HP Forums

Full Version: HP-25 Bugs
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

I have taken my HP-25 out of the drawer and after some new batteries it works fine. This is a very early unit, serial number 1509A13072. (Second week of production?) After I bought this unit I got a recall notice from HP about a defect in the calculator. I did not take them up on the repair.

I don't remember details but if you were using a stored program and saved a value (1 x 10 -25 or something) in some memory location then switched to program mode, the calculator would crash. (It helped if you were standing on one foot.) I was able to reproduce the error back then but I don’t remember the exact sequence now.

Does anyone have any information on this? I will be here pressing keys until I reproduce it.

Michael Holley

There is an HP-25 bug listed in the "Collecting and Using" "Bugs" section of the Museum that might be what you are talking about. Here is a link . The HP-25 link is about half way down the page.

If you store a number from 1e-10 to 1e-20 or do a trig operation that is within that range and then switch to PROGRAM mode, the display will go blank. Switch to RUN and it will say ERROR. I think that is the proper range. I know 1E-14 does it... I discovered this bug during an EE exam and reported it to HP. A couple of months later they sent me an addendum card (but no offer to fix my machine).

The defective code for this bug lies in the ROM0/display driver chip located under the LED display module. Its part number changed when the code was fixed in later machines.

Thanks for the links and responses. I was thinking about the 10e-11 to 10e-20 storage bug. Mine still has that defect, 20 years in a drawer didn't cure it.

Michael Holley

Is it the fact that the display driver chip also contains a program ROM that is responsible for the random flashing display during program execution?

In a way... HP purposely flashes the LED display during program execution so that you know a program is running. They typically blank then unblank the display with each program step (this is discussed in an HP65 Key Note). I think the reason that the display (anode) driver chip also contains a firmware ROM is just to save packages. The chip package already had the required pins and the die was big enough to contain the ROM.

I originally thought that the ROM0 chip only contained display related firmware, but I have an early HP22 that is built on an HP21 circuit board (based upon its 00021-xxxxx part number) that is identical to an early HP25 circut board in all respects except for the ROM0 chip. I also had an HP25 with a bad anode driver (ROM0) chip. When I swapped the chip with a different one, the machine developed a case of the ROM bug.

While I don't have my HP25 anymore (sigh!), I remember that the display patterns during the run of a program are not 100% random. Changing modes (such as FIX-SCI-ENG or DEG-RAD-GRAD) make the leftmost digit to show faint 2-1-0 values...

If somebody has a HP25 at hand, please experiment a little and report his/her observations.

Also "shaking" my HP25 produced a "rorE" display instead of the normal "Error"

All scanned displays except LCDs make interesting patterns if you move either the source or your head. I think what you are describing with "Error" happens because as you move the calculator through space, the machine is energizing the digits one at a time, so the lit digits are not located relative to each other as HP intended. LCDs don't exhibit this because while they might be scanned, their response time is so slow that they don't return to "off" between refreshes. My battery ran out while I was trying your suggestion, so I'll try again later. But I've noticed that sometimes I see "my" numbers flashing by on the display during program execution. If I remember correctly, TI calculators always display "X" or whatever they call it, I think I remember one of their models recommended to include the "=" sign repeated different numbers of times to produce variable length "pause" displays (I don't see how that would work in conjunction with the constant feature, "=" should repeat the last operation?)

I would like to read the HP-65 Key Note that has to do with the display output during a running program. Do you know where I can find it? I am working on something where this would be very useful.

Matt Riehl

Get Jake Schwartz's PPC CD-ROM set. It's a great bargain on a par with Dave Hicks' set. The link on the Museum's Resources page is slightly incorrect, the correct one is http://www.waterw.com/~jakes/index.html.

Ther "rorE" display appeared AFTER shaking the calculator (perhaps a false contact produced a un-sync somewhere), but remained there with the calculator not in movement anymore.

As per some digits (positions) showing faint "2", "1" or "0", I remember it had to do with display and angular modes. What I remember after more than 20 years may be incomplete, and also my English may be not the best to describe the phenomenon, but I think it was (more or less)as follows:

Mode..........Running Display

FIX...........-2#######-##

SCI...........-1#######-##

ENG...........-0#######-##

Something similar happened with the angular mode, but I am not sure about the particular position of the not-so-random digits. Or it may be the other way around...

The "program" to show such display is just an endless loop:

01.....13 01 (GTO 01),

and you should set the display and/or angular mode before running it.

Here's what I found: FIX -> "2" in rightmost digit, SCI -> "1" in rightmost digit, ENG -> "0" in rightmost digit, DEG -> "-" in leftmost digit, RAD -> "1" in leftmost digit, GRD -> "0" in leftmost digit.

Ellis: Thank you for verifying this, I do recall the behaviour but cannot remember the exact relationship between displayed "noise" and calculator modes. By the way, also thank you for a very detailed explanation about HP25 memory chips (other thread).