CXA1145, Chroma Subcarrier, and Crystal Oscillators

Hey all.

First post. I seek knowledge, guidance and fun discussion. I have been discussing this issue with several people via e-mail, and I thought I would open it up to a forum so I could share what I have learned and receive some feedback from others.

This past weekend I modded a model 1 Sega Genesis (NTSC) for S-Video output. It uses the (infamous?) CXA1145 video encoder chip. I used the schematic created by segasonicfan that is circulating the net. The output quality was improved quite a bit: S-Video vs. Composite

The picture hides the true improvement in quality, but it also hides another issue I have encountered - Vertical Bars (non-moving to the naked eye) are present in the output. They are especially noticeable over the blue background in Sonic.

This Picture does a good job of showing these bars. You can see there are technically two sets of bars running through the picture. I am concerned with those running in the direction of arrow #1. I will also point out there are others running in the direction of arrow #2, but those only show up on camera & are not found as easily by the naked eye. Still, #2 is probably worth pointing out.

My main question is - Have any of you encountered these bars from CXA1145 output before (NTSC or PAL I suppose), and what steps did you take to correct them? What is the problem? Is the color too strong?

I have received an e-mail that mentions that some people have attacked this issue by providing an alternate source of a color subcarrier. While this person hasn't encountered the issue himself, he mentioned that perhaps the size of the signal is too big and suggested trying to halve the signal by placing a 4.7k resistor in series with the signal and placing a second resistor to ground at the pin.

An alternative suggestion was to generate the signal by using a crystal, two caps and a resistor (values from the CXA1145 datasheet).

Again I was just looking for anyone to share their thoughts on this issue. I am especially interested in hearing from people that have encountered (and fixed) the same problem that I am having.

Thanks!

(btw - the amp circuit and both pins of the S-Video connector ground to the same place. The L and C ground pins are bent together and share the same grounding cable. This is a model 1 Sega Genesis with the smaller RF modulator & without the RS232 port.)

I see the bars on my model 1 Genesis with the S-video mod as well. You can also see this on the NES 1 and NES Top loader (even worse) with composite. The Neo Geo CD Top loader does something really weird with the bars. In composite it does this:
/ / / /
/ / / /
/ / /

S-video does this but connects solid bars from top to bottom:
/ / / /

I still wonder why this is so. Also, the S-video on SNES models from 1990-96 does the Neo CD S-video type barring, but it is hardly noticeable compared to all other systems with barring effects. All of these systems use different video chips so I dunno what the deal is.

Not familiar with the Sony encoders, Im using ADs.

Are you sure it is a colour issue - the lines dissapear if you kill the chroma ?

Looks like it may have something to do with grounding perhaps. ??

I don't see the lines when I kill the Chroma.

I actually ended up building that Chroma oscillator. Nothing seems to have changed.

I also acquired another Model 1 Genesis modded by someone else and noticed it had the same problem. Their video output was a bit cleaner than mine though. It could be they used a 33 ohm resister and 75 ohm resister for output rather than two 33 ohm resisters (one resister per output - Y & C)

I tried changing the ground and power source on mine several times and didn't notice a change.

One recent theory from another modder was that the RAM banks on the model 1 Genesis were placed too close to the video circuit. This can apparently be confirmed by running ones finger across the bottom of the board, just underneath the ram banks. I have not tried this, myself.

Not quite sure what to do. I just acquired a Sega CDX. Maybe I will give that one a try. I believe it still uses that CXA1145 chip, though.

I did the S mod on my Genesis 3, and I haven't seen any lines. Any different on any other televisions of yours? took two pictures of mine running.

Composite

S-video

I actually do not have an extra TV to easily test it.

I also just acquired a Micomsoft XMD-2 for the Genesis model 1 and the problem is not present. There may be a few odd artifacts here and there (instances of lines), but for the most part it looks really good.

Of course the XMD-2 takes the RGB output and externally converts it to S-Video, so I am still not quite sure about the quality of the Y/C "native" output of the CXA1145 chip.

As for the original problem - I have grounded it in multiple places, but it doesn't seem to make a difference. I did (as many do) bend together the two GND connections for Y/C at the plug. I have no idea if grounding those separately would have made a difference.

FairlanePhantom said:

I did the S mod on my Genesis 3, and I haven't seen any lines. Any different on any other televisions of yours? took two pictures of mine running.

Composite

S-video

Click to expand...

Great pics dude. It really shows the difference.