Signetics 2650

Actually the CPU was designed between June and August of 1972. The management of Signetics had forecast the sales at only 5000 units. The production was then set back to January of 1973. What has always confused me is why the date was set at 1975 when it was first produced in 1973? I think if you look at the design based on 1973 rules, it receives a much better score

After Jim Hall said no, I tried to continue the project further. There was no money for lab technicians, but on the tty I could do some work. I wrote the compiler that you may use. We had at that time some IBM 360's in the building and I used them. If you have used the compiler for the 2650 that runs on the 360 or the 1130 that is mine.

The Signetics S2650 was the CPU in my first home PC which I built from a kit back in 1980. I liked the instructions set and found it really easy to write software for - it also helped that I had access to an assembler on a CDC Cyber at work.

A year or so ago I had a bout of nostalgia and rebuilt it from scratch. The original board design was the Signetics S2650 'Mini computer' published in the May 1978 issue of Electronics Australia. It only had 4k of RAM. It used a cassette interface and a home-built VDU with a 20mA current loop interface. I've now adapted it to use an RS232 interface. It sure is a lot quicker now being able to use a PC with terminal emulator and cross-assembler and upload the hex files from the PC than it used to be typing in hex codes and storing on a flaky
cassette :-)

I managed to obtain a copy of the relevant copies of Electronics Australia off eBay and was astounded to discover that RCS Radio in Sydney still sells the PCBs:

http://www.rcsradio.com.au/
Mod edit 21/02/2013 - RCS Radio seem to no longer be in operation, their website is down, other sources list them as out of business in 2012.

Sourcing the chips via the Internet was not too difficult. I bought a batch of half-a-dozen of the CPUs just in case some didn't work but all did so I'm currently selling some of the spare 2650s on eBay to help pay for the project. The hardest part to get hold of was the PipBug monitor ROM which I couldn't find anywhere. If anybody knows of a source I'd be interested to hear about it.

However, having the source code meant it was possible to burn an EPROM.
Unfortunately the ROM and equivalent EPROMs are not pin-compatible
so I had to make a daughterboard containing an EPROM as a plug-in
substitute. I also adapted the the circuit to use an RS232 port -
amazingly (me being primarily a software person) it all worked! A
photo is here:

http://www.cfbsoftware.com/s2650/s2650.jpg


Edited by terrapindundee - 21 February 2013 at 3:58am

Originally posted by Sig2650 Any interest in this CPU? I am the designer.

Hi,

my first post in this forum just joined today

Ok long story made short, I am Giles from Llamasoft, work as a software developer and have a very strong passion ( illness ) for electronics so during my ( small ) spare time I put things together.

An even longer story made short an ex videogames installer once left me a bag full of "strange chips", that bag been sitting for around 20 years in a space until now.

So I started on a mission/project to put back to life some of that "." HW because "it can't just sit there".

Other long story made short my eyes in this moment felt over this SG2650 chip, I have a couple of those plus even a 2636, so literally a bit there and there during "late nights" I put this together, totally invented/designed from scratch purely using the SG2650 datasheet I found around.



It's basically a signetics attached to 32K of ram, 8K of rom and a 6850 Uart and an IDE interface, so far working at 1 Mhz, this thing has literally 6 days of life about I am still debugging it.

I am writing "my own version" of PIPbug with XMODEM support to upload stuff in.

Not sure yet what to do of it but "something" will come out.

When I'll be finished with all the documentation I'll put it somewhere online in case someone could be interested.

I still have to :

- check a strange thing I am seeing with the ram
- try to put the CPU at 2 Mhz and the UART a 1Mhz using some OPACK logic
- test the IDE interface
- finish to port ( from Z80 ) my "bootloader" in it

I do all this at home for pure fun, there's no commercial intent or so behind, but really is "I want to see those cpus still working".

So far I have the rudimental base of my bootloader SW working, most important I have the uart working I can send messages out on the terminal, I still have a "very weird bug" looks like ( I know it does not make much sense ) the last 223 bytes or so of RAM are failing my rudimental ram test.

Not sure what it is yet but I'll find it out.

So there it is, if anyone is interested let me know

Very nice design. You did that board yourself?