MOD: Mega Drive 2 PAL 60Hz color fix with a 2 pin crystal

Well the jailbars I was referring to was the ones over Composite but so I understand you correctly, you do the same RC circuit minus the Resistor and replace the 100pf with 330pf and keep the 100nf. Each color line get one of these circuits correct? And this is done to the RGB lines just before going into the encoder?

 

The real source of the jailbars is the VDP chip, not the CXA.

 

This is the first I've heard of this, everyone says it's interference in the RGB traces and there was even a person who is active here ArcadeTV (can't find the thread on this forum) who used the THS7314 to amplify the RGB lines off of the VDP and it got rid of the RGB Jailbars here is the link. Are you saying that all these assumptions and some workarounds used are false? Also you did not answer the question about if the circuit I mentioned above for the RGB lines are correct or not.

 

Man, this would be so much easier with pictures of it actually wired up. Circuit diagrams make me feel how people who can't read music feel when they look at sheet music (I can read sheet music, so it can be confusing when people say they can't read it, so with circuit diagrams, I have a point of reference for that feeling). Still, that circuit diagram is helpful. Thanks, Helder. I take it that I don't need the inverter with the 4-pin oscillator? Or is that incorrect?

By the way, do SNES consoles need colour correction for composite with a switchless mod, or do the PPUs handle the clock differences to compensate for that?

 

Explain to me how the jailbars "do not happen" when the CXA1145 chip is used to build a supergun ?

The stripes do coincide with the timing of the refresh of the DRAM chips used as video memory. So obvious source for the stripes pattern is the VDP chip and the video memories, obviously. Board layout actually contribute to make the jailbars stronger or weaker by injecting the signal on the CXA1145 chip.

Because it uses a lot of analog parts and the signal being processed loops outside the chip at least twice all that circuitry is a good antennae for interference. That's why people believe the jailbars are from the actual CXA1145 chip.

 

I can make a simple board that has a place for the oscillator and the circuit I posted a few posts back with all through hole parts so it becomes a painless thing and should cost a few dollars but I wanted to get confirmation for the RGB lines so I can incorporate that as well.

I wasn't saying that the jailbars occur in the CXA or any of the encoders used, I was under the assumption that the interference from the motherboard layout was the culprit since this is all you ever read about when this issue arises.

Thomas do you know what the package size of those oscillators you have are? Or give me some dimensions of it's size.

 

Here's a sample of one of my experiments using a MSX2 computer as source of the RGB signal:


Can you see our old friends in that image ? The MSX2 VDP is also made by Yamaha and it also use DRAM as video memory.

It simply means the CXA1145 is vulnerable to interference in it's inputs.

 

Ok so can you confirm the circuit to remove those is what stated above a few posts ago, basically is this what you're saying is the circuit:







Here Thomas and anyone else interested is the board with 2 different Oscillator sizes and only costs $4.15 USD for 3 boards:

https://oshpark.com/shared_projects/xzXoaWXy

 

Looks good. Keep in mind that the 100nf caps on R,G and B are already there.

 

So all that's needed then is a 330pf to ground? Seems like a simple fix, have you tried this yourself?

 

Not yet. I saw that fix recently on a Japanese MSX Turbo-R computer (Panasonic FS-A1GT).

It has quite the awesome video quality, FYI.

 

Well I can throw it in the board design and give it a go.

Edit: no board needed to try the fix out, here is a picture for reference.

 

Thanks, Helder. That'll be useful in the future. For this particular Mega Drive (don't worry, I have more to mod), I might make my own, though, since I have some spare board I can wire it too. I only need to use the one oscillator (dependant on the console's region), correct?

I'm not exactly sure what you mean by "package size", though. There's a couple of pictures of them in my sales thread if you want to see everything written on them. The dimensions are 18x10x4mm (not including the rectangular "edge" around the bottom, which adds another 1mm to each of those three numbers).

Actually, speaking of those oscillators, I originally got them to do a two-switch mod on my first Mega Drive 2. It's a bit of a mess (although I had a felt from another New Zealander who cleaned it up for me - this was about 4 years ago now, so my soldering skills were rubbish at best), but for it, I had to build a circuit with the oscillator. What's the difference between using that circuit and this one? It seems like it does the same job, after all, although I used to just assume that it was ONLY for changing to the new clock rate (correct me if I'm wrong, but the Genesis/Mega Drive is actually capable of that itself if you give it the instructions to do so with a PIC, like on the SNES).
For reference, here's the circuit diagram on his website that I followed (he also shows it in videos he did on YouTube, which was honestly a bit easier for me to follow): http://www.damomonster.com/Mod_Pages/SegaMegaDriveM2.html If I can use that, then I'd be 100% familiar with it. If I could wire up the PIC for the switchless mod AND the colour correction circuit all on one board, that'd be so much easier, too (although I'll probably only be able to have them on the same board but cut and links between the two parts, because there won't be much connected between the two if I'm getting that right).

 

connect the caps at the other side of the SMD caps, the side coming from the Mega Drive, not the side which goes to the SONY chip.

 

Yes I noticed that after I posted that picture, I'll eventually fix it.

 

So I'm confused. What are the caps on the board for? I'm going to try making the colour fix board within the next week to see if it'll work and make sure I understand what I'm doing, but do I need to add in those caps too, or is that for something else that I've completely missed? Also, does what l_oliveira said mean that I don't need a 100nF cap in that circuit as per the marking on Helder's board?

 

The 331 caps are meant to drain the radio frequency interference to the GND. Such interference is generated by the signal processing that happens inside of the video chip and by video RAM refresh cycles. The point is make it so it doesn't reach the video encoder, which is vulnerable to such interference.

 

So then it's "necessary", regardless of the colour fix, or is it specifically that the interference is generated by the processing that the colour fix circuit does?

 

Not "necessary". More like "something you can try" to help with the jailbars issue.

Edit: keep in mind that his drawing is wrong. Don't follow it.

 

Tried the caps to ground and still had bars but this was through composite but the others wider bars from normal composite is gone after using the board but I did cut the sub carrier trace right below the VDP since it was causing interference. The composite from this board is smoother like l_Oliveira said which I like, any idea on what to do to remove the other jail bars I assume are the RGB jail bars? It seems to be on the Blue color channel.