Another RGB to S-Video / Composite Adapter

[MikeS1]

Hi All,

Just wanted to post the source for a S-Video / Composite Adapter I've been working on the last couple of weeks.

https://github.com/MikeS11/MiSTerCRT

Now, I still have not tested the composite out since I'm waiting on the part but I at least wanted to post the results of the S-Video.

Some notable differences from a couple of the others, I'm using a fairly similar IC in the AD724, but I wanted to use a FSC or 1x NSTC crystal and I added a trim capacitor. I'm hoping adjusting the capacitance on the frequency circuit will help manually minimize as must of the composite dot crawl I've been hearing about. Ill post the results when I get the RCA port in.

So far I have the S-Video working really well on my 1702 monitor.

MiSTerCRT3.jpg (170.12 KiB) Viewed 4232 times

[antoniovillena]

I have tried many prototypes: AD724, AD725 and AD723. The best quality I found was with AD723. There is not difference in dot crawl between use 1x or 4x crystal. I use 4x crystal because it's more stable but I think it's because before I used passive ones. So I think you won't have stability problems. Also you can replace the passive crystal+inverter for active crystal like this to simplify.

https://www.mouser.es/ProductDetail/ECS ... J1cA%3D%3D

[akeley]

The natural dot crawl is not so much a problem, because that's how it is on the original HW. The issue was the alleged extra artifacts these boards added (though it's still rather unclear if, or what they were).

Anyway, a new design is always welcome. Will it require the I/O board?

[aberu]

The natural dot crawl is not so much a problem, because that's how it is on the original HW. The issue was the alleged extra artifacts these boards added (though it's still rather unclear if, or what they were).

Anyway, a new design is always welcome. Will it require the I/O board?

The problem is chroma interference, something that a y-trap should deal with but in this case didn't, that I personally saw. This is something I experienced with Antonio's earlier revision of his adapter (I haven't tested his new one or bought it). I compared to original hardware with my JVC X'Eye and later with both my Sega Genesis MD1VA6 and MD1VA3 for the same games. The video output was not the same.

Unfortunately I don't think there is any way around this at all, you will get minor to major differences in the video output. I've talked to a few people who have done work with video adapters before and they all kinda say the same thing, that relying on the static xtal and a cheap chip like this will continue to cause the problem since each of the cores would require slightly different xtal and circuitry to match original hardware. Even with a pricey chip it would be exceedingly difficult.

I wonder if the trim capacitor addresses this issue adequately.

[akeley]

Unfortunately I don't think there is any way around this at all, you will get minor to major differences in the video output. I've talked to a few people who have done work with video adapters before and they all kinda say the same thing, that relying on the static xtal and a cheap chip like this will continue to cause the problem since each of the cores would require slightly different xtal and circuitry to match original hardware. Even with a pricey chip it would be exceedingly difficult.

It's a bummer, but at least if it's clearly stated people can make an informed decision. If it's minor artefacts then perhaps it wouldn't be that much bother, but everybody's got different tolerance. What would help if there was a decent comparison video vs real hardware. I'd get one myself because I might need it in the future (am using SCART RGB-capable TVs atm) just to see what's up, but if it also needs I/O board then it's bit too much of an investment for me. Does it happen in just composite, or with S-Video as well?

I suppose the problem is more on the MiSTer side, but the interest in this whole thing too niche to warrant making a lot of effort solving it, which is understandable.

[MikeS1]

The natural dot crawl is not so much a problem, because that's how it is on the original HW. The issue was the alleged extra artifacts these boards added (though it's still rather unclear if, or what they were).

Anyway, a new design is always welcome. Will it require the I/O board?

The problem is chroma interference, something that a y-trap should deal with but in this case didn't, that I personally saw. This is something I experienced with Antonio's earlier revision of his adapter (I haven't tested his new one or bought it). I compared to original hardware with my JVC X'Eye and later with both my Sega Genesis MD1VA6 and MD1VA3 for the same games. The video output was not the same.

Unfortunately I don't think there is any way around this at all, you will get minor to major differences in the video output. I've talked to a few people who have done work with video adapters before and they all kinda say the same thing, that relying on the static xtal and a cheap chip like this will continue to cause the problem since each of the cores would require slightly different xtal and circuitry to match original hardware. Even with a pricey chip it would be exceedingly difficult.

I wonder if the trim capacitor addresses this issue adequately.

I've noticed that having the trim capacitor has been really useful because cores like the NES/SNES/C64 and others can be flawless, cores like the Genesis/NeoGeo and arcade cores need to be tuned a bit to reduce some slight flicker. The problem is adjusting for one core will affect others so its a bit of a battle but at least having an adjustable cap makes it so you have flexibility.

Also one of the reasons I wanted to build this was from compatibility issues ive heard with the commodore 1702 monitor. I didn't seem to have any issues with the AD724.

[aberu]

It's a bummer, but at least if it's clearly stated people can make an informed decision.

Unfortunately for me, it wasn't clearly stated before I bought it

[beamrider]

Looks interesting. Is this NTSC only?

[MikeS1]

Its easily converted to PAL, you just need to get the correct crystal and there are a couple changes to the capacitors, I can update Github to reflect those.

I've also tested out the composite and for all of the consoles that had some type of composite out, you can get some decent composite out of them and you can reduce any of the rainbow fringing so its not shimmering all over the screen with the trim capacitor. But for the arcade cords or NEOGEO, its pretty much a mess and there is really nothing that can be done unless you can build in the colorburst right from the FPGA core.

I'm going to look at testing out the 1ChipMSX implantation of composite into the mister framework as a exercise but for the few people who want the feature, I feel like its going to just be taking up needed resources. Those I think S-Video is probably something that can be added without much work. You really need to only combine the Cb and Cr so Ill do some testing there as well.

[darksakul]

The natural dot crawl is not so much a problem, because that's how it is on the original HW. The issue was the alleged extra artifacts these boards added (though it's still rather unclear if, or what they were).

Anyway, a new design is always welcome. Will it require the I/O board?

The problem is chroma interference, something that a y-trap should deal with but in this case didn't, that I personally saw. This is something I experienced with Antonio's earlier revision of his adapter (I haven't tested his new one or bought it). I compared to original hardware with my JVC X'Eye and later with both my Sega Genesis MD1VA6 and MD1VA3 for the same games. The video output was not the same.

Unfortunately I don't think there is any way around this at all, you will get minor to major differences in the video output. I've talked to a few people who have done work with video adapters before and they all kinda say the same thing, that relying on the static xtal and a cheap chip like this will continue to cause the problem since each of the cores would require slightly different xtal and circuitry to match original hardware. Even with a pricey chip it would be exceedingly difficult.

I wonder if the trim capacitor addresses this issue adequately.

I wonder is that why the Analog DAC cost so much?

[lu_source]

Can any damage happed to the CRT when S-Video is not being output. For example, when I'm not in any of the YC cores and navigating the MiSTer menu's I get a black and white picture on my CRT. Will there be any side effects?