• Please review our updated Terms and Rules here

PD32 NS32016 computer

Slightly off-topic, but there aren't many 32016 threads!

Having visited the web page previous highlighted Definicon DSI-32 co-processor emulation (some time ago), I downloaded the file "ns32032_dsi-32.7z". There's quite a lot of interesting stuff in there.

My interest in the NS32000 family dates back to 1985 when I worked for Nat Semi. I was really impressed by the architecture and was able to get my hands on a few parts. I built up a wire wrap board which had a small amount of EPROM and RAM (16k bytes). I programmed a trivial serial boot loader into the EPROMS. UART was a NS16450. I wrote an assembler on my Atari ST in Pascal, but then like many projects it got put on hold. While in Santa Clara in 1985 I did hear the rumour that Atari had been evaluating the 32016, but as we know they made the fatal choice of the 68000. I would have loved a NS32016 based Atari!!! Anyway a year or so ago, I got out the board and played with it. I had to disassemble the original boot loader and found a bug in the boot loader which limited the amount of bytes to be loaded to 256, and another bug introduced by my assembler. I tried to write a simple second stage loader, but I really needed a server on the PC to download the binary (I was having problems with Tera-Term). Before giving up I generated the schematics in KiCAD.

In the past 6 months I have had a number of PCBs manufactured by JLCPCB. These are to do with my 68000 project to run EMUTOS, interest initially triggered by a 68008 project called "Katy", which I tried to build on a breadboard. My real intention is to build a NS32016 board. My 68000 system uses a 40 pin bus and I have a CPU card, a RAM card, and a FPGA card (which does Atari video and serial comms). I use an Mega Arduino to load RAM using DMA (there is no ROM). The intention is to replace the 68000 card with a NS32016 card. In theory (if I could find a working NS32016 compiler) I could run EMUTOS on the NS32016 :D .

Anyway over the past few days I have been playing with the files from the above "dump". In the "set021387" I had success in running the emulator and getting it to compile and then run code. I then tried the GCC compiler. It worked for trivial examples but crashed when compiling Drystone code. BTW The same code compiled using the emulator. So I then fired up a Linux VM, copied over the GCC sources folder and spent an afternoon trying to get the makefile to run. There were a number of problems, including use of I assume now depricated functions. Eventually I got gcc and cpp to compile. While the cpp appears to work, cc1 (the compiler) crashes (aborts) during code generation. I compiled using debugging and there is nothing obviously wrong where it crashes, just an "abort()" in the default of a case statement. So I have no idea how to fix this.

Anyway hello to all the other NS32000 fanboys reading this thread. I am very interested in any projects anyone is still working on using this old CPU.

--migry
 
There was good reason that the NS32K found little acceptance in the US market, though it did find some application in the embedded (e.g. laser printer) segment.

I first learned about the NS16032 from a couple of fellows from the NSC marketing group about 1980. No silicon yet, of course, but what they presented was breathtakingly advanced over even the brand-new Moto 68K. Given that this was the same outfit mostly known for linear ICs and TTL jellybeans, as well as licensed replicas of other manufacturer's work. (the PACE and SC/MP were pretty much obsolete by the time they were introduced), I asked if this was really an in-house NSC design. The two allowed as how it wasn't. Okay, that figured.

So, when would they start sampling pre-release silicon? Un, don't know, but maybe within a year. In reality, by the time NSC had production going on the 16032 (later renamed the 32016), it was late 1982 and the result was buggy as hell. By then, the world had moved on with either Intel or Moto MPUs. There were some very bright spots (e.g. 32532) in the product line, but it was too late.

NSC, true to form, eventually lost interest in the line and ended development and the product line before the decade was out.
 
I have various documents for the NS32016 TDS and DB32016 Multibus card.

Does anyone know if the EPROM binaries can be found anywhere?

I have wanted to disassemble and patch this code to work with my own homebrew board for quite some time.
 
The ETH Ceres workstations were based on this chip : Ceres 1 uses the NS32032, Ceres 2 the NS32532 and the Ceres 3 uses the NS32GX32.
All these are Oberon workstations. Detailed design reports are available on the ETH FTP server.

Jos
 
Since my last post, I did finally read the article about the PD32. Personally I wouldn't want to use DRAM, 1) because the controller is hard to find (possibly meaning it is expensive) 2) there are large SRAMs which would be easier to wire up and would take up less board space. Also although the parallel port interface does make it possible to interface the board with other non-ISA bus computers, the fact that ISA has gone the way of the dinosaur means that there is no advantage to using this interface. One idea would be to use a dual channel FTDI chip, using one half as the serial port for user text I/O, and the other half as parallel I/O for easy interfacing. Anyway just my ideas.
 
Question for the mods (who have to review all my posts anyway - so feel free to answer by inmail and not post).

I have been doing a lot of work on NS32016 software and hardware projects over the past month. I would like to post details for those few who might still be interested(!), but I also have various questions.

Should I open a new thread or threads? Some questions are hardware and some software. Any suggestions as to whether to just continue this thread, or which sub-forums would be most appropriate to use for "NS32106 misc software thread" and "NS32016 misc hardware thread".
 
Hi, probably a necropost, but thought maybe there is still some interest in a PD32 NS32016 revival. It looks like the Dave Rand PD32+ project came to naught, but I've updated the KiCAD capture of the original PD32 NS32016 board from 2017. By default, it is an 8-bit ISA board, however, the design supports a generic 8-bit interface and includes a schematic for a Z80 interface. 3D render image here. Still interested in doing a PD32 project. I think a PCB is doable but need a small team for initial build and test (debugging) and software development. If you're interested in NS32K, even if not the PD32 revival, please respond here or at retrobrewcomputers forum.

 
I built a PC532 years ago and had Minix running on it, I didn't end up using it much and I no longer have the hardware.
 
Hi
Looks like a project page on Hackaday for Public Domain 32000 is starting to take root. Recent activity and people signing up

 
I remember discovering this family of microprocessor chips last year, but I couldn't find any of them available anywhere.
 
Project seems to have stalled, in that I don't see comments at Hackaday or RetroBrew since mid-February -- but Andrew kindly posted schematics, Gerbers, etc on Hackaday.
 
Back
Top