PDA

View Full Version : DG Nova



Qbus
July 29th, 2015, 05:09 AM
As if I did not have enough difficult or just plain imposable projects around the house I am now contemplating diving head long into the Data General Nova way of life, four registers , a program counter and a carry flag! Looking at the internet and this site it looks like the Nova not a big subject or getting a lot of use by collectors, maybe because there are not a lot around and in the past twenty years or so I think I may have seen two total. Anyone know if there is a Nova users group or anything like that? And if I do plod forward into the land of the Nova anyone interested in following the process?

Al Kossow
July 29th, 2015, 05:39 AM
As if I did not have enough difficult or just plain imposable projects around the house I am now contemplating diving head long into the Data General Nova way of life, four registers , a program counter and a carry flag! Looking at the internet and this site it looks like the Nova not a big subject or getting a lot of use by collectors, maybe because there are not a lot around and in the past twenty years or so I think I may have seen two total. Anyone know if there is a Nova users group or anything like that? And if I do plod forward into the land of the Nova anyone interested in following the process?

Bruce Ray
http://wildharecomputers.com/

is the guy to talk to about all things DG

Doug G
July 29th, 2015, 07:54 PM
I'm always interested in Data General talk.

Chuck(G)
July 29th, 2015, 08:31 PM
The Novas were fairly popular as school systems in the early 70s. Pricing was competitive with the small DEC systems. Being 16-bit, rather than 12-bit was definitely an advantage. The Fairchild 9440 was an early Nova-on-a-chip as was the larger MicroNova. The National PACE was patterened after the Nova.

MichaelM
August 2nd, 2015, 04:05 PM
Depending on the direction you decide to take, I would be very interested. I have an interest in PDP 11 having worked with an 11/24 in the distant past. I am aware of the DG Nova and Eclipse processors, and am particularly interested in the simplicity of the Nova design. In fact, I have adapted a part of the Nova I/O architecture as the co-processor interface I am incorporating into my enhanced 65C02 FPGA core.

Qbus
August 3rd, 2015, 12:59 PM
Have agreed to pull the trigger and buy a Rolm 1602B system built by LORAL or as the military calls it an AN/UYK-19 system. I believe this is a Data General Nova system that has been repackaged into a military shell. It has 64K of core memory on four 2011 memory assemblies and the 5605 CPU card along with several other cards for peripherals. Just finished reading the TM 11-7021-201 tech manual and am looking forward to getting the system and what can account for months if not years of frustration in trying to get it going. I do have the local control panel for the 1602 and a fair amount of resources available but have every reason to believe that this will make building Qbus and Unibus systems like a walk in the park in comparison.
I will post more as things happen.

Al Kossow
August 3rd, 2015, 02:08 PM
Have agreed to pull the trigger and buy a Rolm 1602B system built by LORAL or as the military calls it an AN/UYK-19 system. I believe this is a Data General Nova system that has been repackaged into a military shell. It has 64K of core memory on four 2011 memory assemblies and the 5605 CPU card along with several other cards for peripherals. Just finished reading the TM 11-7021-201 tech manual and am looking forward to getting the system and what can account for months if not years of frustration in trying to get it going. I do have the local control panel for the 1602 and a fair amount of resources available but have every reason to believe that this will make building Qbus and Unibus systems like a walk in the park in comparison.
I will post more as things happen.

That certainly wouldn't have been my first choice, but.. whatever

Bruce has all the documentation from Rolm

NF6X
August 4th, 2015, 06:37 AM
Neat! I'm curious about military computer systems, but I've only had experience with militarized PC clones so far. One of these days I need to turn my attention back to my Nova 3 again.

toober
August 4th, 2015, 08:13 AM
Have agreed to pull the trigger and buy a Rolm 1602B system built by LORAL or as the military calls it an AN/UYK-19 system. I believe this is a Data General Nova system that has been repackaged into a military shell.

I think the Rolms are compatible up to the Nova 2, but beyond that, things get funny.

--
Will

dave_m
August 4th, 2015, 09:42 AM
Have agreed to pull the trigger and buy a Rolm 1602B system built by LORAL or as the military calls it an AN/UYK-19 system.

Do you know what power supply it has? Is it a 28VDC system or perhaps a 400 Hz AC system?

Chuck(G)
August 4th, 2015, 10:50 AM
Relevant to the DG Nova architecture, the processor in the Jacquard/Addressograph-Multigraph systems appears to use the Nova architecture as its model. Same 4 16-bit registers and addressing modes, but a somewhat different instruction set (and instruction layout). Also seems to have added a stack for return addresses.

See, for example, The Jacquard J500 Reference Manual (http://bitsavers.informatik.uni-stuttgart.de/pdf/jacquardSystems/j500/JPS500-01_J500_Videocomputer_System_Reference_Manual_Apr7 9.pdf)

Does anyone own one of these things?

Qbus
August 5th, 2015, 05:44 AM
The AN/YUK-19A or Rolm 1602B were available in either a 28 volt DC version or 115 volt AC version. Think the older 1602 was 400 Hz only but the A and beyond are good for 50 to 400 Hz and being that this one is a B the power is not a big problem. The local control panel I have is from an A version and appears to be good from 50 to 400 Hz although I have no documentation on the control itself.
The one I am buying appears to be out of AN/TSC-99 or AN/TSQ-114 ďTrailblazerĒ system being that its configured for the full 64 k core and appears to have the AC power supply but being that the 1602 systems were used in quite a number of systems cannot say for certain. Once I get the system into the shop will be able to tell by the I/O card set what it was used for but hope that it has the fairly standard set of the cards that include the 1642 control panel interface, 3561B I/O card along with the 5605 CPU and the 2011 core cards. The biggest problem so far is that I will need several 55 pin Deutsch connectors for building up the computer to control panel cable and connecting to the I/O port (J7) or any additional peripherals like a paper tape or floppy although thatís way into the future. At this point success would be just to get the local control and serial port up and be able to read and write to memory. Going to assume that maybe one day I can get a copy of paper tape Basic for the DG Nova and then have something going but until then if I can learn some basic Nova would be content to just write simple loops and the like. Never see this as becoming anywhere near as useful as the DEC stuff but have already progressed thru lots of the other military systems like the Grid, SAIC, California Microwave and later Getac in all forms of DOS and Windows and look forward to the challenge of the Nova structure. Also figure this is about the only cheap way of getting into Nova being all this is still less then would be spent on something like a DG Nova4 and donít recall ever seeing a regular Nova for sale in the first place. I have seen years ago the Nova and Super Nova in the AN/TSQ transportable ATC systems where they were used to provide the data displays on Radar screens. They were feed IFF data and drove a CG system that did real time display for flight controllers. The systems were at that time provided with an ASR-33 that had the paper tape assembly and thatís how the operating system was loaded. No drives of any kind.
So at the end of the day this is all a pointless project but one of the things that I have wanted to do for some time. Will see how it goes once I get into trying to find and build up those stupid Deutsch connectors and all there microscopic wiring associate with them.

2571225713

Qbus
August 12th, 2015, 04:32 PM
The local control panel, I pulled it out last weekend and did some cleaning and testing. First row of 16 switches allow setting address or loading data. Second row has selectors to display or change contents of all four registers, allow loading address, data, examine and deposit functions and run and step operations. Had to remove the key cylinder and work over the lock to accommodate a different key but now that’s resolved. The control panel has its own internal power supply and appears to be designed around loading and transferring all the instructions via shift registers to what must be accompanying registers in the control interface card in the system.25791

Qbus
September 29th, 2015, 05:46 AM
The work on the Rolm/Loral 1602 is progressing. So far I know that I have a 1602B with 64K of core and the standard 5605 processor card along with the 1642 Control panel interface card, the 3364 Magnetic Tape controller, a bunch of 3566 Priority load cards that are in effect just dummy cards that carry the signals from one slot to another when you don’t have an option in that slot and finally an 3561B I/O Bus Expander that was intended to carry what I am assuming the data and address busses to a second frame or external system. After cleaning and repairing some minor damage to the system, building a new power cable and doing all the low level power supply test and the like am now at the point of trying to connect the remote control head to the 1642 interface via J1, although the manual (TM-11-7021-201-30) says that all the 55 pin to accessories sockets, A1 to A8 are wired the same have found this not to be true. Slots A1 and A2 are designated for communication being all the cables from that to J6 and J7 are twisted pairs. A3 to A7 are like the manual states wired to plugs J2 to J5 along with J8 and last but not least the A8 where the 1643 control panel interface lives that has its own unique set of connections to J1. It’s also interesting to see that they only use about twenty four wires between the computer and the control panel and not all fifty five that are on the plug. The new issue now is the connection between the fifty five pin J1 on the CPU and the fifty five pin connector on the control head. The manual implies that all communications cables between the CPU and any peripherals are wired pin to pin straight thru but some of my tracings on the control head and CPU have appeared to disprove this. I have only started this work and hope to know more soon. As always if anyone has any ideas or additional data please feel free to share it and if after I get the head to work and get the ability to read and write to memory then the real fun starts in trying to build up some form of communications port. Don’t know why but a lot of this is somewhat like working on S-100 systems only armor plated.

26928

Qbus
September 29th, 2015, 05:57 AM
Additional complications: A Data General Nova has become available, although this would be a Nova three or four going to assume that the basic structure is the same as the original Nova and have to wonder if the system is worth getting just to have a second test bed to try to develop code and test some of the paper tape I have for the Rolm systems?
Its pure speculation on my part but I am going to assume that the Nova three and four are backward compatible or are they? What else makes up the three or four beyond maybe a FPU or expanded memory address? Wonder if they threw in a couple more data registers? And am assuming the civilian systems may also have some form of management port?

firebirdta84
October 8th, 2015, 11:06 PM
I'm always interested in Data General talk.

Doug G and all, I've contributed to another thread where I hope what you say here continues to happen!

I'm discussing the potential similarity between the Microtech machine that I'm working on, and the DG Nova.

http://bit.ly/1Lrc0Pl

Qbus
October 28th, 2015, 06:54 AM
The status so far, after spending much time disassembling the control head, tracing wires and drawing out circuts I now have a fair understanding of how the 1602 control head works with the 1642 control panel interface card in my AN/UYK-19 computer. Somewhat cleaver how they used a bidirectional 16 bit data bus for carrying all the address, data and control back and forth to the system by controlling different latches in the head they make the same bus accomplish all those separate functions and only require a minimum of other discreet control lines. Now have the head all back together along with a connector and cable from the head but still in need of the fifty five pin connector for the front of the system. with this same connector being used on a bunch of military system you would think they would be around but so far have only seen two or three on EBay and they are stupid expensive, see EBay :

http://www.ebay.com/itm/Military-AN-UYK-83-78994-W1-Cable-Assembly-Glenair-805-001-16MT18-55PC-/121664241614?hash=item1c53c0ffce:g:xywAAMXQlgtSxYT x
Or:

http://www.ebay.com/itm/Military-AN-UYK-83-79281-P1-Cable-Assembly-Glenair-USB-DVI-I-O-Audio-VGA-/381277087697?hash=item58c5e2a7d1:g:sPYAAOxyUrZSxX-V

The next step is tracing out the connections from the 1642 interface card to the fifty five pin plug on the front of the system, fortunately it’s only using about twenty five pins not including grounds and if I can come up with the connector will be finishing the connection cable between the two units. One good thing is that in learning how the head works was able to decode much of the instruction set for things like RUN, HOLD, EXAMIN and DEPOSIT along with other functions and now have them in Octal and Binary and will be going back to see how they line up with the same command structure in the regular Data General Nova. Other interesting observation, the Rolm AN/UYK-64 or Rolm 1666 that’s more available disposed of the local control panel and went to all operations being carried out via a con port on the CPU card. Although its pure speculation on my part but think the 1666 is more along the lines of a Nova three or four with the little brother 1602 being more along the lines of the Nova, especially being the 1602 has just four accumulators and the 1666 has a flock of them along with a FPU or math coprocessor. If anyone else wants to abuse themselves with this arcane, pricy and difficult technology there is a 1666 available at:

http://www.ebay.com/itm/Vintage-Rolm-1666B-Minicomputer-Military-Data-General-core-memory-/201455085636?hash=item2ee7a8a844:g:azsAAOSwI-BWLDZj



27557

Al Kossow
October 28th, 2015, 06:58 AM
The status so far, after spending much time disassembling the control head, tracing wires and drawing out circuts

Have you talked to Bruce Ray? I thought he had the tech manuals for the 16xx series?

Qbus
October 28th, 2015, 12:39 PM
I did get a response but not much more. Think he was going to look but no reply so far. That would be uptown having service information! Then I would not have to reverse engineer things that way. The local control panel is the first issue, the next big problem will be trying to build a serial interface and depending on the outcome of how similar the instruction set of this military system is to a Nova that may have a bearing on how to set up the address and interrupt for a serial port in the future. Maybe that way I can use it with a terminal and paper tape system? Long term goal is if I can get that stuff up then try to find the paper tape version of Basic and run that. It’s a pointless goal but a goal none the less.

KC9UDX
November 5th, 2015, 09:59 AM
Not even a very vintage robot.

NF6X
November 5th, 2015, 10:04 AM
I have a vintage ibm system 32 that debuted on january 7, 1975 its very vintage 27674

ANGLETRONICS, you appear to be in the same city as me. I see that your posts here are all about your System/32, but they're all tacked onto inappropriate threads. Is your System/32 for real? If so, you should post pictures of your own machine rather than somebody else's, in a new thread, categorized into an appropriate sub-forum. We can help you if you don't know which sub-forum is appropriate, and I could even come take a look at your machine if you would like, since we're both in Riverside. I have no specific System/32 experience, but I may be able to apply my experience with other vintage computers in some helpful way.



Except that is The Corestore's System/32 (http://www.corestore.org/32-2.htm) and you read like a robot.

If he's a robot trying to build a post count to support spamming us, then a pox upon him and his favorite sports team. But if he's a well-meaning real human who just doesn't know about forum etiquette, then let's help him out. If he's really in Riverside, CA, then I'd be happy to go take a look at his machine and report on what I see.

Al Kossow
November 5th, 2015, 01:19 PM
ANGLETRONICS, you appear to be in the same city as me. I see that your posts here are all about your System/32, but they're all tacked onto inappropriate threads. Is your System/32 for real? If so, you should post pictures of your own machine rather than somebody else's, in a new thread, categorized into an appropriate sub-forum. We can help you if you don't know which sub-forum is appropriate, and I could even come take a look at your machine if you would like, since we're both in Riverside. I have no specific System/32 experience, but I may be able to apply my experience with other vintage computers in some helpful way.
If he's a robot trying to build a post count to support spamming us, then a pox upon him and his favorite sports team. But if he's a well-meaning real human who just doesn't know about forum etiquette, then let's help him out. If he's really in Riverside, CA, then I'd be happy to go take a look at his machine and report on what I see.


Mike sent out a message someone in LA was trying to get rid of one. May be worth contacting to see what's there.

NF6X
November 5th, 2015, 01:22 PM
Mike sent out a message someone in LA was trying to get rid of one. May be worth contacting to see what's there.

I'll send him a PM.

NF6X
November 5th, 2015, 01:24 PM
Hmm, I don't see a link to PM him on his member profile page. Maybe that's disabled for new, low-post-count users?

KC9UDX
November 5th, 2015, 01:36 PM
Hmm, I don't see a link to PM him on his member profile page. Maybe that's disabled for new, low-post-count users?

Yep. ANGELECTRONICS needs to have 10 posts.

NF6X
November 5th, 2015, 01:41 PM
I've posted a visitor message on his user profile. I'll also go respond in a CCtalk discussion that appears to be about the same machine.

Qbus
November 6th, 2015, 05:28 AM
What’s with all this unrelated stuff in my thread???? Anyway I have more to report on the Rolm/Data General front. After much time spent reverse engineering the control head and the I/O systems on the CPU I have learned a couple things. First, despite what I thought originally the control panel interface that’s brought out to J1 does not connect to the control panel. I learned this by tracing out all the connections to the plug on the remote head and then tracing the connections on J1 to the remote interface card and nothing lined up, not just an issue with nothing lining up pin to pin but there were only about half the connections necessary for the head to work on J1. Second, once again I am wrong. The manual states that all the I/O cables are pin to pin, I thought this was wrong. The manual is correct and I am at fault. I discovered that J5 that connects to the I/O interface card had all of its fifty five pins line up directly with the fifty five pins on the control head. The control head uses a sixteen bit bidirectional bus for just about everything except reset and the run light. So I started building a cable to go between the head and J5. Spent three days work in building this and now I have the system up and running, at least with the local control head and can do things like examine the accumulators, read or write to memory locations and lots of other functions that I don’t fully understand yet. Fortunately with the help of Bruce Ray and him providing the operators manual for the 1602 I can move forward with trying to learn some of the basics of Nova operation. That along with building a good looking finished I/O cable once the new connectors get here (MS27473T 16F35P) at $39 each. I intend to investigate the mystery of just what the control interface card dose? Have a speculation that this may have a serial port and a way to get data in and out of the system, all that’s a long way from where I am today, but there is a lot to be said for just having the ability to load and read via the front panel.

dave_m
November 7th, 2015, 11:33 AM
I intend to investigate the mystery of just what the control interface card dose? Have a speculation that this may have a serial port and a way to get data in and out of the system, all thatís a long way from where I am today, but there is a lot to be said for just having the ability to load and read via the front panel.

I am glad you do not need that $500 cable. The control interface card may have been the interface to whatever the real application for the system was. As an AC powered system, perhaps the computer may have been on a Navy ship. I don't imagine that there will be any program/data in the memory. It would have been demilitarized before going to surplus.
-Dave

Qbus
November 17th, 2015, 04:09 PM
Latest picture, have the new fifty five pin plug but waiting for the extended shells for the back of the connectors before final assembly. I have lots more documentation and knowledge then at the start of the project and am looking forward to integration of a serial port and the paper tape hardware that I have. Got paper tape for diagnostics, memory exerciser and the like but the “Holy Grail” right now is the paper tape copy of DG Basic.

27878

dave_m
November 17th, 2015, 04:23 PM
Latest picture, have the new fifty five pin plug but waiting for the extended shells for the back of the connectors before final assembly.

Now this is a Minicomputer! Complete with Octal LED display.

I agree, get the strain relief back shells on before any wires are broken.
-Dave

MichaelM
November 19th, 2015, 04:19 AM
I also echo dave_m's sentiments: a real minicomputer. Great to read about the progress being made with this beautiful machine.

If this machine was used in a Navy application, perhaps the control interface you have written about above may support the NTDS (Navy Tactical Data System) interface standard.

Qbus
November 19th, 2015, 05:05 AM
The manual list several applications along with what the card lineup would be for each:

C-3. LIST OF SYSTEMS
The following is a list of systems that are contained in this appendix:

AN/MLQ-34 (TACJAM)
AN/MSQ-103A (TEAMPACK)
AN/TMQ-31 (FAMAS)
AN/TSC-99
AN/TSQ-84A
AN/TSQ-114A(TRAILBLAZER)
EH-1X (QUICKFIX)
EH-60A(QUICKFIX)

Couples of these are things like mobile tactical switching networks or systems used to locate and identify radar and communications systems. The first 1601 and 1600 series were used to identify and counteract (jam) radar systems that were deemed a threat to aircraft.
The Rolm and later Loral appear to be more centered on use in Army and Air Force applications with the huge ugly computers like the AN/UYK-7 and 20 being more along the Navy lines. Was talking with someone the other day that may be a source for an AN/UYK-44 so maybe that may be the next project?

toober
November 19th, 2015, 07:53 AM
Rolm 1600 series machines saw quite a bit of non-military use as well. They were (are?) fairly popular in the paper industry. Paper mills are absolutely horrible places for delicate electronics, but the Rolms could survive, due to their sealed cases.

--
Will

Qbus
November 24th, 2015, 11:20 AM
I have placed a video of the system on YouTube at:

https://www.youtube.com/watch?v=H0n7YJDdKM8

I am still learning assembler for that system so not able to do anything beyond stomp around in the memory locations and accumulators yet. Have a paper tape reader and several Rolm paper tapes but they are all diagnostics stuff like memory exercisers and the big next project will be some form of terminal for communications. The control panel interface has what looks like a serial port so will start with that next.

Qbus
January 28th, 2016, 09:15 AM
Progress: now have the local control panel working and all that’s good. This is the first binky light and switch system I have owned and worked with and am amazed how easy it is to start to understand loading and examining devices in binary and how you start to get a real feeling for structure and function. Also have the internal communications port (TT) up and running on J-1 The trick to this system is there is no help from the system side, any operations for that port have to be direct addressed via software and never having worked with something so primitive it’s been a real learning experience.