Rolm 1602A mil spec computer

[Roland Huisman]

Okay, something new came along... A ROLM 1602A military spec computer... Also called the RuggedNova. I don't have a clue about what I can do with it. The normal 1602 seems to have a octal switch panel with 7 segment displays. There is no front panel with this machine. But this 1602A has a boot ROM and also a 'boot' switch on the back. (or front? Don't even know what the front or back is from this black box :D )

So I think it should work without the front panel. But, there is a control panel interface board in the machine.




Boards:

2023 ROM


5605 CPU (four 2901) and mem

 

[Roland Huisman]

S375 Control panel interface


3310 Terminal interface


3370 printer plotter controller


3564 Data channel controller


3540 Parallel I/O buffer (16 bit)


3364 Mag. tape controller


3503 Real time clock

 

[commodorejohn]

Very neat

 

[Roland Huisman]

I'm thinking about selling this machine because I have enough projects. Anyone interested here? And any ideas about the value?

Regards, Roland

 

[commodorejohn]

What's the size/weight on these things? I'm interested in the abstract, but shipping from NL may be prohibitive.

 

[Roland Huisman]

It is quite small, 20 x 34 x 38cm and weight is 20 KG

 

[Qbus]

That was the rack mounted or stand alone version. Think it once had the local control panel on the front. They are basically a military version of the Data General NOVA running the same code and all the priffs have the same address. I wasted way to much time on them and have several and have paid as much as $500 for one to one that was just given to me.
I had just sent a head for one over to the UK last year but unfortunately do not have a spare head for one today. Do have a head for a 1601 but don't know how hard that would be to adapt. they look the same but don't have any documentation on the 1601
If you do decide to do anything with it I do have tons of documentation on all the 1602 from both Rolm to the last ones built that were built by Loral. They were used in all type of operations from identifying radar threats and deploying countermeasures, decoding and displaying IFF information on ATC screens to my favorite application as the control system for the cruse missal
one of several videos i have on YouTube on that system:


https://www.youtube.com/watch?v=SMH4bGbC77w&t=10s

Qbus

 

[baigar]

Very interesting machine - have some 1602 (without A or B) and later machines, but never even seen the 1602A! This particular one seems quite nicely stuffed with IO cards, but it is lacking memory. The 5605 CPU card does not (as stated above) contain memory...



Later there have been 8k CMOS modules (e.g. in the 1624) which might go into the empty slot of the 1602A shown, but I doubt it is compatible. My guess is, that one needs one of the bricks below to supply memory to the 1602A. Left module contains 8kx16 and ther right one 24kx16 and they can even be stacked. My bet is, plugging both to the rear plugs of the 1602A would be the way to supply 32k of core. [EDIT 20.6.2021] Left module is a Rolm type 2141 (8k) and right side shows a 2142 (16k) - they contain one or three type 2015 8k x 16 board sets. Each set consists of 4 modules: Core, X-drive, Y-drive and inhibit: [/EDIT]

 

[jdreesen]

Nice system indeed, but isn't this system lacking memory ? There should be some core or ram. On the other hand the box seems to be quite full already. Anyone knows for sure what kind of memory should be located where ?

Jos

{EDIT} ..looks like this system will be heading to me...anyone has a spare membox ? ( fat chance I know )
Am i right in thinking that the first 7 slots are I/O slots ? Then we have frontpanel interface, dual width slot for the CPU, and 2 memory slots, one of which has the (boot?)-rom. Followed by 2 DC37 DSUB for memory expansion.
If correct this would open the possibilty to design in a modern SRAM solution in the spare memory slot...

Any schematics / spare boards / software for this system is welcome. There seems to have been an Algol for this, which I would particularly like.

 

[commodorejohn]

It sounds from Roland's description like the CPU board includes some memory, but I'm curious how much.

 

[Qbus]

All the memory was in 8 or 16K core blocks, the 1601 had an external 4K core block that would mount on the case but you are still missing the local control panel. No memory on the CPU cards, the old systems like that had a nine cards that make up the CPU, 5 slots for I/O or ROM and six slots for 8 or 16K of core.
The last systems just used two slots for the CPU so there was internal space for going up to 64k or 32k of core but as far as I know never saw one with static memory. Big advantage to core is that once you loaded a program it stayed regardless of if it had power or not.
On the old systems like that each peripheral is hard wired to a plug, J3 is almost always TT0 to the card in A4, J4 Paper Tape reader to card in A3 and J6 and J7 are I/O Bus.
The real time card you have provides high stability time base information and the tape controller card make me think that was used in some form of test and logging operation or perhaps some form of automated test and evaluation process.
Fun facts for the 1602, the standard system can tolerate temp from – 0 to +65 C with a high temperature version garneted from – 55 to +95 C. they can withstand vibration up to 10g with the shock mount and 2g hard mounted and are rated for normal operation to 80,000 feet above sea level.

 

[jdreesen]

Anyone has a spare papertape interface board, or a source for the IO connectors ?

Jos

 

[Qbus]

There is a 16 bit bus that goes between the control head and the control interface card, the interface card selects the appropriate buffer depending on if it’s a address, data, command or data address entry via the front switches. The only hard connections are for RUN and RESET with the RUN and EXECUTE leds. Everything else is thru the bus.
Will have to back and look up the connector number for the 55 pin plug. There are two different types depending on if it’s a thread (early) or twist lock (Later). All cables are 55 pin to pin.

 

[baigar]

Congrats to jdreesen for giving this machine a good new home - looking forward to hear on the restoration efforts!

Software: Anything written for the Novas should run out of the box on your 1602A. I know of Algol, BASIC,
Extended (Multi-User BASIC), RDOS, ...

Memory - you should watch out for memory bricks as shown in my postng from 2nd of June here. Their Rolm
part numbers are 2141 (8k) and 2142 (16k) they contain one or three type 2015 8k x 16 board sets. Each set
consists of 4 modules (core, x-drive, y-drive and inhibit). So the core sets are different what is needed for the
1602B (type 2011) and what was on eBay recently.

For the paper tape interface you will have to look for a Rolm part number 3320.

Regarding the comment from Qbus - plugs: In all later machines (1602-B and later), the wiring was indeed 1:1
between the IO-modules and the front plugs, but in the early 1602, there where widely varying ones and in these
realy machines, all plugs where of the "Deutsch" type. In -B only the control panel remained "Deutsch" where the
IO-plugs where Amphenol (easy to obtain today). So as the 1602A alread seems to have the later AMP
plugs, I am optimistic, that you will have got the 1:1 wiring already.


Just to those who might be interested: In my collection I have some 1602s up and running (no -A or -B) and I have
a MSE14/Micro also restored which is a smaller machine from the late 1980ties which is compatible to the Eclipse
series from Data General:

1602-Computers:



MSE14 (compatible to DG S-140, 192k of mapped memory):



Finally using a homebuild (together with a friend) hard disk simulator which also siumlates the AD converter
card, I also have got a version of the DG SpaceWar! running with an old Russian oscilloscope as display:

 

[jdreesen]

There is a 16 bit bus that goes between the control head and the control interface card, the interface card selects the appropriate buffer depending on if it’s a address, data, command or data address entry via the front switches. The only hard connections are for RUN and RESET with the RUN and EXECUTE leds. Everything else is thru the bus.
Will have to back and look up the connector number for the 55 pin plug. There are two different types depending on if it’s a thread (early) or twist lock (Later). All cables are 55 pin to pin.

Thanks for this, would that bus be tristate or open collector ? System IO bus is opencollector I believe, but this specific bus may be different. Would love to have part numbers for the 55 pin connectors.
Does not seem too hard to make a substitute for the headunit...
The core will be more difficult, as I would need a complete block, not just the core sets.
But logically speaking, since there is a ROM card in this it should be possible to have a modern replacement with batterybackup, as the addressbus would necessarly be present in the ROM slot.
Good too see that the unit also take 50Hz supply, a simple stepdown transformer should be OK then.

Jos

 

[baigar]

Anyone has a spare papertape interface board, or a source for the IO connectors ?

Jos

Just rechecked my CPI (control panel interface) board, type 1642 and this one indeed has got a paper tape
reader on board. So I wondered, if the S375 supplied with the box here also has got, but comparing the number
of chips with mine shows, that the S375 is a very basic control-panel-only interface without terminal, PTR and PTP
which all are present on the 1642:





Regarding the plugs - as said before, the Deutsch plugs for the panel are un-obtainable), but the IO plugs are
still sold e.g. at Mouser. Having crimp tooling I prefere the crimp variant of the MS27473 series. One example
is the MS27473T16F35P which requires a strain relief (e.g. MS27506F16-2). But also something from Glenair
like the G3703-165-1.75P will also be OK. Search e.g. Mousers page for circular mil plugs, 55 pins, size 16,
insert arrangement 35 and "P" for pins. That should give a selection ranging from USD50 to USD171 (yes, they
are expensive) ;-)

 

[Qbus]

I have a 4 K core block but thought that was for the 1601, never saw a 1602 that used external memory and all had the memory internal but that may explain the 4 K block I have. Both my 1602 have plenty of memory so think I can be talked out of it. Will dig out and take pictures. Where are you located? Just sent a 1602 head to
UK and spent more on shipping then the head cost. Think the head was $150 and the cable $25 but shipping was well over two or three hundred dollars.

If I can figure how to feed big copier at work the fold out drawings I can send you schematic for head and I/O cards. Just for fun this is the pin out for the control interface.
I have a 1666 that I just picked up and just seeing what everyone else has done think getting that up and running may be the next project.

 

[jdreesen]

Thanks for the pinouts and offer for the 4K block. I am located in Switzerland, and will therefore suffer the same shipping issues. I will first examine the machine, check out powersupplies, and get it running with a SRAM board before looking into getting & fitting an original Core board. If I cannot get it running it would be a shame to have relieved you from your spare parts.

Schematics, of course, would be very welcome !

More to follow when I have taken possession of the machine, which is some weeks off.

 

[jdreesen]

This machine is now at my place undergoing a cleanup. Next stage is checking the powersupply...

 

[baigar]

Good luck! I never had a problem with the 5605 CPU boards, so crossing my fingers that you will be lucky as well ;-)