Atari's Color Circuits --- Why???????

[reifsnyderb]

I am working on another 800XL board and have it mostly together.  (https://forums.atariage.com/topic/355940-thinking-of-getting-a-new-atari-800xl-board-made/ )  However, I can't get the color circuits working yet.  I have gotten desperate and am going to connect my board to an external breadboard to get this stupid color circuit working.

 

Looking through Atari's video circuits, it's as though they were either clueless or had some sort of drunken party to design the color output on their next computer system board.  I mean I can literally imagine a bunch of electronics guys sitting around a table, with a keg of beer, breadboards on the table, piles of electronics parts nearby, working as fast as possible to see who can get some circuit working first just to win a prize.

 

For example, here's the video for the 800:

I decided to use the 800 video circuits because I got it working on my 600XL, because it has a minimal number of components, and because the S-Video isn't affected that much by the composite video.

 

Here's the 600XL video.  Another transistor and a bunch of other components were added.  S-Video works best without composite video.  However, the colors aren't as nice.  The color section of the 800XL schematic is almost identical.

 

 

The 800, 600XL, and 800XL have a lot in common from the 3.9pF capacitor to the components to the right of it. 

 

Next, we have the 130XE.  A 100uH inductor and 6.8k ohm resistor replaced a lot of components.

 

I am not even going to mention the mess on the 1200XL schematic.  I think that color circuit design party was either a two kegger party or they were playing beer pong before they got started.

 

Does anyone have any insight as to why Atari couldn't use one color circuit design and just stick with it?

 

 

 

 

Edited November 5, 2023 by reifsnyderb

[+bob1200xl]

I think they had two objectives in mind - minimize RF transmission to the 'outside world' (meet FFC regs) and boost the color during part of the frame cycle. The 1200XL had a pretty effective color boost, but that's about it. Other systems were somewhat South of where you would want them to be.

 

The 130XE has no color boost, just too much Chroma.

 

Bob

 

[phaeron]

As someone who knows little about analog circuits but sometimes reads schematics, I wonder if the 1200XL was an attempt to fix the shortcomings of the 800's video output, and everything after that was cost reduction.

 

Firing up an 800, the first thing I noticed is how dim the video was relative to the 800XL and 130XE. The 800 outputs a strong sync pulse relative to the rest of the video, and by my calculations that puts luma 15 at only around 60% intensity. This also results in relatively saturated color. I wonder if the color boost circuit in the 1200XL was meant to try to maintain saturation while fixing the luma range. I actually prefer the 130XE video for how crisp it is, but it also has the most overshoot in the signal compared to earlier models.

 

BTW, there seems to be an error in the jsobola 1200XL schematics, which show the color boost circuit looping back on itself: it has the output of CR19 connecting to Q16's emitter instead of Q19.

[xrbrevin]

i read somewhere that there was no 'standard' for luma/chroma back in the 8bit days, so computer manufacturers made signals to suit their own hardware. they all did it slightly differently so we ended up with a lot of variety!

[reifsnyderb]

On 11/4/2023 at 10:30 PM, bob1200xl said:

I think they had two objectives in mind - minimize RF transmission to the 'outside world' (meet FFC regs) and boost the color during part of the frame cycle. The 1200XL had a pretty effective color boost, but that's about it. Other systems were somewhat South of where you would want them to be.

 

The 130XE has no color boost, just too much Chroma.

 

Bob

 

I would have thought the 800 would have been the best to meet FCC regs. 

 

 

 

[reifsnyderb]

8 hours ago, xrbrevin said:

i read somewhere that there was no 'standard' for luma/chroma back in the 8bit days, so computer manufacturers made signals to suit their own hardware. they all did it slightly differently so we ended up with a lot of variety!

I remember seeing something, written somewhere, regarding how the video sync signal is sometimes positive and other times negative but TV's would work either way.  So, yeah, too much variety.

[reifsnyderb]

Well, I've tried all sorts of changes, on the breadboard, and still can't get color.  I've got the scope running dual inputs so I can look at both luma and chroma at the same time.  They didn't line up.  Then, I checked a working 800XL board and they don't line up either.

 

So far, there has been way too much time spent on this circuit.  (I have been working on this d*&n board since Wednesday.   😞  )  I am seriously thinking of just going with the 130XE color circuit and being done with it.  If I had that 100 uH inductor, I would have tried it already.

[+kheller2]

Is the problem perhaps upstream?  Color Pot? GTIA?

[Rybags]

Not sure.  I think the overall video generation has problems.  Isn't the colour signal constant amplitude?  Which means low luma gets oversaturated and higher ones get washed out.

Then there's the luma issue - certain transitions, esp 7 to 8, can have a black line because the rise time is a fraction of a cycle but fall time is almost instant.

[reifsnyderb]

47 minutes ago, kheller2 said:

Is the problem perhaps upstream?  Color Pot? GTIA?

Color Pot works and adjusting it changes the voltage.  Voltage generated by the color Pot circuit, when the Pot is centered, is about the same as a working board.  Swapping GTIA chips around results in no change.  GTIA chips work fine in a factory board.

Edited November 6, 2023 by reifsnyderb

[reifsnyderb]

23 minutes ago, Rybags said:

Not sure.  I think the overall video generation has problems.  Isn't the colour signal constant amplitude?  Which means low luma gets oversaturated and higher ones get washed out.

Then there's the luma issue - certain transitions, esp 7 to 8, can have a black line because the rise time is a fraction of a cycle but fall time is almost instant.

GTIA Pin 21, the color signal, appears to be the same as that on a working board.  Luma output looks normal.  In desperation, I even changed the Luma back to the normal XL Luma.  Still didn't work.

Edited November 6, 2023 by reifsnyderb

[ClausB]

11 hours ago, reifsnyderb said:

I've got the scope running dual inputs so I can look at both luma and chroma at the same time.  They didn't line up.

What do you mean, they don't line up? Color burst should appear after every horizontal sync pulse with fixed amplitude. Then color info of same or no amplitude and variable phase depending on the display. Can you post scope photos sampled at various points in the circuit?

[reifsnyderb]

2 hours ago, ClausB said:

What do you mean, they don't line up? Color burst should appear after every horizontal sync pulse with fixed amplitude. Then color info of same or no amplitude and variable phase depending on the display. Can you post scope photos sampled at various points in the circuit?

The color burst and horizontal sync pulse were not lining up.  Oddly enough, and after moving the circuit back to the system board, they now line up.  Still no color but the amplitude of the color burst is way too high.

[351cougar]

wasnt the fcc specs in constant flux during that time? certain groups banding together to get them to revise the specs? They had a ;ot of design balls up in the air and ship deadlines to boot. remember Apple for the most part just said f'it and shipped with NO tv out. you had to buy a separate 3rd party add on for tv. not being a commie guy, I dont know , but has anyone done a similar check of the vic20 to c64 to c128 etc video designs?