The CSYNC on Genesis

32x's SCART cable is the exact copy of MD1 SCART cable named "adaptateur RVB". Both take the csync signal instead of CVBS.

 

I don't think MDII suffers of the problem with the CSYNC pin the original MD had.

Edit: Gosh I am repeating myself that the Mega Drive has a design error on the SYNC output circuitry. Is that hard to sink in ?

Because of that, things like this do exist:

http://www.assemblergames.com/forum...Analog-RGB-Unit-for-Megadrive-What-does-it-do!

What happens is that if you draw current from the CSYNC pin, you "kill" the CXA1145 as it will lock down in the clamp state and all video outputs will be shorted to GND. (that's only supposed to happen during the blanking phase of video signal...)

 

But then, if the French Mega Drive has a design error on the SYNC output circuitry, how could we manage to get it work with a 32X ? We had lots of problems when 32X was launched in France to make it work when the player had a Megadrive I instead of a Mega Drive II (No problems with the Mega Drive II which has the same AV out as the 32X and the same SCART cable).

That's why SEGA very quickly delivered a SCART cable only dedicated to the use with a Mega Drive I with a grey DIN connector in the 32X packagings in order to make the difference between the one to use with a Mega Drive II and the one to use with a Mega Drive I.

Hence my question. Is there a difference between those two SCART cables ? There has to be a difference, or else why bothering to put one dedicated SCART cable to the use of the 32X with a Mega Drive I ? And I mean the SCART cable ON THE 32X !

 

I explained but people didn't care about reading, right ?

I said common TVs have a 75 ohm resistor inside them to act as a terminator for the video signal. Normal TVs deal with the CSYNC pin the same way they deal with CVBS. On some standards even the same pin functions as CVBS or CSYNC. (Isn't SCART like that?)

And SEGA made the CSYNC circuit driven by an digital open collector output with a very weak pull up (2k2) resistor. The 75 Ohm resistor inside the TV will overdrive that and in the end you will have no SYNC pulses at all. There's no such resistor on the 32X as it's designed to take whatever a MEGADRIVE console outputs. Then it obviously work with the 32X but not when connected straight to the TV. This must be the third time I'm explaining this ... lol

 

Sorry if this discussion is getting frustrating, and I hope my questions aren't off-topic, but...


1) MD2->32X presumably uses the 'correct' CSYNC as provided by the MD2.
2) MD1->32X uses the 'incorrect' CSYNC as provided by the VDP

->I modified a MD1 to use the CSYNC from pin 11 of the CXA chip (to be entirely honest I can't remember if I used raw signal from pin 11, or put it through a cap+resistor - I *think* I actually tried both)
->32x no longer displayed video correctly (couldn't handle the sync signal anymore).
->Reverted CSYNC back to VDP and 32x worked again.

I don't really understand how this can be possible, given point 1. Unless the MD2->32x gets the sync from the composite video?

Is there a way to modify a PAL MD1 such that you get true CSYNC from the CXA pin 11, and ALSO retain compatibility with the 32x?

 

I don't know if it can help you more, Dr.Wily, but let's try (I think Fluxcore understood what Dr.Wily wanted to demonstrate and I think it would be interesting to answer this ultimate question from Fluxcore. I am sorry, l_oliveira, to be blind. I promise to read better for the next times )

http://www.sega-16.com/forum/archive/index.php/t-19347.html

 

That thread has some "naming problems" which can make people even more confused.

For example, COMPOSITE SYNC = Horizontal SYNC + Vertical SYNC mixed on the same wire.
People at the thread are talking about using Composite VIDEO as Composite SYNC when they say Composite SYNC...

The whole issue is about the Mega Drive not being designed to work with SCART but with Japanese RGB21 standard, instead ... http://junkerhq.net/xrgb/index.php/RGB21

 

Good point, l_oliveira ! But in France, everything was ok with our SCART boxes on Master System 1, 2 and Mega Drive 1. Why is it such a mess now to connect the 32X to Mega Drive 1 whereas SEGA was ok with the SCART cables they provided in their packagings ? Why did they mess up to offer RGB to 32X with the use of Mega Drive 1. Mega Drive 1 alone did work beautifully with its SCART cable in France. Why was it not the case for the 32X ?

Can you tell me again what SEGA should have done for the French market to have no problems at all in dealing with the Mega Drive I and the 32X ? Should they have modified the video output of the 32X or, better, launched out a 32X for Mega Drive I and a 32X for Mega Drive II ? By the way, I better understand why SEGA quickly wanted to set the Neptune project...

 

Honestly to be really helpful with this I would need to see *live* what is happening at the OP's hardware and propose a solution.

SYNC signal is troublesome because it's a digital signal but on a lot of standards it's dealt with as an "analog" signal so it can be kept compatible with CVBS. That's what happens with the SCART system, for example. Because deep down CVBS and CSYNC can both be used as source of synchronization a lot of stuff do skimp on SYNC and use CVBS as SYNC signal source. But you mentioned degraded video quality... That can happen when the color carrier waveform irradiates inside the cable or within the TV circuitry (after reaching the SYNC area) and the bars people see are just that, interference from color burst signal.

Such economy of wires while can be practical on "usability" of devices it always have some kind of detrimental effect. Now look this compatibility issue with the MD1 sync, in the end it boils down to voltages and electrical current on the SYNC line.
All because SEGA tried to save some money on a 75ohm resistor and a 220uF capacitor so they could use the CXA1145 CSYNC output buffer...

 

x

 

Unfortunately the amplifier delays the signal and that causes the screen to shift. Remember, the monitor synchronizes to the "SYNC" signal it receives so if you play with that signal the changes are obviously visible on the output.

I had a similar problem with a MSX2 computer (funny VDP is also from Yamaha on that one) and to solve it I used discrete transistors to amplify the SYNC signal (faster switching perhaps ?)

 

if anyone comes up with a schematic for such an amp for the md I'll try it , I happen to have a mk1/VA6/pal waiting a recap (preemptive one, works fine)

 

There is no naming problem here. I know what is difference between composite sync and CVBS.

Don't forget, French MD1 is not PAL MD1.

- PAL MD1 has CVBS and CSYNC from pin 42 of VDP + RF module
- French MD1 has only CSYNC from pin 11 of CXA1145 and no RF module, no CVBS signal

32x take the CVBS signal as reference for video clock, but there is no CVBS on french MD1. On PAL MD1 there is a CVBS, but french RGB cable uses CSYNC only and AFAIK the csync line on 32x does not carry this signal to genlock, it's just a direct pass through to 32x's video out.

PAL MD1 with Sega official RGB cable that uses CSYNC gives this with 32x.

On GEN1 Sega official RGB cables use CSYNC. On GEN2 it uses CVBS. Are you using a homade cable ?

 

I said that having no sync pulses on it's input causes the CXA1145 to keep the outputs shorted to GND, didn't I ? A black screen with no scenario (only the output from 32X is shown on that picture) is consistent with what I said. There's still enough of the pulses to sync the 32X but the CXA1145 is VERY PICKY with the sync voltage variation as it has a clamp (DC level restoration for cable capacitors) circuitry inside.

Seems to be a PAL LAND problem because I never saw that happen with NTSC hardware.

 

Yes, I am OK with that. But on the csync case, I am more sceptical. How works the genlock circuitry on 32x ? Is the 315-5788 need to recieve a sync signal (csync or comp) to synchronise with it ? Because csync does not go to 315-5788 chip.

No, this happens on NTSC hardware too. I think you never tested with official RGB cable, only homemade cable with CVBS or CSYNC with 75ohm+220µf cap. I have to check 32x's schematics to see if the csync goes to 5788 or not.

EDIT :

Checked 32x video schematics and... csync goes directly to 32x's video out. BUT, the trace goes through pin 23 of 315-5788. This pin is labelled C Sync (in, out ?) and I don't know what is its purpose...

 

I NEVER, EVER used a 32X in RGB. Only in CVBS. Remember we don't use RGB on this side of the Atlantic ... lol

Anyway, seeing that they have separated mux for RGB and CVBS does explain some of the problem now.

If you look again at the schematic you see that there's no component connecting anywhere to GND through the whole CSYNC line. Were you using the ORIGINAL MD1 to 32X adapter ? I then never used a 32X on a MD1 without the official cable. And it's well possible that it's not a straight pass through. There might be circuitry in it.

I no longer have access to 32X hardware so I can't really study that.

 

I too no longer have access to the 32x, otherwise I would make a video explaining the problem I saw. But I don't believe my earlier post shows any confusion between CSYNC and composite video, so I'm not sure if you were referring to me or not.

 

No, I think el_oliveira was meaning the link I posted to help Dr.Wily which was confusing indeed given that the guys there were mixing csync and sync on composite.

 

I was talking about the thread at the SEGA16 forum.