• Please review our updated Terms and Rules here

ROI DCJ11 ISA card

hanshuebner

Member
Joined
Aug 27, 2020
Messages
23
Location
Berlin, Germany
Hello,

a friend gave me this ISA card:

PXL_20221206_074055018.jpg
PXL_20221206_074110260.jpg

Peter Schranz suggested that it could be manufactured by ROI (Rolf Obler Industrieelektronik), but that company seems to have folded before the age of the Internet. He also mentioned that it might be possible that the serial port is likely to be wired as the DCJ11 console and the card could be made to boot into ODT by manipulating the jumpers correctly. JU02 connects to ISA interrupts, and changing JU01 does not cause anything to appear on the serial port, however. I tried two different terminals with all four jumper combinations, and I've sent a break with each of them to see whether that'd bring up ODT.

The card uses a lot of programmable logic parts - GALs, but also two interesting Am29CPL 154H Field-Programmable Controllers. To me, that seems to make reverse engineering pretty much a lost cause. I wonder whether anyone here has seen the board before or has other ideas how one could bring it back to life. Using it as a building block for a little PDP-11 based FORTH would be a fun exercise to me.

Thanks!
Hans
 
The Osprey/PC card has additional I/O connectors on the top an Altera FPGA for the glue logic, so it is different and probably much newer.
 
There was a posting about an ROI board on the CCTalk mailing list a few years ago, CCTalk. (okay, more than a few years ago).

Search for ROI-931007 in the text.

CW
 
@Bitly Thanks for the pointer, I've written a note to the guy to see whether he had any success with the card. I've given it a closer look now and it seems to be a relatively straightforward ISA card with two I/O registers and a memory mapped segment into the PDP-11 RAM. If I manage to get it to run, I'll report here.
 
>>> If I manage to get it to run, I'll report here.

That would be really interesting. Thanks.

Dave
 
I'm a bit farther with analyzing the board, but what I've found does not make me optimistic that I can get it to run.

The card is designed as a coprocessor running in a PC running as the host. The host controls the J11 and presumably provides disk and other I/O services. When powering up the board in a passive ISA backplane, it sits idle with /INIT held low. /INIT is generated by one of the GALs, so it likely needs to be released by the host.

There are two 74ALS996 registers right above the ISA bus connector. They are readable and writeable from the ISA bus - Presumably, one of them is mapped as the Maintenance Register on the J11 side, but I'm certainly not sure. There are also three 74ALS996 paired with three cascaded 74HCT688 comparators on the upper right of the board. This apparently is a programmable 22 bit address comparator connected to the J11 bus. The J11 can write an address to the registers, and a signal is generated on the output of the comparators. Maybe this is some kind of hardware debug register, but how would I know?

There are no provisions for DMA to the ISA bus, so all I/O is controlled by the host. One of four 16bit IRQs can be selected using the JU02 jumpers.

I'll try finding the I/O address by putting the card into a PC, but I first have to set myself up with a suitable machine. In the end, however, the card is probably more of a curiosity due to its limited I/O capabilities (just the console) and dependence on a PC to run it. I might just rip out the CPU and build a system around it that is more fun.
 
If you want a J11 chip I will be quite happy to send you one rather than ripping it out of this board!

Dave
 
If you want a J11 chip I will be quite happy to send you one rather than ripping it out of this board!
Alright alright, I'll try my best to get this board to work, and if I can't, I'll pass it on to the next person who wants to give it a try :) - I quite like the challenge, but figuring out how to release /INIT might be tough.
 
very interesting! i don't think it's an Osprey, Strobe Data seem to be pretty good about marking their products from what i've seen. if you have an NT4/W2k machine you can install their software and see if it is detected, though. there is a hardware diagnostics script that you can use.

Osprey/PC is an interesting product. But, if you have to "Call for Pricing" then you cannot afford it.
i am sad to report that (at least according to my research) Strobe Data seem to have also folded some time ago. i am on an email thread with some of the folks that were there as i have a DCJ11 Osprey myself and they (apparently) do not function without a license- been trying to get that sorted, and also see if they are willing to open source their designs.
 
>>> I don't think it's an Osprey.

Correct. I was looking at the OPs photographs on my desktop machine (rather than my phone) and I could clearly see the "ROI" marking. I couldn't see that clearly on my phone...

We have a load of Osprey's at work. We updated from 11/XX QBUS machines a long while ago. However, we still use the QBUS I/O chassis - so the Osprey is interfaced to the QBUS I/O chassis.

Dave
 
Hello folks,

i did some phone calls to see if i can get some information about the card.
This card has been a special development for automation purposes and to get rid of the not so reliable anymore original DEC PDP11 systems for eventually only one customer with a bigger number of systems.
Without the original software one needs for bringing it up and running again it is just a nice looking brick... :-(((
 
>>> Without the original software one needs for bringing it up and running again it is just a nice looking brick... :-(((

It just requires more work to disassemble how it works and to re-create the software to drive it :). What else is the OP going to do over Christmas?!

Dave
 
Yes -- you're right :)
Maybe the OP has some skills and fun to do it...
I also will try to figure out more if possible - to get the hands on original software would be the best...
 
I have found the I/O addresses now, but the PDP-11 memory does not seem to be mapped into any memory area between 640k and 1024k by default. The ODT debugger starts when the ISA bus is reset, so at least I know that the J11 is running. Even if I can figure out how to control the memory mapping, I will still need to have a host on the ISA bus that implements all of the I/O for the J11. As the bus mastering signals are not connected to the J11 board at all, it will always need to be hand-fed everything from the host, which kind of puts a limit on the time I want to spend trying to reverse engineer it.
 
If it is anything like the Strobe Data card, the PDP-11 memory is not mapped onto the PC bus (I don't think).

What I will do is to see if I can find the data interface for the Strobe Data card from one of our sites and we can see if one company has copied another...

The Strobe Data card had 'stub' drivers (from what I remember) that communicated via a 'bridge' between PDP-11 land and PC land.

EDIT: I have sent an email...

Dave
 
Last edited:
If it is anything like the Strobe Data card, the PDP-11 memory is not mapped onto the PC bus (I don't think).
OK, if the memory is not mapped, how does the host get software into the card? Programmed I/O through the four registers would be an option, but as the MRD/MWR signals are connected to some of the GALs, it seems more likely to me that some form of mapping is going on.
 
The memory could be mapped both ways - either mapped into the PC memory space (expanded or extended memory for instance) or via the I/O registers (similar manner to EXAMINE and DEPOSIT for instance).

The PALs could be used either way, so I would expect the MRD/MWR signals to go via some sort of programmable logic in any event.

You would normally only need to download an initial bootstrap from the PC to the PDP-11 device - and this would normally be a very small piece of code. In my opinion, mapping the PDP-11 memory into the PC memory space (in this case) would seem to be overkill. That is not to say it isn't done this way of course...

Just a thought... If you have the PDP-11 console port working, you could use PDP11GUI to download programs to the PDP-11 from a PC via the console serial port initially. It would be interested to see what devices respond to 'probing' in the I/O space.

I am awaiting a response to my email from site regarding the Strobe Data PC card.

Dave
 
You would normally only need to download an initial bootstrap from the PC to the PDP-11 device - and this would normally be a very small piece of code. In my opinion, mapping the PDP-11 memory into the PC memory space (in this case) would seem to be overkill. That is not to say it isn't done this way of course...
Maybe I'm missing something, but it seems to me that as the card does not have bus master capabilities, all transfers to it need to be initiated from the host anyway. Even an initial bootstrap would have to rely on the host to get the rest of the system loaded into memory. Also, the address bus is completely wired through to GALs, which would not be needed if everything would be going through the I/O registers.
If you have the PDP-11 console port working, you could use PDP11GUI to download programs to the PDP-11 from a PC via the console serial port initially. It would be interested to see what devices respond to 'probing' in the I/O space.
That's a very nice suggestion! I'll be away from my basement for two weeks but will try PDP11GUI when I return. I'll have to waste my time with something more portable over the holidays :D
 
Back
Top