Early 'video cards' for 'personal computers'

[falter]

Hey! I've not been around much due to work, but I have been chipping away at my Youtube channel here and there as I get time. I've got a project I'm working on that will attempt to trace back from an arbitrary point in history to arrive at the 'first video card ever for a personal computer'. To try and simplify matters, for the purposes of the video I've defined 'video card' as a device/circuit offered in a commercial context (ie. for sale directly as a kit or finished card, or offered as a project in an electronics magazine). I've defined 'video' as anything output on an electronic screen (ie. CRT). And I've defined 'personal computer' as a computer a person of average means could purchase for their own personal use, be it home or business.

I'm starting from the early 80s and working my way back chronologicall, featuring cards I find interesting for one reason or another. This is not the complete list, just cards I've found. I thought I'd post this here in case anyone wanted to critique my choices, or offer anything I missed. I'm particularly interested in any offerings from prior to 1976 and have been poring over every periodical or newsletter I can find from around 1974 or so - the earliest card I've found so far was produced that year. I'm doubtful there's anything earlier than that, but that's why I'm asking here.

I've chosen commercial context because I want to eliminate home garage inventors, whose work can't really be date verified and wasn't typically offered for sale. I'm sure someone somewhere in 1972 rigged up a TTL computer for themselves and figured out a way to produce video, but we can't verify that.

Anyway here's my working list so far in no particular order:

OSI 440
IBM MDA
Digital Group Packet #1 Video interface
VDM-1
ECS May 1975 Digital Graphic Display Oscilloscope Interface
SWTPC GT-6144
Cromemco Dazzler
Australian ETI-640
TVT 6 5/8

To qualify for the list it has to be either the first video card ever, or something that came later but was significant in some way (hardware vs software controlled, text vs graphics, etc).

 

[johnx993]

That someone/somewhere would be Don Lancaster and the TV Typewriter in the 70's.

-J

 

[Eudimorphodon]

I’m sure you know this, but IBM MDA is like seven years too new for this list…

That someone/somewhere would be Don Lancaster and the TV Typewriter in the 70's.

This brings up a very good point, which is the distinction between a “video card” and a “terminal”. To pick an example, the GT-6144 was technically something in the netherworld between those things; it doesn’t have a bus connector, it has a parallel port; you basically “printed” to it. A TV typewriter with a hacked connection to a microcomputer definitely would predate it.

Edit: I guess if we’re looking for “first” than I guess it doesn’t matter, I’m pretty sure the GT-6144 loses to several S-100 video boards like the VDM-1.

 

[Eudimorphodon]

To try and simplify matters, for the purposes of the video I've defined 'video card' as a device/circuit offered in a commercial context (ie. for sale directly as a kit or finished card, or offered as a project in an electronics magazine). I've defined 'video' as anything output on an electronic screen (ie. CRT).

FWIW, a more useful definition, or at least one more in line with what we consider a “video card” today, would be the first commercial memory mapped raster-scan video system. X/Y Oscilloscope output is “interesting”, but… among other problems it goes way, way, WAY back, into the 1950’s (if not earlier), so the argument would mostly just be over what counts as a “personal computer”.

 

[1944GPW]

I had the DG-640 VDU (also the PCG board) in my Applied Technology machine. I had read somewhere that this was pretty much the SOL-20 display turned into an S-100 card.
I recall the Target game even said 'Processor Technology' at the top, I had thought AT would have changed that. I think the Dazzler pre-dates it though.

The DEC PDP-11/05-based GT40 graphics terminal had a vector display card in it in 1972, and it has been posited that the 11/05 was a desktop machine that could have started the personal computer revolution three years earlier than the Altair if not for one giant problem - the price. The basic machine was $5000 USD in 1972 dollars.

 

[falter]

I’m sure you know this, but IBM MDA is like seven years too new for this list…



This brings up a very good point, which is the distinction between a “video card” and a “terminal”. To pick an example, the GT-6144 was technically something in the netherworld between those things; it doesn’t have a bus connector, it has a parallel port; you basically “printed” to it. A TV typewriter with a hacked connection to a microcomputer definitely would predate it.

Edit: I guess if we’re looking for “first” than I guess it doesn’t matter, I’m pretty sure the GT-6144 loses to several S-100 video boards like the VDM-1.

The GT6144 is definitely not the winner in this contest - the Dazzler predates it.

And you do have a good point - the line between video card and terminal gets blurry in places. In some of the magazines I've read what could properly be described as a video display unit or video card instead is tagged as a TV Typewriter.

I am trying to leave explicit terminals out like the original TV Typewriter. And anyway that wasn't one card, there were several together to generate the display. And I agree the GT-6144 is kind of influence a grey area. I don't mind having a couple of those just to stir discussion.

The digital group Packet 1 card has sometimes been referred to as a TV Typewriter also. So far I haven't found anything earlier than that. But I am interested in anything notable in between.

The reason I picked the MDA as a starting point is because it's what Google says is the first video card. That was my extremely arbitrary starting point for the quest.

 

[falter]

FWIW, a more useful definition, or at least one more in line with what we consider a “video card” today, would be the first commercial memory mapped raster-scan video system. X/Y Oscilloscope output is “interesting”, but… among other problems it goes way, way, WAY back, into the 1950’s (if not earlier), so the argument would mostly just be over what counts as a “personal computer”.

Yeah that's why I'm supplying my own definition, hehe! I realize it can get into the weeds really fast trying to establish firsts in tech. I've had a guy claim the Apple 1 should be considered the first 'true' personal computer on various criteria. I mean, you could consider an LGP30 a personal computer if you were Howard Hughes.. it ran on 110V and only cost $110000.. That's why I went with 'computer a person of average means could buy'. Sorry Howard.

 

[falter]

That someone/somewhere would be Don Lancaster and the TV Typewriter in the 70's.

-J

Well.. thing is that wasn't designed from the outset for computer use. It was literally just a typewriter for your TV. Other capabilities came from other expansions later. And it's a stack of cards, not one card. Otherwise yeah it's probably the earliest thing to put digital text on an electronic screen in one's home.

 

[falter]

I had the DG-640 VDU (also the PCG board) in my Applied Technology machine. I had read somewhere that this was pretty much the SOL-20 display turned into an S-100 card.
I recall the Target game even said 'Processor Technology' at the top, I had thought AT would have changed that. I think the Dazzler pre-dates it though.

The DEC PDP-11/05-based GT40 graphics terminal had a vector display card in it in 1972, and it has been posited that the 11/05 was a desktop machine that could have started the personal computer revolution three years earlier than the Altair if not for one giant problem - the price. The basic machine was $5000 USD in 1972 dollars.

From what I understand, the Proc Tech VDM-1, which was an S100 card, was integrated into the SOL. Some websites cite the VDM 1 as the official first video card for personal computers. But I think the digital group card, which was developed for the Mark 8 and other 8008 machines, potentially trumps it.

 

[Eudimorphodon]

And it's a stack of cards, not one card.

Strictly speaking the Dazzler is a "stack of cards", so...

I've had a guy claim the Apple 1 should be considered the first 'true' personal computer on various criteria.

I'm going to guess you're referring to that slavering that fanboy that nominated that the IEEE should call it that based on utterly laughably cherry-picked criteria? Whatever, everyone's entitled to their opinion. Even objectively wrong ones.

But I think the digital group card, which was developed for the Mark 8 and other 8008 machines, potentially trumps it.

If you read the "Packet #1" document the Mark 8 device is explicitly compared at several points to interfacing to a TVTI/TVTII, so it's hard to not come away with the conclusion that someone *had* actually done that first, so... again, I think we have to think pretty hard about what makes a video card a "video card" before we hand out rewards, and... I guess, I dunno, this one leaves me really torn.

In its favor:

• it is a thing that makes video come out on a TV screen
• its form factor was a PCB board that was the same size and shape as the other boards in a standard Mark 8 stack

Hits against:

• The original Mark-8 construction article has instructions on page 5 for interfacing it to a TV typewriter. So you definitely cannot claim interfacing "Video" to it was a novel idea... but maybe that's not important.
• The capabilities of the card are very limited; it has no direct cursor addressing and the video can't be read from; it only has a "home" function that resets the address counter to the front of the buffer. In other words, operationally it's just a TV typewriter, its main advantage over hooking up a TVTI would be you can fill it up faster.

I guess I just wouldn't personally put this in the same category as a modern video card, for the same reason I actually don't actually consider the built-in video on an Apple I "modern" either; both are write-only buffers with no random access.

"Modern" video cards have random access to the memory on the card, so... as far as I'm concerned the VDM-1 is probably the winner here; it has 1K of dedicated onboard RAM that the CPU does not have to update in a "serial" fashion. Although if we're going to be picky about what went on sale first things get more complicated, because there were ads at least for the Sphere I computer before I could find any for the VDM-1, and the Sphere I had a fully modern video architecture. I would actually kind of consider it more modern than the VDM in one kind of tongue-in-cheek way, in that it's simpler; it relies entirely on the CPU to move stuff around. The VDM has this weird "windowing" feature for doing hardware scrolling and masking in hardware, which while kind of useful is more reminiscent of a hardware terminal, and most of its descendents/knockoffs (like the Polymorphic video card that came out just a couple months later) dispensed with. I guess we could flip the table and say it's the first accelerated video card...

 

[falter]

Being this is for a video, I'm not entirely against giving my best opinion and then letting commenters duke it out. That's sometimes the best way to find the best answer.

To me the line between a 'TV Typewriter' and 'video card' is that the former can be operated independently (offline) of a computer, whereas a video card cannot. Bryan Blackburn did end up creating a general purpose TV Typewriter out of his digital group Packet 1 board, but I think he had to do substantial modifications for it to operate as a standalone.

I think I'd be willing to include it in my list but with an asterisk. But what I really want to do right now is make sure I've not missed anything else that arrived at the same time or earlier. I don't think so.. but you never know. Still trying to compile a decent list of all pre-BYTE microcomputer periodicals from 1975 backwards.

 

[Eudimorphodon]

But what I really want to do right now is make sure I've not missed anything else that arrived at the same time or earlier. I don't think so.. but you never know. Still trying to compile a decent list of all pre-BYTE microcomputer periodicals from 1975 backwards.

Just going to chuck this out there: the Datapoint 2200, which came out in 1970, was the design basis for the Intel 8008 CPU, and if you look in the hardware manual for it it’s essentially architecturally identical to a Mark-8 with a TV typewriter interfaced to it. (IE, the video system driving its built in CRT was interfaced at a single port address and accepted character bytes that were written into the video buffer at the current cursor position, in addition to offering extremely limited cursor/buffer control commands).

This predates the Mark-8 by four years. Obviously it mostly fails on the “affordable to a mortal person” axis, but if we’re honest and take inflation into account even a Mark-8 was a pretty expensive toy, so… shrug. I dunno. I guess I’m pointing this out to emphasize my gut feeling that there is a meaningful distinction to be made between “terminal I stuck in the same box as the computer” and what really deserves to be called a “video card”.

 

[MiaM]

Interesting topic!

Requiring it to be on a separate card puts up a division where you kind of have both guys in suits and guys with a unix beard on one side arguing about whatever is "serious" computing and what is a "toy", and "youth" on the other side would be people having computers with everything on the same PCB, probably not even any built in expansion slots, and shaking their heads why the other group seem so grumpy, perhaps?

How about each more or less important change in how video hardware were done, no matter if it was a "TV Typewriter"/terminal, a separate card for a computer, built in on the main board of a computer (or a single board computer), and possibly even give a mention for video games?

 

[jlang]

Looking at the evolution of display technology would be a more interesting video than "What was the first modern video card for ancient computers."
With both modern and ancient being poorly defined.

Why select/exclude based on single board ,memory mapped ,read write ,not a terminal?

Did anybody build a TV typewriter just to type to the TV screen? that would get old pretty quick.

 

[vwestlife]

Did anybody build a TV typewriter just to type to the TV screen? that would get old pretty quick.

The Magnavox Odyssey 2 had a "game" which did nothing except display what you typed on its flat membrane keyboard on your TV screen. If you were a kid in 1978, the novelty of that might have been amusing for a while.

 

[Eudimorphodon]

Looking at the evolution of display technology would be a more interesting video than "What was the first modern video card for ancient computers."
With both modern and ancient being poorly defined.

I second this, IE, I think it makes better narrative and historical sense to attack a project like this as a documentary about the evolution of video technology in small computers (featuring historically important milestones, discussing the different technologies, etc.) instead of trying to crown some particular PCB the "first video card". If you want to make it more clickbait-y arrange them as some kind of a tier/top ten list ranking them on how close they are to, I dunno, pick an arbitrary "this is definitely a video card" video card like IBM CGA(*).

(* Asterisk here because I need to nit-pickily point out another wrongness with this:

The reason I picked the MDA as a starting point is because it's what Google says is the first video card. That was my extremely arbitrary starting point for the quest.

CGA was both designed and went on sale before MDA. The prototype proposal doc for the 5150 that's floating around documents a kind of proto-CGA that's an almost exact ripoff of the Apple II graphics system, MDA came later as the "serious business text" option.)

You can measure the "moderness" of video cards on several different axises, which makes putting an exact tier list together complicated. It's nothing like a straight line, but they can to at least some degree be lumped into piles. One of those piles is "TV-Typewriter-like", IE, devices that essentially behave like dumb terminals but are "internal" to the computer. That heap roughly starts at 1970's Datapoint 2200, and ends(*) at 1976's Apple I, with the Mark-8 TVT board conveniently sitting at around the halfway point. If you really want to crown something in this category the "first video card" because that's the narrative you want to paint then, sure, go for it, but I feel like it's kind of ignoring the bigger picture of how personal computing actually evolved away from the "glass teletype" model to something more dynamic. A glass teletype is still a glass teletype even if there isn't a long cable between it and the CPU.

(* Well, when I say "ends", I don't really mean that, nothing really ends, the Apple I is just the last thing like this that people try to posit as being some kind of a first... which, again, is pretty ironic because not only was it way too late, in some of its technical details like the shift register video memory it was an absolute *dinosaur* even for the time. There were a number of S-100 boards that were basically just TVT/terminals on a card that were accessed via I/O ports instead of memory mapping, and by 1978 or so there are CRTC ICs that are essentially "terminals on a chip" that were in use through the 80's in embedded applications and cheap/bad CP/M computers. It's not a useless paradigm, it works fine if terminal level output is what you want, but nobody calls terminals "video cards".)

Another reason why I'm inclined to give the crown (if there is one, which we know there actually isn't) of "first video card" to the VDM-1, besides the technical nitpicking of it being one of the first personal computer video systems (sold as a third party accessory verses coming from the computer manufacturer) that offered direct access to video RAM instead of being forced to treat it like a terminal, is it's widely credited as one of the first personal computer video systems that had real-time action games written for it, like TREK-80 and TARG. (The real-time-edness of these games was, of course, enabled by direct memory access.) Real-time interactivity is one of the things besides physical size and price that *really* distinguishes "personal computers" from their predecessors.

(And like I mentioned earlier, I'd actually put some very closely related boards like the Solid State Music VDB-1 and Polymorphic Systems VTI a half-step higher on the modernity ranking, not only because they dispense with some dumb-terminal-like features the VDM-1 has tacked onto it, making them cheaper and more efficient, they also have TRS-80 style semigraphics at a 128x48 resolution, which made them even more capable for games and other real time displays that, unlike with the Dazzler, were easily mixed with text. They obviously don't displace the Dazzler for its "first", or near it, as a very early color raster display video card, but they are a milestone in the convergence of text and graphics capable CRT displays, which going back to the vector oscilloscope days tended to be different animals, into the unified systems we have today.)

 

[1944GPW]

"offered as a project in an electronics magazine" - tick
"anything output on an electronic screen (ie. CRT)" - tick
"computer a person of average means could purchase for their own personal use, be it home or business" - (University environment) - tick
"prior to 1976" - tick
"work can't really be date verified" - (verified though published article with screenshot) - tick
"somewhere in 1972" - (November 1971) tick
"rigged up a TTL ..." - (less than $200 in unspecialised parts) tick
"figured out a way to produce video" - tick

The published article that satisfies arguably all of the above is something I've posted about before, which I find fascinating for two reasons:
Convert your scope to a display terminal https://archive.org/details/bitsavers_ElectronicignV19N2319711111_83573468/page/n75/mode/2up

First reason is it's a published article with a schematic, description of parts required, principle of operation and a photo of the resulting display. It would very likely have been built on a wire-wrap board for their university's minicomputer which they had enough access to to add this hardware. It does require an interrupt-driven software driver much in the fashion of Don Lancaster's "Cheap Video" board for the KIM-1, seven years later. It pre-dates the TVT by two years.

My second reason is that other people's reactions to this is an underwhelming "meh" which I can't understand as I am unaware of a low-cost build-it-yourself computer video output device article published prior. Maybe no-one else built one, but there it is.

 

[jlang]

I worked on a display in the early 70's that's very much what the article describes.
It was common at the time for system displays to be vector displays rather than raster.
All descrete parts no integrated circuits. It was massive. TWO full racks. about 100 cards.
1K by 12 core memory for the display refresh. The display was a long persistence orange phosphor.
About 12 inch diameter.
It even had the same funny looking 8.
The keyboard... Worst ever I swear. The keys required a couple of pounds to depress. No touch typing here.

It was retired in the mid 70's. I was glad it was. It was a bear to align all the DACs and integrators.

If this project turns into a display timeline your article should be a part. @1944GPW

 

[Eudimorphodon]

My second reason is that other people's reactions to this is an underwhelming "meh" which I can't understand as I am unaware of a low-cost build-it-yourself computer video output device article published prior. Maybe no-one else built one, but there it is.

It's a pretty neat device, particularly in the use of a psuedo-character generator to streamline writing glyphs. I think the drawbacks it has for this particular exercise is:

A: A PDP-8 was arguably a "personal computer" in terms of usage, but it definitely wasn't a "home computer" in price? I mean, I guess you could call being able to play with it at the university a workaround, but...
B: It is an oscilloscope interface, not a raster display. Again, I think they're cool, but they're not really relevant today.

FWIW, the PDP-8 had oscilloscope display interfaces pretty much from the day it was born. For instance, the Type 34 oscilloscope display from 1965. It looks like the original version cost north of $3,000, so yeah, $200 is definitely an improvement.

If this project turns into a display timeline your article should be a part. @1944GPW

I agree. And again, I think this is a great example of how pointless it is to try to declare some arbitrary thing the "first" personal computer video card. The technology was improving and the prices of components were falling so fast from the end of the 1960's through the first half of the 1970's that huge advances were being made left and right, everywhere from the offices of corporate giants, to university labs to random garages and basements. I mean, I get it, people just love heroic inventor narratives that let them point at one guy and declare them the "father of the personal computer" or whatever, but that's just not how it works. Everybody who could get their hands on magazines like this was devouring what they saw and remixing bits from different designs, figuring out ways to do them cheaper (aided by new chips coming out all the time and the cost of everything getting cheaper as volumes ramped up)... etc. The rising tide was raising dozens of boats all at once.

worked on a display in the early 70's that's very much what the article describes.
It was common at the time for system displays to be vector displays rather than raster.
All descrete parts no integrated circuits. It was massive. TWO full racks. about 100 cards.
1K by 12 core memory for the display refresh. The display was a long persistence orange phosphor.

I've said it before (including pretty recently), but it's *really hard* to overstate just what a game-changer MOS memory technology was. Vector instead of raster made sense because 1K-words of a memory can describe a pretty complicated diagram made of line segments; the same number of bits arranged as a raster bitmap display (without a character generator) translates to a whopping 128x96 monochrome pixels. Homemade video terminals weren't rare before 1973 because nobody had "invented" them, they just cost an arm and three legs before then.

 

[pbirkel@gmail.com]

The published article that satisfies arguably all of the above is something I've posted about before, which I find fascinating for two reasons:
Convert your scope to a display terminal https://archive.org/details/bitsavers_ElectronicignV19N2319711111_83573468/page/n75/mode/2up

First reason is it's a published article with a schematic, description of parts required, principle of operation and a photo of the resulting display. It would very likely have been built on a wire-wrap board for their university's minicomputer which they had enough access to to add this hardware. It does require an interrupt-driven software driver much in the fashion of Don Lancaster's "Cheap Video" board for the KIM-1, seven years later. It pre-dates the TVT by two years.

Thanks for pointing out this article. In 1972 we had a DEC PDP-8/L with an AX08 (https://www.pdp8online.com/ax08/ax08.shtml) and a Tektronix RM503 that could be programmed to do this trick; I believe that it was user-developed software distributed via DECUS. Our actual user interface was a M33 ASR Teletype; on-screen text was only practical for limited use in labeling plots. A storage scope would have made it a lot more useful.

Later, from https://bitsavers.org/pdf/dec/decus/programCatalogs/DECUS_PDP-8_Program_Library_Catalog_Jun79.pdf:

SIMBA: A PDP-8/E Oscilloscope Symbol Generator 8-766
Author: Meluyn George Fishel, Free University Brussels, Brussels, Belgium
Source Language: PAL-8, PAL-III, Memory Required: 400 8 words, Special Hardware Required: EAE and VC8E Oscilloscope.
Abstract: SIMBA is a fast, two-page oscilloscope character generator. A 6 x 4 dot matrix is used to generate the symbols. The subroutine takes care of full-line, full-page and end-of-file conditions. Tab characters are automatically expanded.
Media Price Code: 02, F5, G5

EAE: Extended Arithmetic Element, which included hardware-accelerated multiply/divide
VC8/E: https://www.pdp8online.com/pdp8em/vc8e.shtml