Cybervision 2001 (Montgomery Ward)

[MattPilz]

The Cybervision 2001 was an extremely limited catalog-only release through Montgomery Ward from 1978. It was powered by an 1802 CPU and featured 4K RAM (2K video, 2K data), 1K ROM, four color display and a 2000-baud cassette interface.

 

Most impressive were the cassette data types ("Cybersettes" as they called them). They were recorded and read in stereo, where the left channel houses studio audio recordings and sound effects to complement the application, and the right channel would silently load to the system. This enabled them to present what appeared to be much more than 2K of data to the user by continuously loading more data in the background as the voice narration continued. It also had the capability of powering on and off the tape motor to handle user interactivity and program cycles. There were an estimated 40+ applications released for it through the catalog mailing list including games, educational applications, home utilities and visual storybooks.

 

I have been repairing this set for a number of weeks and will eventually make a video to showcase it once done. It has been an interesting repair effort given the complete lack of reference material so is largely reverse engineering and tracing paths to understand how things operate. The son of one of the co-founders wrote two stories back in 2014 that details the history of this system more extensively (Part 1 - Part 2). The core team of five wound up in bigger companies after the demise of Cybervision including for Atari and Sega. 

 

This system is notable for a few reasons. It was one of the few 1802 home computer systems, had full 40-button alphanumeric keypads, and 2000-baud transfer rates were relatively unheard of at the time. It was one of the earliest examples of interactive children storybook telling, where as the stories are read the screen updates dynamically. The following year they put out a remodeled variant of it known as the Cybervision 3001 but few if any were sold. At the following Consumer Electronics Show they demonstrated a working Cybervision 4001 that sported a full mechanical keyboard, integrated BASIC and dual 1802 CPUs...but that model was never released publicly.

 

 

[carlsson]

Hm, this quote was interesting:

Quote

We also developed a prototype of a much sleeker looking 4001 model, which included an extended adaptation of Ron Cenker's floating point BASIC in ROM.  That model was never manufactured in quantity.

It seems all the RCA 1802 designs more or less shared the same BASIC implementation, sometimes adjusted for the target hardware. The original article mentions an integer BASIC made by John and Joe, which might differ from the other implementation.

 

Otherwise this machine doesn't seem to have a lot in common with RCA's various offerings or the later models. The graphics remind me of the Interact Home Computer, a.k.a. Victor Lambda, precursor to Hector but those were Z80 machines. I didn't understand if this CyberVision has any RCA video chips or some other solution.

[MattPilz]

20 minutes ago, carlsson said:

I didn't understand if this CyberVision has any RCA video chips or some other solution.

It uses a Texas Instruments SN76430N Sync and Video Generator. Which has very little information out about, but I found pin-out and specs in a 1978 book on video game repair. That's the only non-RCA chip in it but the rest are common gates and similar. It communicates from the ROM (2708 compatible) to four different 1852 I/O ports.

 

The SN76430N utilizes low-power Schottky TTL technology to provide a color composite video output signal, video summing, and horizontal and vertical synchronization pulses in a complete video game system. Clock reference is a 3.58 MHz crystal. Of course, the chip is TTL and CMOS compatible.

 

The circuit generates horizontal and vertical sync pulses from a 3.58 MHz internal oscillator. The chip also contains a video summer accepting video information from five video inputs (pins 16, 17, 18, 19, and 20). A composite video waveform is generated containing horizontal and vertical blanking, horizontal and vertical sync, color burst background video information, spot video information (players, walls, balls, score, etc.), and serration pulses.

 

20 minutes ago, carlsson said:

It seems all the RCA 1802 designs more or less shared the same BASIC implementation, sometimes adjusted for the target hardware.

That's interesting too. Original promotional material said the 2001 incorporated Tiny BASIC, which if true must've been on another cassette that I've never seen mentioned. It'd be really awesome to learn enough to develop something for this system as an experiment, but there are multiple levels of complexity of that namely that zero reference materials currently exist on any aspect of the data layout.

Edited February 14 by MattPilz

[+FarmerPotato]

That 76430 rings a bell for me.  I saw a TI application note for a "TV Terminal" that had a sync generator chip as part of a set. Might have been that. 
 

I found one mention of a 76410 in this story of the Pong-in-a-chip makers. 
 

IC Master (in Bitsavers) is another place to check. The 76430 gets one line in IC Master 1978. Go back a few years to when it first   appeared, might reveal an application note. 
 

I scanned and posted  the 1982 TI price list yesterday--definitely not in there.   You can see  a consumer electronics/TV theme in the remaining 76xxx chips.  
 

[MattPilz]

Thanks @FarmerPotato (also for that scan!) I found some sequential ICs from another early-80s catalog, which listed these other curiosities that definitely seem catered to a hobbyist DIY game system:

• SN76423N - Automatic Random English
• SN76424N - Wall Generator
• SN76425N - System Regulator and Sync Generator
• SN76426N - Dual Character Generator
• SN76427N - Wall Ball Generator
• SN76428N - Video Game Logic
• SN76429N - Hockey/Tennis/Hand Ball Game Logic
• SN76430N - Sync Generator / Color Generator / Video Summer
• SN76431N - Complex Character Position Controller
• SN76432N - Hockey/Tennis/Handball Character
• SN76440N - Space War Game Logic Complex Character
• SN76442N - Race Car / Rocket Ship / Universal Man Character
• SN76443N - Complex Character Generator
• SN76444N - Rocket Ship / Hockey / Tennis
• SN76445N - Gun Fighter / Universal Man Character
• SN76446N - Exploding Rocket Character
• SN76449N - Exploding Helicopter Character
• SN76460N - Zero to Win at 20 Digital
• SN76462N - Scoring
• SN76477N - Complex Sound Generator
• SN76483N - Space War Obstacles Generator
• SN76484N - Space War Switching Logic

Edited February 16 by MattPilz

[+FarmerPotato]

Wow, that list is fascinating!

 

The article that I found talks about those a bit--not successful, microprocessor era had just begun. 
 

Seeing this list, it puts the 76477 Complex Sound Generator in context. I have a soft spot for this WEIRD little sound chip. It lives on to appear in the 1982 catalog. The 76477 was used in the toy Sound Gizmo. 
 

The 1975 price list (scan coming next) has no 764xx yet. But the 76xxx range is all CON, defined as Consumer Integrated Circuits.

(Exception: 76514 is LIN. But it has the same price as  Linear IC  75514.)  The 1975 price list is compact and doesn't have any chip descriptions. 
 

[Hwlngmad]

On 2/14/2023 at 3:21 PM, MattPilz said:

The Cybervision 2001 was an extremely limited catalog-only release through Montgomery Ward from 1978. It was powered by an 1802 CPU and featured 4K RAM (2K video, 2K data), 1K ROM, four color display and a 2000-baud cassette interface.

 

Most impressive were the cassette data types ("Cybersettes" as they called them). They were recorded and read in stereo, where the left channel houses studio audio recordings and sound effects to complement the application, and the right channel would silently load to the system. This enabled them to present what appeared to be much more than 2K of data to the user by continuously loading more data in the background as the voice narration continued. It also had the capability of powering on and off the tape motor to handle user interactivity and program cycles. There were an estimated 40+ applications released for it through the catalog mailing list including games, educational applications, home utilities and visual storybooks.

 

I have been repairing this set for a number of weeks and will eventually make a video to showcase it once done. It has been an interesting repair effort given the complete lack of reference material so is largely reverse engineering and tracing paths to understand how things operate. The son of one of the co-founders wrote two stories back in 2014 that details the history of this system more extensively (Part 1 - Part 2). The core team of five wound up in bigger companies after the demise of Cybervision including for Atari and Sega. 

 

This system is notable for a few reasons. It was one of the few 1802 home computer systems, had full 40-button alphanumeric keypads, and 2000-baud transfer rates were relatively unheard of at the time. It was one of the earliest examples of interactive children storybook telling, where as the stories are read the screen updates dynamically. The following year they put out a remodeled variant of it known as the Cybervision 3001 but few if any were sold. At the following Consumer Electronics Show they demonstrated a working Cybervision 4001 that sported a full mechanical keyboard, integrated BASIC and dual 1802 CPUs...but that model was never released publicly.

 

 

Very cool and so obscure.  Appreciate you sharing as it is always fascinating to look at all of the various individual computer platforms released before PC compatibles took over.

[MattPilz]

1 hour ago, Hwlngmad said:

Very cool and so obscure.  Appreciate you sharing as it is always fascinating to look at all of the various individual computer platforms released before PC compatibles took over.

Thanks! Amazing story about it is I wound up chatting on the phone with the actual co-founder this week, who engineered the software and OS behind it. Always awesome to get that opportunity! He cranked out the 1802 operating system on a 1K ROM typing manual HEX code into a monitorless DECwriter. Not quite as easy as dragging assets into Unity these days to churn out an app

 

Eventually I will put together a video demonstration of it and some other things, after the very long restoration process concludes.

[carlsson]

I suppose many of these chips equal function calls or data sections in a CPU based system, so you would combine several in different ways to get a certain function. The thought of a chip to select a random English word is fun, but I suppose that is how e.g. a Hangman game in a CPU-less system would work.

[Hwlngmad]

On 2/16/2023 at 5:12 PM, MattPilz said:

Thanks! Amazing story about it is I wound up chatting on the phone with the actual co-founder this week, who engineered the software and OS behind it. Always awesome to get that opportunity! He cranked out the 1802 operating system on a 1K ROM typing manual HEX code into a monitorless DECwriter. Not quite as easy as dragging assets into Unity these days to churn out an app

 

Eventually I will put together a video demonstration of it and some other things, after the very long restoration process concludes.

You are most welcome and looking very much forward to a video demonstration of the machine.

[Bruce Tomlin]

On 2/17/2023 at 5:46 AM, carlsson said:

a chip to select a random English word

I'm pretty sure that's "english" as in ball spin.

 

Is there a link to the ROMs? It's fun for me to disassemble code these days. I've seen a few 1802 things so far, really interesting how they tend to have a few registers dedicated to inline functions, and some kind of byte code or threaded macrocode to support a return stack, etc.

[MattPilz]

16 hours ago, Bruce Tomlin said:

Is there a link to the ROMs? It's fun for me to disassemble code these days. I've seen a few 1802 things so far, really interesting how they tend to have a few registers dedicated to inline functions, and some kind of byte code or threaded macrocode to support a return stack, etc.

Not yet. But the ROM is 2708-pin compatible and when I have a little time I'm going to build a proper dumper for it since my one legacy EPROM programmer that'd support that chip is out of commission.

 

Have you ever disassembled 8080 ROMs? Unrelated but last year I dumped ROMs for an MDT-870 police terminal (as seen in Terminator 2). At some point in my repairs I believe static memory that housed the actual applications got wiped, and the ROM is just now sitting in a perpetual "Initialize MDT" state so I can't even type. Before that at some point I had gotten to where it was just a black terminal screen and at least I was able to type, but every so often it would pop-up saying Radio Disconnected. I blogged my efforts and details on that here: https://mattpilz.com/mdt-870-mobile-data-terminal/

 

 

The photo above is when it was still functional with software somehow engaged in it.

 

 

And this is where it's at now. Here are the ROM dumps: https://mattpilz.com/wp-content/uploads/2022/09/MDT-870-ROM-Dumps.zip

 

The CDVR, I believe is the driver/interface to the CRT, while MDX-ROAM is the main ROM. CGROM was the character generator ROM that was corrupt/wiped so I had to rebuild it without knowing how the original structure was.

 

Edited February 21 by MattPilz

[Bruce Tomlin]

Here's a basic disassembly. I think that chargen rom you should only put the FF underlines in every other character, it's a little easier when you can see it as pixels rather than bytes.

mdt870.zip

 

EDIT: also it would be good to have a thread about this, either on vcfed.org or I guess the "classic computing" section is appropriate enough

Edited March 2 by Bruce Tomlin

[MattPilz]

1 hour ago, Bruce Tomlin said:

Here's a basic disassembly. I think that chargen rom you should only put the FF underlines in every other character, it's a little easier when you can see it as pixels rather than bytes.

mdt870.zip 129.41 kB · 0 downloads

 

EDIT: also it would be good to have a thread about this, either on vcfed.org or I guess the "classic computing" section is appropriate enough

Wow, thanks for that! I'll have to dig more through it when I have the chance and agree I should create a post. I blogged about my experience on my own site at the time but don't think I shared anywhere else. I should also become more active at VCFed as I have quite a few offbeat and obscure systems I don't write much about but am still exploring individually. I'll send you another note if I get a thread created here or there.

 

From your disassembly of the ROAM.BIN, the sticking point on launch seems to be:

 

011A:    CD1D03            CALL    L031D
011D:    215C01            LD    HL,D015C

 

Where D015C refers to the "INITIALIZE MDT" text I see on screen.

 

But at some other point I was able to get the "NO HOME CHANNEL IN LIST" or similar, it was just a blank terminal I could type in but that message would pop up intermittently. Some amount of "button mashing" as I was cleaning the keys when it was powered on seemed to had wiped it and now it just is stuck on that initialize screen. But also I definitely lost the rest of the software that must had been stored in some static RAM or chip that got wiped, as it had utilities for plate checks and so on including the same utility seen in Terminator 2 and see no semblance to them anywhere in the ROM. Elsewhere in the disassembly (2F0A) you can see it says "PRESS C TO ENTER RAD. CHANNELS" but no key input has any affect. It's been a while since I delved into that project I know I had ran the code through some disassemblers and tried to parse through it but 8080 assembly is really foreign to me compared to Z80. (Just realized the disassembly is in Z80, which is more convenient for sure).

 

Back on the Topic of Cybervision...

 

I had a major setback while working on diagnosing and troubleshooting some of it after I posted last. The cassette deck impacted the logic board and seems to have shorted some thing out. Initial scope checks show concerns around the ROM to CPU data bus where one output of the ROM is stuck on high and the other outputs are very abnormal looking and no sign of life with or without the ROM in, just garbage screen. My obvious concern was with the ROM that it did not get corrupted since i hadn't been able to dump it yet. I managed subsequently to dump it using three different solutions and at least do see vaguely recognizable 1802 opcodes in there. In fact there are a few matching bytes at the same locations as to the original Cosmac VIP ROM. I plan to burn the data to a 2716 if I have one and lift the necessary pins (and have some other 2708s on their way) and am just overall hopeful nothing got corrupted on it before dumping. I haven't had enough time to meticulously test a lot on the board since the failure occurred, but will have more time this weekend.

 

I'll update this post when I can. 1802 assembly is something very unusual to me even to make heads or tails of how things work. But I believe the ROM basically initiates a video driver out of sorts, then has to handle cassette signal and control as well as display the "Cybervision" text on screen. That's about it, 1KB worth.

 

 

Edited March 2 by MattPilz

[Bruce Tomlin]

39 minutes ago, MattPilz said:

(Just realized the disassembly is in Z80, which is more convenient for sure).

It's in "Z8085", because Intel's opcodes are more annoying. You can tell it's 8085 because of the vectors for RST 5.5/6.5/7.5.

And 1802 code tends to be weird. I've disassembled two or three big things, and the code tends to have a few registers dedicated to call/return and bytecode processing, shifting in and out of native code.

[MattPilz]

Thanks @Bruce Tomlin - Here is the world premiere 1KB dump of the ultra elusive Cybervision 2001, ROM. With no other system or chip dump to compare it to, I can only cross my fingers no bytes are corrupt. Otherwise I'm out of luck!

 

I dumped it using a homemade 2708 EPROM reader, a TL866II (using PC power supply described here) and a 1980s EP-1 programmer on a 386 and all three matched. Just wish I had done this when I had a known working system before the hiccup, so would be curious what you can make of the output.

 

Also pasted below as plain text. I know one other person who has a ROM-functional Cybervision but I think they'd be reluctant to mess with pulling chips or dumping to try and get a second version for comparison. 

 

Around #0250 it becomes a lot of INC data, "19 1C 11 15 10 15 18 1C" etc... I know too little about 1802 to make any helpful observations yet, but presume this is a candidate for byte data / draw data, which somehow ends up drawing the logo "Cybervision" on the screen while awaiting cassette data. The ROM also, somewhere or other, checks for invalid cassette data or valid data and will display two other messages including "CHECK TAPE" or similar.

 

The system does not have any dedicated character generator, the only other chip is that TI video chip (SN76430N) that can receive five video inputs to compose. But it does interface with four CDP1852s and one CD4034B I/O chip.

 

But also one of the software engineers involved indicated the tape motor start/stop is controlled by passing a series of hex A5s to the system, but I don't see any A5s in this code (there is one 5A, however.)

 

F8 90 B0 FE BC F8 FF BF F8 17 BF F8 15 BB 9C BB 
F8 91 B1 F8 31 B1 F8 93 B3 F8 B9 B3 F8 18 B4 9C 
B4 F8 93 B5 F8 BC B5 F8 50 B9 F8 30 B9 D1 F8 91 
B1 F8 55 B1 F8 93 B5 F8 BC B5 F8 1B B4 F8 DA B4 
F8 93 B6 F8 D1 B6 D1 F8 90 B8 F8 D0 B8 FD 7B 30 
5A D4 D4 D4 D4 D4 D4 D4 D4 D4 9C BA F8 17 BD F8 
FF BD D8 F8 B5 F3 32 90 9A FB FF 3A 72 F8 1B B4 
F8 DA B4 F8 93 B6 F8 F2 B6 F8 91 B1 F8 55 B1 D1 
D4 F8 91 B1 F8 3A B1 D1 30 10 D4 D4 D4 D4 9C BA 
D8 F8 B5 F3 32 9D 9A FB FF 3A 72 30 7D D0 93 F2 
D8 F0 B6 D8 F0 B6 D8 F0 B7 D8 F0 B7 D8 F0 BE F6 
D8 97 5F 96 FF F3 32 BC F6 70 30 B0 97 5F 96 F3 
3A B8 FD 9E 3A 9E 7A 30 F5 D4 14 36 34 98 D4 D0 
3D D0 7A 1A 30 DF 98 98 98 98 98 98 98 98 98 98 
98 98 98 98 D4 FF 9B 73 9B 73 F8 D0 73 F8 FF BF 
FD D0 17 FD F8 7D B0 D0 D4 D0 F0 FE 9C D4 D4 D4 
3B 14 FC 11 B6 F0 FE FC 10 F6 B6 F8 91 BE F8 D4 
BE FE 96 F3 32 1C 70 9E FB FC 3A 12 FF 9E FF D4 
5F F4 B6 FD D0 90 F9 D4 D4 D0 79 3E 3B 79 30 39 
D0 94 B8 94 B8 99 BA 18 F3 F2 F5 F1 58 18 3A 9A 
3A 37 94 FC 50 B4 3B 5C 94 FC 11 B4 39 99 3A 31 
FD 30 30 16 D0 94 BE 94 BE 16 7E 33 53 56 B7 F8 
92 B7 57 B8 17 B8 9C BA F5 58 F1 54 14 58 F1 54 
14 58 F1 54 FD 1A 9A FB 13 3A 94 9E FC 13 BE B4 
9E B4 30 59 94 FC 3E B4 3B 78 94 FC 11 B4 30 78 
D0 F8 18 B4 99 B4 34 D4 99 BE 3E 9E 32 90 94 FC 
50 B4 3B 9A 94 FC 11 B4 30 9A D0 9C 3A B1 F8 11 
BC F8 BF BE F8 BF BE F8 F0 FC 11 3A B9 9E FC 11 
BE 3A B7 9E FC 11 BE 3A B4 9C FF 11 BC 9C 32 BA 
30 B1 D4 D4 17 1E 16 36 1D 15 35 1C 14 34 18 10 
30 50 10 19 11 31 51 11 1A 12 32 52 12 1B 13 33 
53 13 54 14 55 15 56 16 1F 17 37 57 10 10 10 58 
92 50 92 59 92 72 92 7B 92 74 92 7D 92 96 92 9F 
92 98 92 B1 92 BA 92 B3 92 BC 92 D5 92 DE 92 D7 
92 F0 92 F9 92 F2 92 FB 93 14 93 1D 93 16 93 1F 
93 38 93 31 93 3A 93 53 93 5C 93 55 93 5E 93 77 
93 70 93 79 93 92 93 9B 93 94 93 9D 93 B6 93 BF 
19 1C 11 15 10 15 18 1C 10 12 15 10 10 15 10 18 
1C 10 19 1C 11 19 1C 10 1C 1C 14 19 1C 11 18 1C 
11 18 1C 10 15 10 15 1C 1C 15 10 10 14 1D 1C 14 
1C 1C 11 18 1C 10 19 1C 11 17 1C 11 18 1C 10 1C 
1C 15 10 35 10 10 14 10 19 1C 11 19 1C 11 18 1C 
10 19 1C 11 18 1C 15 18 1C 10 19 1C 11 1D 1C 15 
14 10 14 1D 1C 11 1D 1C 11 1C 1C 10 19 1C 11 15 
10 10 18 1C 10 1D 16 10 15 10 15 1C 34 10 1D 1C 
14 1D 1C 10 1C 1C 14 1D 1C 14 1D 1C 10 14 10 10 
19 1C 11 15 18 15 18 1C 10 15 10 15 1D 1C 15 14 
10 14 18 1D 10 10 15 10 18 1C 10 10 1E 14 11 3A 
10 18 34 10 15 12 14 1D 11 10 14 30 14 15 10 10 
15 10 10 1C 1C 14 17 12 15 15 10 15 14 10 14 17 
10 15 15 18 15 14 10 14 1D 1C 15 15 10 15 1C 1C 
14 1D 1C 11 1D 1C 10 14 10 10 1D 1C 15 15 10 15 
1C 3C 16 1D 1C 11 1D 1D 10 14 30 14 19 1C 11 18 
1C 11 18 1C 10 1C 1D 14 10 15 10 10 14 10 15 10 
15 15 10 15 18 1C 10 15 10 15 15 10 15 30 34 10 
15 10 15 15 15 15 1C 1C 14 15 10 15 32 36 10 14 
10 14 15 10 15 30 35 10 10 14 10 1C 1C 15 12 34 
10 1C 1C 14 3D 33 3E 3F 32 3F 3F 3B 3F 10 15 10 
10 14 10 10 14 10 10 10 10 10 10 10 14 10 10 10 
10 10 10 10 10 10 10 10 10 10 50 90 D0 3F 7F BF 
FF 5E 5E 18 54 16 1C 36 3E 34 38 34 30 3E 5E 5E 
5E 98 5E 38 3A 14 3E 1A 5E 16 54 5E 5C 5C 5C 5E
5E 98 18 32 1C 18 38 5E 3A 32 1C 5E 3A 14 32 1C
5C 98 F8 DA B4 D1 F8 90 B8 F8 D0 B8 D8 D0 90 B0

 

cybervision2001.bin

Edited March 2 by MattPilz

[carlsson]

@etxmato might be interested in this too, perhaps eventually add support for the Cybervision 2001 in Emma 02?

[etxmato]

@carlsson Definitely interested to add a Cybervision emulator in Emma 02!

 

I have been wanting to do so since 2016 when I first heard mention of the Cybervisions, but minimum requirement is the ROM and it looks like the above is indeed corrupt somehow. Data bit 4 is always 1 which is suspicious but there might be other issues. Also need to dig into this video chip to be able to get going...

 

Cheers, Marcel.

[Bruce Tomlin]

On 3/2/2023 at 3:02 PM, MattPilz said:

I can only cross my fingers no bytes are corrupt. Otherwise I'm out of luck!

Sorry to tell you this, but it looks like the 10H bit is stuck high on every byte. I was wondering for a while why there were so many PHI instructions and no PLO.

[MattPilz]

1 hour ago, Bruce Tomlin said:

Sorry to tell you this, but it looks like the 10H bit is stuck high on every byte. I was wondering for a while why there were so many PHI instructions and no PLO.

Hi Bruce! I was just about to post an update here. The high bit makes sense, as that corresponds to what I observed when checking the ROM outputs with my scope while connected to the machine. Which is definitely unfortunate, but echoes what Marcel said as well and what I found upon converting the HEX to binary where it was very obvious.

 

But... Now the good news:

 

The amazing folks in the COSMAC Elf group including @etxmato have been tearing through it this past week and through a collaborative effort tonight I was able to use the modified ROM and get back to the boot screen with good and bad program data also being detected appropriately.

 

The mostly corrected disassembly from Marcel and the group can be reviewed here: https://groups.io/g/cosmacelf/attachment/40510/1/Cybervision.asm

 

A fellow with a second known Cybervision may also be working to get a dump this weekend with assistance from VCF, and if so that will be a great reference to compare to this repair effort.

 

Thank you for the help as well on this and the MDT-870!