HP-25C Eumulator

[rprosperi]

(04-20-2017 07:02 PM)PANAMATIK Wrote:  The transition from HP-25 to HP-25C by HP was ingenious and made a very fast time to market possible at these competitive days. The firmware remained bit by bit identical, even the "Clear Memory" instruction, which cleared the external RAM chips at initialization, was left in place. There were only two hardware changes made: the two NMOS RAM chips were replaced by two pin compatible CMOS RAMs, and battery plus was connected to a previously unused pin of these RAMs. As the NMOS RAM chips itself (and not the ACT processor) decoded and performed the "Clear Memory" instruction, the CMOS RAMs decode unit was modified and just didn't execute this instruction.

The HP-25 firmware used the two RAMs (each containing 56 bytes) for storing 49 program steps plus 8 registers and the LastX register, 49+9*7=2'56=112 bytes, and placed the stack and display format inside the ACT chips volatile memory, the HP-25C consequently could not save the stack and display format after power off, because only the external CMOS RAMs were made non volatile.

I may allow to mention: The HP-25E since Rev 1.10 newACT replacement (HP-25 repair kit) also saves the stack registers and display format together with up to 111 programs and 88 registers in nonvolatile memory, and keeps their state and data even when batteries are removed.

Thanks for explaining that Bernhard, I have wondered why the items preserved isn't simply 'everything'. It never occurred to me they could have created the 25C from the 25 with no f/w changes, only h/w. And let's face it, they did get all the really important stuff. Ingenious indeed!