HEPAX internal EPROM coding. (Maybe too) Techie post... ;-)

[Diego Diaz]

Hi all,

While waiting for the next episode of the Steen's HEPAX Treasure Chest, I decided to go a bit deeper inside my HEPAX module; namely inside its EPROM content.

It uses a standard UV erasable 27C256 chip. Later models use the OTP PLCC version.

The ceramic PCB has pads on its sides that allow reading and even in-circuit programming so I built a "dirty" fixture in order to access the EPROM directly.

   

(Probably unnecessary, but just in case: Do not try this at home! :-D)

I knew HEPAX only use 5 bits (b0-b4) of every byte to build every HP-41 10 bit word.

The way it handles this arrangement was however a bit surprising.

Most EPROM boxes place the MSB on the even byte of the EPROM and the LSB on the odd byte; much like the .ROM file format.

HEPAX on the contrary places all (5) LSbs of a given page on the lower 4K of this page while the upper 4K hold the (5) MSbs.

An example of that may help to clarify.

   

The four banks of the HEPAX code are sequentially stored in the 32K of the EPROM.

Best wishes from the Canary Islands and enjoy your weekend.

Diego.

P.S.: The HEPAX module in the picture above survived the procedure and is comfortably sitting in my CX port 1. ;-)

[SteenPetersen]

Hi Diego,

great - you found one of the features we wanted in the HEPAX module from a warranty point of view. It was a design feature to be able to re-program HEPAX module in case of firmware updates (yes, we still needed to replace the plastic case), however there was an issue (I do not remember what it was - it was hardware related) which made it impossible for us. This is why we changed to OTP in the second batch of modules.

Regards,
Steen

[Diego Diaz]

Hi Steen,

I see a potential source of problems to reprogram the EPROM directly on-board.

Since the EPROM power (Vcc) is not separated for the RAM & ASIC chips, chances are that some spurious Address/Data signals conflicts with those comming from the programmer. But that's just a guess.

The EPROM reading however has worked as expected for me.

I've even written a short program to extract the 4 .ROM files from the EPROM dump .BIN and compared ok with currently available ones.

Best wishes.

Diego.

[Edit] Correction: After checking the 27C256 datasheet it becomes evident that the EPROM cannot be programmed on-board since it has CE/ and OE/ tied together.