and certainly on this one), the sqaure pad indicates the +ve lead of an electrolytic capacitor, or pin 1 of an IC (or connector), or the banded end of a diode, and so on.

Possibly Related Threads...
Thread Author Replies Views Last Post
  Misbehaving 42s Multirow Menu Annunciator? Dallas Osborne 4 698 02-26-2008, 05:00 PM
Last Post: Dallas Osborne
  Dyslexic (agnostic?) HP-67 Michael Meyer, M.D. 15 1,857 11-07-2001, 04:02 PM
Last Post: Tony Duell (UK)

Forum Jump: