• Please review our updated Terms and Rules here

Harris RTX-2000 Forth Stack CPU: Need Info

mmruzek

mmruzek

Experienced Member
Joined
Sep 28, 2011
Messages
232
Location
Michigan, USA
Hi, I recently was able to obtain 3 pieces of the Harris RTX-2000 CPU. This CPU has been used on numerous spacecraft by NASA, and in various embedded or development cards for computers in the early 1990's. (For example by Silicon Composers.) I would like to create a PCB to do some development work and tinkering with the devices. I have found the datasheet from Harris, but I have not been able to find any good circuit diagrams or application notes that would give me a quick start on creating something for development work.

I did find mention of a databook called the "RTX-2000 Hardware Reference Manual". There is a radiation hardened version of the chip called the RTX-2010. Does anyone have, or recall seeing some resource that could get me going? Thanks! Michael
 
The RTX-2000 was a version of the NC4000. I believe the major difference was the addition of interrupts. I suspect the bit fields might be different but it still had many of the same things such using two data buses at the same time. Usually refer to as the operation stack and the processor return stack. Both processors could easily implement some version of CMForth. It was an interesting version of Chuck's thinking about Forth.
My experience is more related to NC4000 than the RTX2000. I've not has the opportunity to fiddle with the RTX2000 chip. There are interesting things I've done with my board ( as I recall, it is a Silicon Composers' board but it has been a while ). I've added to it things like a hard drive and floppy drive, using old XT controller cards. One thing that I found annoying is to handle bytes was always a pain. I found the the address -1 was rarely needed in real memory. I made it a byte swap location. Since -1 could be a short constant, one could write and read -1 address in just two instruction cycles. It only needed one more instruction if it was important to remove the high order part of the 16 bit value. For my xt cards is wasn't important, they ignored the high order bits.
Having a working hard drive was great. I could make changes to my version of CMForth and have it working in a short cycle time. I'd do the entire recompile in about 20 milliseconds. I could then run it in RAM to ensure I didn't miss some important part.
I used to call it my single board computer. It was a single plywood board with the computer on top and a power supply under the board.
Dr Ting published a number of books about the NC4000 and I think there was one on the RTX2000 as well. I have most of the series someplace. It started with "Footsteps in an Empty Valley".
Dwight
 
I wish I could find the page that a student at one of the universities was looking at a stack processor and how efficient it was. Compared to an highly optimized RISC or CISC processor, a stack processors are about 1/2 the speed of an optimized processor. So, one might wonder why one would want to invest effort in a stack processor.
There is at last two reasons that most miss. They run at a fraction of the power needed deal with all the tricks needed to make a RISC or a CISC processor work fast and take up a fraction of the space. Now that people have gotten used to dealing with multiple processors, it seems like a waste of silicon to continue down that road but who am I to say what is a waste. They still make a lot of money with those other designs.
Dwight
 
I spent many hundreds of hours in grad school working on an MRI system that was built around an RTX-2000-based controller plugged into a Windows-3.1 machine. It was running a custom version of MPE FORTH from ROM.

I found my PDF of the programmer's manual. It helps clarify the function of some of the pins and peripherals. If you're interested, PM me.

MPE has a FORTH compiler for the chip, as well as a C-to-FORTH compiler, which was used by the MRI system.

It sure would be nice if you had some kind of reference design, though. If this is for hobby purposes, then you should be able to build up an SBC using the datasheet and programmers manual, with some tinkering. If you are doing this for a professional project, then consider getting in touch with MPE. They might be able to get you set up with a development system of some kind.

Dave
 
mmruzek said:
Hello, Thank you everyone for pointing me to some very useful resources! As Dwight mentioned, the links above have some especially useful information about interfacing RAM and ROM memory to the RTX family. I also found 2 online resources for Dr. Ting's work. They are at:

http://forth.org/OffeteStore/4001-footstepsFinal.pdf

and

https://wiki.forth-ev.de/doku.php/projects:ting_s_electronic_forth_bookshelf
Click to expand...

The NC4000 implementation of CMForth would of course need low level fixes for the RTX2000 parts. I don't think all of the bit fields are the same. Most of the operations are the same though.
Dwight
 
You must log in or register to reply here.
Back
Top