FB2 S-Video Output?

[Hornpipe2]

I'm buying a new TV with an S-Video input and I'm wondering if it's possible to mod the FB2 to output Y/C signals. There's really nothing quite like extra-crisp giant pixels, and I'm looking for new cables for all my systems (or soldering new cables for the ones that aren't available already!).

[Math You]

I know what you mean about crisp giant pixels! They sure look good when they're clear and colourful

 

You can buy a Composite to S-VHS/S-Video adapter for around £6 / $12 in the UK. It might be worth searching Google or an online auction for 'Composite to S-VHS adapter' and see what there is near you.

 

There are also some converters available though these cost a lot more.

 

Matt

[Hornpipe2]

I'd rather mod the FB2 if possible, since I think video signal is already degraded when converting internally from S-Video to composite. (Many devices have separate choma/luminance on board and then mix them into the composite signal later).

 

I haven't gotten a chance to look at the innards of my FB2 yet, but if anyone has taken a good peek at the board, I have a good idea of what we're trying to do. Take a look at this page (about adding S-Video to the C64DTV): http://galaxy22.dyndns.org/dtv/common/svideo/index.html

 

So if the FB2 really mixes those two signals together, we can separate them before it happens. Anyone who's opened your FB2 up, can you tell if there is a capacitor merging two signals together? Or does the chip output already-mixed composite signal?

[policelimit]

As well as just the modders, Curt probably knows too.

[Hornpipe2]

Anyone? ...Curt?

[Math You]

It might be worth taking a look inside the FB2 to see if you can follow the composite video output back to the video chip. If it goes to 1 pin then chroma and luma are mixed together. If it leads to 2 pins then chroma and luma will be separate.

 

There is the possibility that separating the two signals will weaken the signal strength and reduce the brightness of the picture.

[nyder]

I will be doing the cart mod on mine soon, so while I'm doing that, I will see about getting S-vid &| RGB.

 

mainly since I'd be using it on a Commodore/Amiga monitor with seperate chroma & lum rca plugs. Well, it also has a db9 for rgb, so if I can get rgb out of it also. =)

 

I also got a BK Precisions 15MHz Triggered Sweep Dual-Trace Oscilloscope this morning, and am waiting on someone who's dropping on an old Tektronics Vacuum Tube Oscilloscope, plus a couple other things.

 

Of course, I don't know how to use those yet, but thats what getting them was for.

 

Craigslist is so nice. Ask and you shall receive. =)

 

So, wht I'm saying in my usual round about way is that there's very little left to stop me from working on this stuff. I guess i'm just excited. =)

[antron]

I will be doing the cart mod on mine soon, so while I'm doing that, I will see about getting ... RGB.

 

not possible.

[supercat]

I will be doing the cart mod on mine soon, so while I'm doing that, I will see about getting ... RGB.

not possible.

 

I don't think it would be impossible, but it would be an interesting technical challenge. The 2600 would be easier than the A8's and the 7800 because each chroma cycle just as one pixel rather than two.

 

The TIA generates a three-bit luminance output (which is straight binary--the SECAM machines use this signal directly), and a chromanance output which should nominally be a square wave shifted relative to the chroma clock. To identify the hue of a pixel, one must identify how much its chroma signal is shifted. Modern analog or digital circuitry should not have much trouble measuring this.

 

If one wanted to stay mostly in the analog domain, one could use two triplets of integrating amplifiers (red/green/blue), one for even pixels and one for odd pixels. A couple of timing circuits would subdivide each pixel into three subintervals, called "red", "blue", and "green"; a simple divide-by-two counter would distinguish between even and "odd" pixels.

 

During the "even" pixels, the red, green, and blue "even" integrators would be allowed to integrate during their respective sub-intervals. The "odd" integrators would be sampled during the red time and then discharged during the green time. During "odd" pixels, the roles of the even/odd integrators would be swapped.

 

Basically all analog (except for the even/odd "counter") but it should work reasonably well for a 2600 or for a FB2.

 

A more interesting digital approach would be to use a CPLD to intercept/monitor/adjust accesses to the TIA chip so that it would always output a luminance of x00 for the background, x01 for the playfield, x10 for player 0, and x11 for player 1. Writes to COLUxx could be latched into four seven-bit registers, thus allowing a seven-bit digital output which could be turned into any desired color format. Such an approach would be especially useful if the output needs to be converted to another video format, since it would allow the signal to be captured digitally.

 

Capturing the TIA writes would probably require that the CPLD watch A12, A7, A5-A0, D7-D1, phi2, and RW (17 wires). D1 and D2 would need to be intercepted (2 wires), and L0-L1 would need to be captured (2 wires). There would then need to be a data output (7 wires). So 28 wires total--a nice fit for something like a small-package 95C72XL (the logic wouldn't quite fit in a 95C36XL).

 

Actually, I rather like that approach, though I doubt there'd be enough market to justify designing and building such a thing. Further, it would only work if it's possible to intercept signals going to/from the TIA. No problem on a socketed-TIA 2600, but not likely to work in an FB2.

Edited May 12, 2007 by supercat

[nyder]

um, wow.

 

I was figuring since it's not the actual atari 2600 chips, and most chips deal with color in the rgb realm, that you could tap it before it got converted to composite.

 

 

If it doesn't deal with it like that, or there isn't any practical place to tap it, then I would think getting S-Vid would be no problem because it already has a Black & White/Color switch on it.

And since the easist way to do black & white & color on the same display is to split them into Chroma & Luma, or, S-Vid.

 

So it would seem to me, that the Black & White/Color switch just toggles it to either combine the signals, or just send Luma down the composite cable. So all you would need to do is tap the chroma to a seperate plug on the unit, and then use the Black & White/Color switch to Black& White, then have the cables going from there to your montior/tv. Of course you probably are going to need to make a special cable for it.

Edited May 14, 2007 by nyder

[BigO]

um, wow.

 

So it would seem to me, that the Black & White/Color switch just toggles it to either combine the signals, or just send Luma down the composite cable. So all you would need to do is tap the chroma to a seperate plug on the unit, and then use the Black & White/Color switch to Black& White, then have the cables going from there to your montior/tv. Of course you probably are going to need to make a special cable for it.

 

It is my understanding that the B&W/Color switch has no connection to the video circuitry. It is simply an input that is acted upon by the software at the whim of the programmer. It can be used for purposes other than changing the displayed colors. Testament to that is the fact that many games do not exhibit color differences when switched between the two modes.

 

EDIT: Yeah, here it is in the schematics - http://www.atariage.com/2600/archives/sche..._2600_High.html

The B&W/Color switch is attached to a line in a port on the 6532 "RIOT" chip. Don't know the addressing scheme off the top of my head, but the switch state can be read as a bit.

Edited May 15, 2007 by BigO

[nyder]

um, are those the schematics to the Flashback 2? I don't know if CX2600 was the protoname or something.

 

Well, if the switch doesn't do what I would of had it do (after all, isn't how I said the easist way of switching between black & white and color? probably not...), then I guess that theory won't work.

 

I won't know more until I open it up and take the time to poke around and look.

 

I'm not really sure S-video would help the graphics anyways. I mean, after all, how many different colors can it display on screen at once? The picture looked great on my commodore 1902 monitor as it was.

But hey, getting better vid signals out of it would be something to do.

 

I'm trying to get my place cleaned up and my electronics workstation setup. got a lot of stuff to do and no where to do it. =(

[BigO]

um, are those the schematics to the Flashback 2? I don't know if CX2600 was the protoname or something.

 

Well, if the switch doesn't do what I would of had it do (after all, isn't how I said the easist way of switching between black & white and color? probably not...), then I guess that theory won't work.

 

I won't know more until I open it up and take the time to poke around and look.

 

I'm not really sure S-video would help the graphics anyways. I mean, after all, how many different colors can it display on screen at once? The picture looked great on my commodore 1902 monitor as it was.

But hey, getting better vid signals out of it would be something to do.

 

I'm trying to get my place cleaned up and my electronics workstation setup. got a lot of stuff to do and no where to do it. =(

 

That schematic is for the original 2600. The FB2 appears to have been built to be functionally as much like the original 2600 as possible. It was designed with support for a cartridge slot to play the original games. If the switch was not functionally the same as on the original 2600, I think there'd be little point in including it given the minimal use of B&W TV's these days.

 

The switch doesn't directly affect the video in the FB2. There aren't many games on the FB2 that do any color change in response to the switch. You don't even need to launch a game to satisfy your curiosity about this. Just power up the FB2 and see that the initial menu doesn't respond to the B&W/Color switch.

 

 

BTW - if you want an inexpensive, functional, but cosmetically imperfect FB2 to experiment with, or even just a semi-functional one to peek into, I can tell you who has some at a reasonable price (not me).

Edited May 18, 2007 by BigO