Analog Video Issues - Composite Modding Atari Super Pong

[Reckless Maker]

Hello All,

Today I tried adapting the Atari 2600 composite mod (using

schematic) and was actually mostly successful! There's severe color banding around everything (I believe the original signal is monochrome) but I'm thinking that diagnosing this might be a no-brainer for someone familiar with analog video signals. So here are some photos of the output:

 

And the super pong schematic for reference:

[BrianHG]

Try a different TV.  If the video out has no color burst, the TV supposed to go into forced B&W only mode.  It could be bad firmware in the TV which doesn't respect old fashioned B&W NTSC signals.  This doesn't necessarily mean a B&W movie/TV show inside a color NTSC signal, those broadcasts take special considerations to avoid coloring around the edges of vertical black / white angled line patterns which would cause purple/green crawling color patterns.

If the output of that chip is a color NTSC signal, then you need a TV which better decodes 240p, the most likely video mode of that IC, not the standard 480i which the TV's color comb filter was designed to operate with.

[Reckless Maker]

The video output is actually supposed to be color. It's not proper color but rather a rainbow that gets overlaid on top of the white elements to give the illusion of color. The effect of which can be seen here. The overlay actually comes through perfectly if I remove R10:



Could it be that perhaps L1 and L2 are necessary as some kind of filter?

[NiHaoMike]

Can you show sheet 2? It might be expecting an impedance other than 75 ohm.

[Zero999]

The video output is actually supposed to be color. It's not proper color but rather a rainbow that gets overlaid on top of the white elements to give the illusion of color. The effect of which can be seen here. The overlay actually comes through perfectly if I remove R10:



Could it be that perhaps L1 and L2 are necessary as some kind of filter?

Have you tried an old analogue CRT TV? Those separate colour bands might smooth into one, on an old fuzzy CRT display.

[Reckless Maker]

Can you show sheet 2? It might be expecting an impedance other than 75 ohm.

Here is sheet 2:

[NiHaoMike]

It does look like the impedance is going to be considerably more than 75 ohms. Connect a scope to pin 18 on the chip and note the voltage levels with the original circuit connected, then try replacing the original circuit with resistors of varying values until the voltages look similar. Also note the peak voltage observed and that will tell a lot about how to get a standard composite signal out of it.

While I can't confirm, it looks like the "rainbow" is made by feeding a little signal from the oscillator into the output, tuned slightly off frequency in order to get "Never Twice the Same Color".

[BrianHG]

Show us your composite mod circuit...  The 7404 oscillator at the crystal feedback is capacitively coupled to the video output.  A strong load here would cause the oscillation/color to drift.  The TV modulator doesn't have a hard 75 ohm load like NTSC video.

Other things to try:
Step 1, try tuning C7.
Step 2, try changing the crystal for a more modern 3.579545 Mhz one.
Step 3, I would also add a 2200uf and 0.1uf cap across IC A1's VCC and GND.

How are you powering this device?  It has a simple 7404 oscillator which is also used to generate 2 power supplies, +/-9v for the Atari pong chip.  All this may skew it's quality of oscillation.

[Reckless Maker]

Ok so I finally had a chance to work on this again. To remove as many variables as possible I hooked the composite mod up to my spare unit with everything but the RF modulator connected.



This improved the image quality marginally but there are still plenty of noticeable artifacts.



I tried looking at the video waveforms on my scope but they looked normal. I only have an old analog scope though so I couldn't zoom in very far. I tried tuning C7 and that didn't do anything. As for the oscillator, it's being powered by batteries so all of the supply voltages on the board are clean (I also inspected them with the scope to verify this.) Would anyone with a nicer scope be willing to have a look at the video signal? I would be happy to cover shipping costs.