Designing a new Atari Video mod

[ReilyReed91218]

Howdy, so basically I've been designing a new atari video mod based around some old schematics and testing the values myself. Unfortunately my girlfriend sold all of my testing supplies, soldering iron, etc for drugs just recently (needly to say I'm done with her) and I can't breadboard the blanking signal amp at all. I was wonder if the community could give me some advice on whats needed here in this picture. Don't know if I got it right because I couldn't finish my design. 

 

Thanks everyone.

Reily

 

Image wouldn't embed for some reason but here's a picture of the schematic and a circle around where I was going to start testing designs a few weeks ago, but can't any longer

 

https://imgur.com/8kA93VF

 

 

 

[the_crayon_king]

6 hours ago, ReilyReed91218 said:

Howdy, so basically I've been designing a new atari video mod based around some old schematics and testing the values myself. Unfortunately my girlfriend sold all of my testing supplies, soldering iron, etc for drugs just recently (needly to say I'm done with her) and I can't breadboard the blanking signal amp at all. I was wonder if the community could give me some advice on whats needed here in this picture. Don't know if I got it right because I couldn't finish my design. 

 

Thanks everyone.

Reily

 

Image wouldn't embed for some reason but here's a picture of the schematic and a circle around where I was going to start testing designs a few weeks ago, but can't any longer

 

https://imgur.com/8kA93VF

 

 

 

Well that's unfortunate...

 

Anyway, I thought the chroma couldn't be buffered. Since chroma is analog?

When I say 'buffered' I mean you can't use a hex buffer on C; you can use an op-amp/transistor.

 

I think blanking is just an offset applied to C., so you can buffer blanking then feed it to C.

However, I don't see how blanking applied like it is on the atari could wipe C if it was transmitting., which is how I usually see blanking used (like an OE pin). Where Cout would be 0 when blanking is active even if Cin had voltage.

Anyone else feel free to illuminate that area because I do not know.

 

You are going to have a hard time getting measurements without an Oscope (im assuming if you had one she stole that too).

I guess you could build the circuit in here: https://www.falstad.com/circuit/

Just concentrate on the Y video portion for now. I believe the scale for the resistors should stay more or less the same.

You should be seeing 700-714 p-p for video and 300 p-p for sync (sync can be a higher P-P if you don't want to stay super close to spec because if its too small then you won't get video).

 

Here is an example of Y: http://tinyurl.com/y2j7totq

This can be used if your buffer goes from near 0 to near 5V any deviation from that messes up the p-p;

right now the P-P goes 400, 200, 100 and 300 for sync. I am not sure how that FMS IC will like the impedance with that so you may need a transistor or op amp to buffer to 2X gain (6db).

I am also not certain if the resistor ladder I made is correct but its sums up to 0.7 for Y and 0.3 for sync or 1V overall; which is what Y+sync is. If you want to change the scale you can do it with a pot at the op-amp/transistor end.

Basically, we are making a DAC in the simplest way possible; there are better ways to do it.

 

You can test Y just by itself first to make sure you get video. Then start messing with C.

I would buffer blanking and then apply it to C through a resistor., just like Y the FMS IC probably will not like the impedance so you may need to buffer with an op-amp/transistor. C should probably have a resistor as well. I don't know how you can get it into spec if you don't have an Oscope. I don't know the specific voltages output there. I guess just eyeball the video output and hope you don't break anything. The FMS is sensitive to voltages above about 2V p-p (I think?) so make sure you are going from low to high not high to low.

 

If you got rid of the FMS and just wanted S-Video you could use the THS7314/THS7316 as it doesn't seem to have an issue with impedance. It's also tolerant to 5V so you might not have to worry about breaking something., maybe your TV but at least not the breadboard.

 

Any of this helpful? Anyway, I am sure there is someone on here with more knowledge about this system than I have.

 

EDIT:

It appears like the FMS can go +0.3v above VCC or -0.3V below ground.

I am confused about that whole thing. Either way you shouldn't be feeding 5V signals to either IC. Or in the case of the FMS exceed 1V p-p; I have definitely melted one before.

 

Edited October 4, 2020 by the_crayon_king

[ReilyReed91218]

On 10/4/2020 at 5:31 PM, the_crayon_king said:

Well that's unfortunate...

 

Anyway, I thought the chroma couldn't be buffered. Since chroma is analog?

When I say 'buffered' I mean you can't use a hex buffer on C; you can use an op-amp/transistor.

 

I think blanking is just an offset applied to C., so you can buffer blanking then feed it to C.

However, I don't see how blanking applied like it is on the atari could wipe C if it was transmitting., which is how I usually see blanking used (like an OE pin). Where Cout would be 0 when blanking is active even if Cin had voltage.

Anyone else feel free to illuminate that area because I do not know.

 

You are going to have a hard time getting measurements without an Oscope (im assuming if you had one she stole that too).

I guess you could build the circuit in here: https://www.falstad.com/circuit/

Just concentrate on the Y video portion for now. I believe the scale for the resistors should stay more or less the same.

You should be seeing 700-714 p-p for video and 300 p-p for sync (sync can be a higher P-P if you don't want to stay super close to spec because if its too small then you won't get video).

 

Here is an example of Y: http://tinyurl.com/y2j7totq

This can be used if your buffer goes from near 0 to near 5V any deviation from that messes up the p-p;

right now the P-P goes 400, 200, 100 and 300 for sync. I am not sure how that FMS IC will like the impedance with that so you may need a transistor or op amp to buffer to 2X gain (6db).

I am also not certain if the resistor ladder I made is correct but its sums up to 0.7 for Y and 0.3 for sync or 1V overall; which is what Y+sync is. If you want to change the scale you can do it with a pot at the op-amp/transistor end.

Basically, we are making a DAC in the simplest way possible; there are better ways to do it.

 

You can test Y just by itself first to make sure you get video. Then start messing with C.

I would buffer blanking and then apply it to C through a resistor., just like Y the FMS IC probably will not like the impedance so you may need to buffer with an op-amp/transistor. C should probably have a resistor as well. I don't know how you can get it into spec if you don't have an Oscope. I don't know the specific voltages output there. I guess just eyeball the video output and hope you don't break anything. The FMS is sensitive to voltages above about 2V p-p (I think?) so make sure you are going from low to high not high to low.

 

If you got rid of the FMS and just wanted S-Video you could use the THS7314/THS7316 as it doesn't seem to have an issue with impedance. It's also tolerant to 5V so you might not have to worry about breaking something., maybe your TV but at least not the breadboard.

 

Any of this helpful? Anyway, I am sure there is someone on here with more knowledge about this system than I have.

 

EDIT:

It appears like the FMS can go +0.3v above VCC or -0.3V below ground.

I am confused about that whole thing. Either way you shouldn't be feeding 5V signals to either IC. Or in the case of the FMS exceed 1V p-p; I have definitely melted one before.

 

What an excellent idea, thanks a lot for the info. I know of a different way of amplifying Blanking on chrominance. But I don't really like the idea of doing it that way any longer (using a 3.3k pullup on blanking with 470Ω of termination before chrominance. 

 

Anyways, I'll give the THS7314 idea a shot and see if I get better results at all. 

[+batari]

From my analysis of the COL (Chroma) line on the TIA, it is actually a digital signal, even if in NTSC it is technically analog. On the 2600 schematics, this line appears to have a couple of simple, cascaded RC filters that likely are designed to smooth out the signal to something more sinusoidal. However, on video mods we typically pull the signal directly from the TIA, so we are actually likely getting something closer to a square wave, so should be able to buffer it.

 

Because a square wave can be represented as an infinite series of harmonic sine waves, a low pass or band pass filter can eliminate these higher harmonics and basically turn a square wave into a more of a sine wave. So there is a good chance that the FMS chip can accept a square wave and will filter out the harmonics from the COL line, as it supposedly has 4th or 5th order filters (much better than whatever is on the 2600, which appears to be second-order, and not a great second-order filter at that?)

[cwilkson]

 I don't come on here very often anymore, but I'll try to give some quick help.

On 10/4/2020 at 6:31 PM, the_crayon_king said:

I think blanking is just an offset applied to C., so you can buffer blanking then feed it to C.

However, I don't see how blanking applied like it is on the atari could wipe C if it was transmitting., which is how I usually see blanking used (like an OE pin). Where Cout would be 0 when blanking is active even if Cin had voltage.

Anyone else feel free to illuminate that area because I do not know.

 

The COL signal from the 2600 is a (poor) square wave.  It can be buffered or not, but it *must* be filtered for best chroma results.

Ideally chroma is a sinewave with 0v offset....it goes positive *and* negative.  Loosely stated, you need to use filters to absolutely kill everything below 3.5MHz and kill everything above 3.6 MHz.

 

Blanking controls the amplitude of chroma -> color saturation.  The amplitude should be bigger when /BLANK is high.  Otherwise the colors look pale and washed out.  The way Atari did this was to use an added resistor to reduce the amplitude of COL when /BLANK is active.  Unfortunately, this also adds 2 nice big vertical edges to the COL signal which shows up as a single pixel wide vertical colored line on each side of the visible display.  These will be different colors, and one will probably be shown more strongly than the other one.

 

On 10/4/2020 at 6:31 PM, the_crayon_king said:

You can test Y just by itself first to make sure you get video. Then start messing with C.

^THIS^

Absolutely make the Luma works first!  This is quite easy to make arbitrarily good.  If you want to, you can make it look "perfect".  And it's not difficult.

 

Chroma is hard.  On the 2600, it's really hard.  On an NTSC 2600 it's ***REALLY FRIGGIN HARD***.  It can be done but there is a brick wall where you can't make it look any better and there still are obvious visual artifacts.

 

You can use magic bullets like the FMS series of devices or you can go old school with transistors, opamps, logic gates, etc.  Or somewhere in between.   Whatever choice you make just try to get the Y and C signals from the Atari as clean as possible before you try to rebuild them into proper NTSC waveforms.

 

 

[Shawn]

On 10/4/2020 at 12:34 PM, ReilyReed91218 said:

 Unfortunately my girlfriend sold all of my testing supplies, soldering iron, etc for drugs just recently (needly to say I'm done with her)

 

 

 

 

Why would you include this in your first posting on a forum to a bunch of strangers?  

 

[the_crayon_king]

10 hours ago, Shawn said:

 

Why would you include this in your first posting on a forum to a bunch of strangers?  

 

Flavor-text.

[ReilyReed91218]

So i am using the THS in the end and it works perfectly and ideas on how to matrix composite through a series of passive components? And yea, she did. Stole all my shit. And sold it. Sorry if that caught you guys off guard.

[ReilyReed91218]

NVM using an op amp for composite. Might just redo my design and use op amps instead.

[ReilyReed91218]

So I need an expert in opamps to help me out. Using a 250MHz op amp to amplify Y and C and to matrix the signal into composite. Here's my schematic so far with resistor values that I believe to be correct. Not certain how to configure this for composite. Please do help out. If other could try this, I'd be ecstatic.

Edited October 15, 2020 by ReilyReed91218

[cwilkson]

Are you familiar with LTspice?  Or another analog simulator?  If you don't want to do the algebra to design the resistor networks and amplifier circuits then using a simulator will help you a great deal.  It's usually not a good way to go...simulator is only to verify the behavior of a completed design before you purchase the components.  But maybe you can use it for experiments?

 

What is the intent of C2 and C3?  Their only effect here is to block the DC input to V+ of the opamps.  This won't work.  The V+ input is basically floating.

Also, the intent of the 2 pots is not clear.

 

I suggest you share a block diagram of what you want to build.  Then we can offer better advice.  Also label the inputs on your schematic.  I can understand what they are by analyzing the circuit and/or looking at the other image but it's much easier if it's all on one diagram.

 

 

 

[+batari]

8 hours ago, ReilyReed91218 said:

So I need an expert in opamps to help me out. Using a 250MHz op amp to amplify Y and C and to matrix the signal into composite. Here's my schematic so far with resistor values that I believe to be correct. Not certain how to configure this for composite. Please do help out. If other could try this, I'd be ecstatic.

When you mentioned op-amps, I figured you were going to use the op-amps as filters. I see you are using them as low impedance drivers for the output signal, and to provide gain in your circuit. From my observation it seems you only need the gain because the 75 ohm resistor to ground on the COL side and the pot on the LUMA side may be attenuating the signal a lot.

 

From what I have seen, in general, video mods do not need a whole lot of gain (if any at all) as we are basically starting with 5V digital signal and we want to end up with 1v p-p or something like that. So what I have seen more of is circuits that use a simple BJT amp configured as an emitter follower, which at best provides unity gain, but that may be enough. It is very simple to implement while still having the benefit of being able to drive low output impedance you need.

[cwilkson]

As batari said.  You need an output stage with an overall gain of less than 1.  (5V logic -> 140IRE NTSC = 1Vp-p into 75 ohms)  You can get that in many ways, including the resistor attenuator + compensating amplifier like you are using.

 

It depends on how "correct" you want the video output signal to be.  If you just want something that will work with most TV receivers, then a transistor buffer amp (with correct resistor network on the input and output) will do.  Or even just a naked resistor network.  If you want something that meets all the NTSC level and driver specs for any given amplifier device (Vbe varies a lot), you'll need more transistors.  If you're trying to include controls for brightness, contrast, and saturation and still generate a "correct" waveform then you'll need a lot of careful design and op amps will probably work better/easier. 

 

Is this something that you are doing for your own learning/curiosity?  Or are you trying to develop something as a product for the masses to enjoy?  Again....giving more details on your goals and design strategy will help us give the most useful answers.

[the_crayon_king]

12 hours ago, ReilyReed91218 said:

So I need an expert in opamps to help me out. Using a 250MHz op amp to amplify Y and C and to matrix the signal into composite. Here's my schematic so far with resistor values that I believe to be correct. Not certain how to configure this for composite. Please do help out. If other could try this, I'd be ecstatic.

I don't know why you would not just have your op-amps being voltage followers. Like if you are setting gain on your op amps at all then your inputs must be being attenuated to less than 2V; I don't see a reason to do that.

Just change your 75 pulldowns on the video inputs to a higher amount.

 

I also don't see the point in having two potentiometers on the Y line. Like that is the point of the resistor ladder those resistances should already be correct.

 

I mean I see you are trying to augment the Luma without effecting the sync. My point is you shouldn't need to do that; these should already be weighted correctly. If when adjusting the other pot you lose video (sync is less than 300mv) then you have moved video out of the correct P-P anyway.

 

Also those input capacitors block DC right ?

So then your op amp MUST have a negative power rail. Else your 0 to 2V inputs without DC would be -1V to 1V right ?

If you don't have a negative power input then your output will be clipped at ground unless I am missing something.

 

As the others have said we don't really know your motives;

just messing around ?

trying to make an open/easy circuit for others ?

Trying to make a mod better than exists currently ? < this is unlikely

 

There are hundreds of ways to build this kind of circuit. I have done it just separating the Y/C on the Atari after the Atari's own dac ladder. Then fed those signals to a pot to ground then into the THS7314.

I may have used input caps and pullups or something and of course output caps but that's it.

 

If you want to sum for composite you should probably just double up the inputs to a THS7374 then sum two of the outputs (one Y and one C) with a 470pf capacitor. That way you still get S-Video outputs. There is also the FMS6410B but I don't like using it personally.

 

[ReilyReed91218]

BUT has anyone actually tried this op amp circuit I built? Can someone test it for me and give me some advice on it? Still can't test it myself, have no equipment is why.

[cwilkson]

8 minutes ago, ReilyReed91218 said:

BUT has anyone actually tried this op amp circuit I built? Can someone test it for me and give me some advice on it? Still can't test it myself, have no equipment is why.

I already told you it won't work.  Period.  There's no need to build it.

 

Please answer my questions in the previous posts before asking for more help.  We want to help, but we need more information from you.

Let's go one at a time, starting with the most obvious thing:

 

What are the 2 pots for?

 

 

 

In the larger sense, what are you trying to achieve with this circuit?  What are the goals?

• High quality output?
• Low cost?
• Easy to build?
• Lot's of user controls (contrast, brightness, tint/hue, saturation)?
• Learning how video works?
• Learning how op amps work?
• Something else?

the_crayon_king said it best....there are 100's of ways to build a 2600 A/V circuit.  Some good, some bad. 

To pick the best approach and to help you make it work, I need to know what you're trying to do. 

 

[ReilyReed91218]

thanks for all of the advice. sorry about the two pots, was testing my circuit with a friend and didn't know the values of resistance. Final product won't contain them. This is how I matrixed CVBS with a transistor. Not great. I hate the fact that there's more circuitry involved, but it should work. Constructive criticism wanted.

 

https://imgur.com/zmivV1B

[ReilyReed91218]

Just added the Pause mod to circuit. It stores the low/read high/write signal from the TIA/6507. So it's technically storing the instruction to read or write from the 6507. You cannot HALT on a write cycle which is effectively what the 6507 is doing on a low cycle. The added transistor mod I got to work and it amplifies composite not as well as I thought so I'm prototyping new ways of doing that.

 

Edited March 3, 2021 by ReilyReed91218