Megadrive RGB / softness / caps

[bloodflowers]

Hi - I made my cable to match this diagram here:



I'm seeing vertical bands where the picture is much softer - i'd say the 1st and 3rd quarters, the rest of the picture is pin sharp.  Two megadrives are doing the same thing.  I tried modifying a spare cable to resemble this instead:



Note the value and positioning of the caps is different.  The effect is still there, but not as bad.  Someone on the Sega-16 forum says he had the same issue until he removed the caps.

What's going on?

[bloodflowers]

As an update - I removed the caps from my cable and the problem persists, no better or worse really.

Could it be related to the sync perhaps?  This TV objects to PAL modded machines but native 60hz have looked fine until now.

[RGB32E]

As an update - I removed the caps from my cable and the problem persists, no better or worse really.

Could it be related to the sync perhaps?  This TV objects to PAL modded machines but native 60hz have looked fine until now.

Try CSync instead of composite video for scart pin 20.  You might need to add a series 100uF cap on the sync signal though (works for some monitors).  For some TVs, you'd need a sync stripper (LM1881) and/or hex buffer to get a good result.

I've never understood why viletim's scart diagrams show resistor before capacitor, even though the datasheet for the respective RGB encoder ICs shows otherwise.    I make my Genesis cables with capacitor then resistor.    But then again, some things are different between PAL and NTSC systems... 

[bloodflowers]

Another question - why do I see 5.03v on pins 8 and 16 despite the fact that there's a 100ohm resistor between them?  I checked on another cable someone else made and see the same thing - I assume I'm just measuring them wrongly?

[bloodflowers]

Well - I modified the Megadrive to output csync out on pin 7 instead of csync in (surely a mistake during design?), on the cheap cable with 470uf caps on it, this resulted in no change at all.  On the nice cable I made now with no caps, this resulted in moving interference waves.  Both cables are set with pin 7 -> pin 20 on scart.  

The only conclusion I can make from this and the softness - the difference between the two cables hasn't changed at all in that respect, is that the actual wire I used is garbage - and I'm going to have to make a new one.  I'll make sure all wires are coax cores this time I think - in theory I can start running the audio in the same cable again without the 50/60hz hum?

Also - should I be following the schematics and using 470uf caps, or are 220uf better?  I think someone said 470uf will make the picture brighter, does that mean using a 470 on the sync is going to send a stronger sync signal which might interfere with the RGB ones, or will shielding look after that?

If I use 75 ohm coax does it mean I don't need the resistors?

Sorry for all these possibly dumb questions.

[marqs]

I've never understood why viletim's scart diagrams show resistor before capacitor, even though the datasheet for the respective RGB encoder ICs shows otherwise.    I make my Genesis cables with capacitor then resistor.    But then again, some things are different between PAL and NTSC systems...  

The capacitors are probably there to block DC so the exact value (whether it is 100u, 220u or 470u) shouldn't matter too much. Similarly, it should make no difference whether you put the resistor or capacitor first.

Another question - why do I see 5.03v on pins 8 and 16 despite the fact that there's a 100ohm resistor between them?  I checked on another cable someone else made and see the same thing - I assume I'm just measuring them wrongly?

I guess the cable isn't connected to TV (which has 75ohm input resistance on RGB status pin)? Try measuring the voltage when connected or put a 75ohm resistor between 16 and gnd (temporarily).

Well - I modified the Megadrive to output csync out on pin 7 instead of csync in (surely a mistake during design?), on the cheap cable with 470uf caps on it, this resulted in no change at all.  On the nice cable I made now with no caps, this resulted in moving interference waves.  Both cables are set with pin 7 -> pin 20 on scart.  

The only conclusion I can make from this and the softness - the difference between the two cables hasn't changed at all in that respect, is that the actual wire I used is garbage - and I'm going to have to make a new one.  I'll make sure all wires are coax cores this time I think - in theory I can start running the audio in the same cable again without the 50/60hz hum?

Also - should I be following the schematics and using 470uf caps, or are 220uf better?  I think someone said 470uf will make the picture brighter, does that mean using a 470 on the sync is going to send a stronger sync signal which might interfere with the RGB ones, or will shielding look after that?

If I use 75 ohm coax does it mean I don't need the resistors?

Sorry for all these possibly dumb questions.

Strange, I've done a MD rgb-cable from a pretty standard wire and it has worked flawlessly (no humming or interference). Are your Tv and MD connected to the same electricity outlet?

I had once a secondary, newer revision us snes (not snesjr) in addition to a first revision model. The new model had a lot more interference with the same cable and there was a lot less shielding around the mainboard (which may or may not have caused that). I don't know about MD revisions, and it sounds unlikely that your two MDs would suffer from the same problem.

The cable impedance is another matter, so you still need the resistors. Just remember to connect the coax shieldings to ground in both ends.

No problem for the questions, just feel free to ask. Let's hope we'll find a solution

[viletim]

Well - I modified the Megadrive to output csync out on pin 7 instead of csync in (surely a mistake during design?), on the cheap cable with 470uf caps on it, this resulted in no change at all.  On the nice cable I made now with no caps, this resulted in moving interference waves.  Both cables are set with pin 7 -> pin 20 on scart.  

That's no design mistake. The input signal to the encoder IC is TTL (greater than 2.4V = high, less than 0.8v = low) while the output is a 75 ohm sync output (capacitively coupled 0.3Vpp when terminated) -- basicaly a composite video signal with no video information.

The only conclusion I can make from this and the softness - the difference between the two cables hasn't changed at all in that respect, is that the actual wire I used is garbage - and I'm going to have to make a new one.  I'll make sure all wires are coax cores this time I think - in theory I can start running the audio in the same cable again without the 50/60hz hum?

The wire/cable doesn't matter too much as long as you have a good, low resistance ground connection and it's no longer than two meters or so.

What's kind of monitor are you using?

[MKL]

I've never understood why viletim's scart diagrams show resistor before capacitor, even though the datasheet for the respective RGB encoder ICs shows otherwise.    I make my Genesis cables with capacitor then resistor.    But then again, some things are different between PAL and NTSC systems... 

It's a matter of consistency: on the composite video output (in the console) the Megadrive has resistor+capacitor. The RGB outputs follow the same pattern.

[albino_vulpix]

I'd say it's more likely that your monitor is the problem.