HP Forums

Full Version: HP42s / HP35s Schematics
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

Has anyone compiled (at least partial) schematics of the HP42s or the HP35s?


I take it (by the lack or response) that these schematics do not exist. I will start on the HP35S.

Is there any interest out there?


I think that they are not readily available, however, that master of retro-engineering, Tony Duell might have created the circuit diagrams as well as analyzed the logic of the system.

He has done this on many machines that lack any available original diagrams.

Try emailing Tony directly.

Cheers, Geoff

You will find his address here:

hacker in residence

Edited: 30 Oct 2009, 12:40 p.m.

Hi, Tom;

It is not a lack of interest, I´d guess it is more a lack of possibilities. I opening at least ten HP42S and saw some pictures of opened HP35S´, like this one (by Lyuka). The HP42S has both a controller and a RAM plus a few SMD components, and also provides room for another SMD chip (RAM/ROM?) right? Because of the type of chip it uses, the HP42S is, somehow, 'serviceable'.

The HP35S has some more SMD´s and two chips under resin bubble (not directly identifiable) plus any LCD controller, if any, as we can see. I remember that the HP42S layout was easier to figure out by inspecting, but the HP35S seems to be somehow harder. Because the HP35S has not been what we could call a successful product (at least amongst us here), I see no reason to hack it and try something more than what it already has. On the other hand, the HP42S could only offer a way to add an I/O port, but some guys have been trying to do that for some time. It does not mean some more research is not desired... 8^)

I would go for the hP42S schematics, if I may. And if there is anything I can do for helping you, just let me know.

Cheers and thanks.

Luiz (Brazil)

Thank you all. I haven't contacted Tony (yet), but to my knowledge (via the 'Schematics Disk') these do not exist. I have opened both and am not (yet!) intimidated.

Thanks for your interest. I hope this rises on my priority list.



I do have the reverse-engineered 42s schematics.

I'm currently out, and not sure if the drawing is in one of the CD's/Flash cards I carry with me...

It'll take a while to check out, but will mail back either case.

Cheers from the Caribbean Sea.


Thank you Diego:

Looking forward to it!


Hi there,

Regrettably I have not a copy of the 42S schematics with me. However, if you're not in a big rush, I'll mail you the file as I get back home about Xmas.

Sorry for the bad news.

Regards from Dominican Republic.


FWIW, there are 42S block diagrams in the patent application. They can be found on Eric's site. It doesn't take much to go from there to actual schematic. Once you have one, I really have no idea what you'd do with it... Let's see. One processor, one static ram, a voltage doubler, an IR LED, a capacitor and and LCD. Yep, that's a 42S.

I detailed the keyboard years ago...

That would be a very interesting patent to read, but the patent page at Eric's site seems to have no entries for the HP42S. There's lots of HP32S (not 32SII) entries.

Edited: 4 Nov 2009, 7:17 p.m.

Some of the patent docs appear useful to those who like to take calcs apart or to fix.

The one I had in mind had no model listed. It applies to the original Pioneer keyboard design and the block diagram is of the 42S/17B format hardware.


Thanks, Randy! I just printed it out after seeing your citation.

I'm going to start looking through all the patents. Eric has assembled an impressive collection (Thanks, Eric).

Hi all,

I've checked the patent .PDF and seems to have mayor differences with the actual schematic.

I'm speaking from my mind, but I do not remember any DATA line connected to the keyboard matrix and there are no resistors either.

On the other hand, Randy's diagram and description of the keyboard is (as usual) impecable. :-)

One of the relevant points regarding the 42 ans 17 schematic is that the extra ROM Data and Address wirings are not "alligned" with the ones on the RAM chip. So you can find, for example, that D1 in the RAM is connected to D2 in the ROM...

This is of the main importance if you're planning to write some Saturn code to place in such extra ROM.

I built the 42S-MLDL box in such way that the addresses and data are conveniently arranged to overcome that "disorder".

If there is a general interest in these issues I'll place the documents and pictures in a dedicated page at www.clonix41.org as I get back home.

Best wishes.


Information in patents doesn't necessarily match the details of any given product. The "preferred embodiment" in a patent is a description of how the inventor (or patent attorney) thought the invention could best be recognized at the time the patent application was written.

The fact that some of the patents match actual products exactly, or nearly exactly, might lead one to expect that, but it is the exception, not the rule.

The wiring of the data and address busses of static RAM chips are arbitrary. Unlike a masked ROM, there is no functional difference between any two data pins of a SRAM chip, or between any two address pins. This is true to a lesser extent of dynamic RAM as well. It is typically not true of flash memory due to the protocol used to control writing to the chip.

Edited: 8 Nov 2009, 2:39 a.m.

Hi, there.

I drew an 'HP 42S partial schematic' just now.
The original schematics have drawn previously while making some modification to my 42S.

HP 42S partial schematic

Note that some connectors are ignored and connections to the LCD are omitted.
Though this is incomplete, I hope this helps you.

have fun!



When you found a typo or an incorrect part in it, please inform me.

Hello Lyuka:

That is most helpful!

As I am considering interfacing to the keyboard, the information on your schematic is most useful to me.

However, as I recall, this circuit relies on specific resistance(s) between the column and row circuits.

Have you by any chance measured the resistance for each key in the keyboard?

Thank you,

Hello TomC

The contact resistance of the key matrix had been measured to be 10k~12kOhm(11kOhm typ., Ta=25C)


Hi Lyuka, all...

Excelent job! (to say the least) :-)

One thing that calls my attention was the numbering order of the Saturn CPU pins. After looking at your pictures I realized that the plastic protective frame aroud it has a number "1" stamped in the location you choose to start numbering.

However a closer look to the PCB denotes the "PIN 1" indicator dot just in the corner between VSS and VDD (probably not visible in the pictures but I remember that since I removed said protective frame in the unit I took the schematic from... ;-)

Also some of the not-connected pins are tied together on the PCB... certainly this is irrelevant for technical use; just for the sake of accuracy.

Again my congrats for a nice and clean work.

Regarding the question on the keyboard contact resistance. This is just a design characteristic of the keyboard construction itself and not a functional requirement.

The Pioneer's KB is made of carbon traces, this causes some resistance to be present when a key is pressed; but a clean contact (0 Ohms) is also perfect for operation.

One of the most annoying troubles with heavily used Pionners is the referred CPU lock-up due to carbon dust between frequently used key contacts. Dissasembling, cleaning and reassembling the keyboard is, though possible, not trivial.

Hope this helps.

Best from the middle of the Atlantic Ocean.


Bom dia Diego!

Voçês moram em os Açores? Debe ser uma locação muito linda.

Com os melhores cumprimentos,


(Recommendation to our monolingual readers: try to learn a 2nd language ;) It's fun :)

Hi Diago

IMHO the small dot mark on the PCB might be the orientation indicator for the automated parts mounter. Each parts which have polarity such as diode or solid tantalum capacitor, or specific pin allocation such as RAM, also have the dot mark nearby.

Regarding the key contact resistance, it should be not less than several Kilo Ohm, to avoid signal conflict, since each signal have multiple purpose such as address line to the RAM.
Take a look on the patents page at Eric's site.

On the other hand, the contact resistance of the [ON] key-switch need to be less than several kilo Ohm by my experiment. The PWR-ON input of the processor seems to have some pull-down or require some power, and have threshold to activate the processor.


Also some of the not-connected pins are tied together on the PCB... certainly this is irrelevant for technical use; just for the sake of accuracy.

Good design practice. Many a design has produced errors in high EMC environments due to pins left unconnected or floating ... even if the IC datasheet names them as N/C (no internal connection).

Hi Walter,

Currently I'm in the "other side" of the Atlantic Ocean, in the Dominican Republic. I spent roughly half of my time here and the other half in the Canary Islands... (again in the middle of the ocean... ;-)

Yo no hablo portugués, pero se parece lo suficiente al Español para poder entenderlo o leerlo fácilmente.

(I don't speak Portuguese myself, but its close enough to Spanish as to understand or read it easily)


Edited: 10 Nov 2009, 5:21 p.m.

The schematic was updated. It was found that the PCB of HP-27S and HP-17BII are almost identical to that of HP-42S.

Left:HP-17BII(original), Center:HP-27S(original), Right:HP-42S(modified)


Edited: 14 Nov 2009, 11:17 a.m.


Thank you for your fine work! It is most helpful.

What I am considering is bringing out the row and column signals to the outside world to implement program input to the 42s.

The empty socket pcb pattern makes a tempting access to the address lines to bring the keyboard signals to the outside. I do notice though, that K14(R0) does not seem to be connected to this socket pattern.

I wonder if anyone has tried anything like this before. (?)


I wonder if anyone has tried anything like this before(?)

The 42S Input Project

Search archive 6 for more threads on the subject, the one above was the starting point.

Edited: 14 Nov 2009, 12:55 p.m.

Hi, TomC

K14(R0) does not seem to be connected to this socket pattern.

However, there is a via hole near the processor.

Please notice that the wiring for the key switch are used as the address bus lines of the memory as well.

Additional wiring to the bus might cause some harm such as increased possibility of data error, or decreased EMI tolerance. Take care.



Lyuka / Randy:

Thank you for your responses. I've followed all of the threads (I could find), and it appears that some were close to this, but did not ever complete it. I'm not yet totally confident that I can complete it either!

I am aware of the sensitivity of these lines; certainly keeping the switching matrix within the calculator package would be ideal.



Then, if you really want to switch the key-switch matrix electrically, how about using two 8:1 analog multiplexers such as

MAX4617 or MAX4638 to make cross-point switch matrix?

(A resistor of 10KOhm is needed between its common pin to common pin as the contact resistance.)



On a realted note.....

I did this project a few months ago. A USB interface would be simple to add and it would work on any Classic or Woodstock calculator.