Programming Bipolar PROMs

Brent Hilpert bhilpert at shaw.ca
Mon Sep 27 12:52:18 CDT 2021


On 2021-Sep-27, at 8:23 AM, Tom Hunter via cctalk wrote:
> While restoring and repairing a Data General Nova 2/10 I found a bad
> bipolar PROM on the CPU board. The PROM has open-collector outputs and is
> organized as 32 words by 8 bits. It appears that one of the open-collector
> driver transistors is faulty (but it could also be that a fuse has
> "healed").
> 
> The part is an Intersil IM5600CP, but these were also made by others, for
> example Signetics and Philips made the 82S23 and TI and NTE made the faster
> SN74S188N. Some vendors still sell these parts and there are even a few on
> Ebay.
> 
> How do I program these PROMs? I found one somewhat obscure description of
> the algorithm in the NTE datasheet, but I suspect that each manufacturer
> had (somewhat) different algorithms.
> 
> Is there an affordable commercial programmer out there which can program
> these PROMs?
> 
> Is there a simple design out there which I could breadboard for a one-off
> programming job (maybe using an Arduino to control the programming
> sequence)?


The hardware & software for a one-off breadboarded programmer is, or should be, easy enough with a ~ few-hours/1-day effort if you wanted to go that route.

I did this 8 years ago for a PROM in a HP9830 - a 256*4 Intel 3601 PROM, for which I burned, and replaced with, a TI 74S387.
(In the end in turned out it was an unnecessary effort, one data-line from the PROM was showing bad but the failure wasn't the PROM output, it was the input of the IC the PROM was driving.)

Programmer hardware was 3 TTL ICs, 2 zeners, 6 transistors (more specifically 2+numDataLines) and a few resistors on a solderless breadboard, and a bench power supply for the programming voltage. Software was a python program running on a RPi. All program-pulse timing was done in user-land on the RPi.

I can send along the schematic and program if you wish, there will be some adaptation required of course. Details of the 74S188 programming/burn algorithm is in the 1975 TI Memory Databook for Design Engineers (available at bitsavers) (I expect the 188 is the most likely type you'll find NOS today).





More information about the cctech mailing list