Bad NeoGeo CD sync signal?

[tomwaits]

I'm trying to connect my NeoGeo CD to an NEC XM29.  I bought a custom SCART cable with the console's sync output fed into the monitor's csync pin, but I just get flashing horizontal lines like I'm getting a bad sync/no sync.  I checked the sync pin with the frequency setting of my multimeter and it reads 29kHz??  My other consoles are running fine and they read around 15kHz on the csync line.  Is there something weird about the NeoGeo CD's sync signal, or is there maybe a problem with my console?  Any other way to get csync from the NGCD without building an LM1881 circuit?

Thanks!


EDIT: Checked the sync line on my Saturn and that reads around 30kHz but it displays fine, so maybe that's irrelevant.   After searching more old posts here, I saw threads about either adding a 100 ohm resistor to the sync line or re-routing the csync pin to come directly from an IC leg in the NGCD.   I got some 100 ohm resistors tonight and I'm going to try adding one on the sync line next...

[RGB32E]

Verify that you are actually using CSync to begin with.  Standard SCART cables use composite video instead of CSync.  If that's the case for your cable, the NEC XM29 will NOT strip sync from composite video.  So, check to see if you're getting Sync from pin 3 or 7:

http://gamesx.com/wiki/doku.php?id=av:neogeoav

[tomwaits]

My SCART is non-standard and definitely connected to pin 7 for csync.  (not composite video)  I confirmed a good connection with a multimeter.  I've been tapping sync directly out of the consoles for all my cables instead of building sync separator circuits.

Didn't get a chance to try the resistor yet.

[RGB32E]

Another weird thing I encountered with a NEC XM29 recently is that it wouldn't sync with a RGB modded TurboDuo!  The Turbo Duo was using the GamesX video amp (amplified Red, Green, Blue, and CSync signals), yet it wouldn't display a stable image!  However, using an external sync stripper (Extron brand), a picture would display! 

The same RGB'ed system works just fine with my PVM-2030 and XRGB-3...  though both units will strip sync anyways.  I'm curious to see what the issue is here... one friend has a Neo Geo CD, and another has the XM-29... Hmm...

[tomwaits]

Adding resistors on the sync line didn't help.  I tried 1, 2, then 3 100 ohm resistors in series on the sync line and couldn't get stable video.  Hard to tell if they were making any difference at all.

Guess I either have to try tapping sync directly off an IC leg or building a sync separator circuit.

[RGB32E]

Adding resistors on the sync line didn't help.  I tried 1, 2, then 3 100 ohm resistors in series on the sync line and couldn't get stable video.  Hard to tell if they were making any difference at all.

Guess I either have to try tapping sync directly off an IC leg or building a sync separator circuit.

Yeah, I'd try removing whatever through hole component exists before pin7 on the 8 Pin din and tap composite sync from pin 11 of the Sony CXA1145P chip (pin11->220uF electrolytic -> 75 ohm -> pin 7 of output connector.  The data sheet shows a 75 ohm resistor to ground or +5VDC, however I don't think you'll need to do this, just the cap and resistor in series to pin 7 (without old signal connected to pin 7).

http://nfggames.com/games/neorgb2/

[RGB32E]

For comparison, Micomsoft's XMD-3 uses pin 11 (CSYNC) output and uses a 220uF and 75 ohm resistor in series to the RGB output connectors:

[viletim]

Yeah, I'd try removing whatever through hole component exists before pin7 on the 8 Pin din and tap composite sync from pin 11 of the Sony CXA1145P chip (pin11->220uF electrolytic -> 75 ohm -> pin 7 of output connector.  The data sheet shows a 75 ohm resistor to ground or +5VDC, however I don't think you'll need to do this, just the cap and resistor in series to pin 7 (without old signal connected to pin 7).

All of these CXA encoders and clones have a TTL C-Sync input and a 75 ohm C-Sync output. The 75 ohm output won't drive a TTL input of a monitor any better than a composite video signal will. The TTL input to the encoder  chip can be used but it's best to buffer it first.

[tomwaits]

Thanks... so, I'd use pin 10 instead of pin 11 to get TTL sync, how should I buffer it?

[RGB32E]

Yeah, I'd try removing whatever through hole component exists before pin7 on the 8 Pin din and tap composite sync from pin 11 of the Sony CXA1145P chip (pin11->220uF electrolytic -> 75 ohm -> pin 7 of output connector.  The data sheet shows a 75 ohm resistor to ground or +5VDC, however I don't think you'll need to do this, just the cap and resistor in series to pin 7 (without old signal connected to pin 7).

All of these CXA encoders and clones have a TTL C-Sync input and a 75 ohm C-Sync output. The 75 ohm output won't drive a TTL input of a monitor any better than a composite video signal will. The TTL input to the encoder  chip can be used but it's best to buffer it first.

Shenanigans!  What a load... I think viletim is just confusing the point, and not offering any useful info...  not the first time thats happened.  Remember Tim, the NEC XM29 DOES NOT accept composite video on the sync input... I think you're missing that point.

I've tested the output from pin 11 (wired to the RGB output connectors on the XMD-3) and it works great!  Like I said before, if you connect a 220uF capacitor and 75 ohm resistor from pin 11 (Sony) to pin7 (8pin din), and remove whatever was wired to pin 7, you'll fix the issue.

[tomwaits]

ok... just have to pick up some 220uF caps and I'll give it a shot.  Pin 11 first, and if that doesn't work for me I'll switch it over to pin 10 and try that.

[RGB32E]

ok... just have to pick up some 220uF caps and I'll give it a shot.  Pin 11 first, and if that doesn't work for me I'll switch it over to pin 10 and try that.

MLK described the proceedure in this post:
http://nfggames.com/forum2/index.php?topic=3387.msg22704#msg22704

In case you want to restore sync (so you can use standard cables) all you need to do is desolder the blue thing (ferrite) shown here:

and solder a wire like the yellow one here (ignore the other things):

Remove this part (in yellow box):


Connect pin 11 (left of pictured connection - pin 10) to pin7 using a 220uF and 75ohm resistor in series:


I'd imagine that you could just connect pin 10 sync input, but I think your compatibility would increase by using the sync output (pin 11).

[RGB32E]

Another interesting note is that SNK actually used the audio in and out pins (8 & 9) on the 1145!

[viletim]

Just for you RGB32E:

Measured from a CXA1145 in an SMS2.

Composite sync input (pin 10)


Composite sync output (pin 11)

[RGB32E]

Just for you RGB32E:

Measured from a CXA1145 in an SMS2.

Composite sync input (pin 10)


Composite sync output (pin 11)

Thanks!    So, you made the assertion that the output from pin 11 wouldn't work, yet I've used that output and it works just fine...?  What specific buffer circuit were you implying when stating this:

The TTL input to the encoder  chip can be used but it's best to buffer it first.

[tomwaits]

No luck with pin 11 -> 220 uF -> 75 ohm -> sync input.  Hard to tell if it was any different than what I was getting from standard pin 7 of the A/V port.

Tried pin 10 -> 220uF -> 75 ohm -> sync input and that didn't work either.  Looked a little different, but still a mess.  I also tried pin 10 directly to the sync input and that seemed closer.  The image was still distorted horizontal bands, but the screen was definitely more stable and wasn't flashing on and off.  If viletim has an easy to build 'buffer circuit' from pin 10 I'd give that a shot.  Otherwise, I'll have to track down an LM1881 or just stick with s-video for the NGCD.


Also, the Genesis 1 uses a Sony CXA1145P and I'm getting stable video using the Genesis 1's standard sync pin.  Could I somehow duplicate the Genesis 1's output path from the CXA1145P to the sync pin?

[RGB32E]

The XM29 is an ODD beast... I used to own one till I sold it!  I tried hooking a 2SC1815 amped TG16 to a XM29 recently and couldn't even get a stable image using CSYNC!  However, using a Sync separator device from Extron yielded a stable image.  I think you're best bet at this point in time would be to build a LM1881 circuit.  The LM1881 should do the trick!

EDIT:
Try fiddling with the dipswitches on the back of the XM29 related to sync settings..?

[tomwaits]

I think the only sync related switches on the back of the XM29 are for sync-on-green/composite sync/RGBHV.  I'll have to climb behind with a flashlight later to confirm.

I'll try an LM1881...



I'm really happy with the XM29 so far.  The picture quality is great, and it supports 15kHz & 31kHz so I can get 480p from my previous gen consoles.  So far, I've got RGB connections for:

15-pin: PC, Amiga 1200 (all modes), Atari ST (color or mono), Dreamcast VGA Box
XCM component to RGB converter (set for no upscaling - no lag):  Wii, Gamecube, PS2, Nuon, PSP
SCART (sync modified): Supergun/MVS, Saturn, Genesis 1/32x/CDX/PBC (SMS), Jaguar

Haven't had to use an LM1881 yet, just tapping syncs directly from the A/V outputs.
(Update: The NeoGeoCD and SNES both required LM1881 circuits to get a valid sync for the XM29)


A few issues though... the XCM converter triggers Macrovision protection on retail DVDs, so I can't watch DVDs on the PS2 or the Nuon.  I have to re-connect via s-video if I want to watch a DVD in my game room.  It also causes weird color distortion on the PS3 if I try to use any progressive scan mode.  The PSP menu/UMDs have the color distortion too but PSP games look fine.  I also tried my Sony DVP-NS700H DVD player and that has both color distortion on the menus in progressive scan mode and Macrovision distortion during 480i DVD playback.  Some limitations, but the converter was a lot cheaper than an XRGB and it looks great with the previous gen consoles which is the focus for my game room anyway.

The 16-bit Sega setup worked out well though... I'm using a sync modified Genesis 2 -> SCART cable and it covers every console variation.  To use the bare Genesis 1, I just use the adapter stub from my 32x connected directly to the SCART cable input.  32x or CDX can plug directly into the SCART cable input.  Everything looks great.

The NGCD is the first 'difficult' system that I've hit.   Still have to work on SNES (waiting on a gamebit) and maybe N64.  Probably going to stop there since any other RGB mods are beyond my soldering skills.  Everything else can just connect via s-video or composite...

[viletim]

Thanks!    So, you made the assertion that the output from pin 11 wouldn't work, yet I've used that output and it works just fine...? 

I don't know the monitor so I can't say it simply won't work. It isn't a valid TTL level signal (it never goes below about 1.8v) so you should not be surprised if it doesn't work but anything is possible if the monitor has tolerant sync input.

What specific buffer circuit were you implying when stating this:

The TTL input to the encoder  chip can be used but it's best to buffer it first.

Open up your bin of parts and fish out some chip marked 74LS??. Most logic gates can be used provided they don't have open collecter outputs and you understand boolean logic. Take a 74LS32 (quad NOR) for example... In this case I have connected all the inputs together and all the outputs together (it's ok to connect outputs together when they all carry the same signal and transition in unison) for more current drive ability. An inverting gate such as the 74LS00 (quad NAND) whoud have one gate by itself, inverting the signal  and the rest connected in parallel to re-invert the signal and do the buffering.

74LS aren't the only type which can be used. There are 74HC and 4000 series parts but I'll leave the topic here...


tomwaits,

An LM1881 will almost certainly get you a good sync signal. Make sure you connect the input of you LM1881 circuit to pin 11 or the composite video output. A full sized C-Sync signal from pin 10 may cause problems.

Pictured - 74LS32 wired as a simple buffer:

[RGB32E]

Thanks!    So, you made the assertion that the output from pin 11 wouldn't work, yet I've used that output and it works just fine...? 

I don't know the monitor so I can't say it simply won't work. It isn't a valid TTL level signal (it never goes below about 1.8v) so you should not be surprised if it doesn't work but anything is possible if the monitor has tolerant sync input.

What specific buffer circuit were you implying when stating this:

The TTL input to the encoder  chip can be used but it's best to buffer it first.

Open up your bin of parts and fish out some chip marked 74LS??. Most logic gates can be used provided they don't have open collecter outputs and you understand boolean logic. Take a 74LS32 (quad NOR) for example... In this case I have connected all the inputs together and all the outputs together (it's ok to connect outputs together when they all carry the same signal and transition in unison) for more current drive ability. An inverting gate such as the 74LS00 (quad NAND) whoud have one gate by itself, inverting the signal  and the rest connected in parallel to re-invert the signal and do the buffering.

74LS aren't the only type which can be used. There are 74HC and 4000 series parts but I'll leave the topic here...


tomwaits,

An LM1881 will almost certainly get you a good sync signal. Make sure you connect the input of you LM1881 circuit to pin 11 or the composite video output. A full sized C-Sync signal from pin 10 may cause problems.

Pictured - 74LS32 wired as a simple buffer:

Ahh... that's along the lines of what I thought you were implying.   Similar to the hex buffer contained on my Redmond Cable buffer PCBs:

[tomwaits]

I built the LM1881 circuit and it works fine.  Games looks great.  Took me about an hour to wedge all the parts into a SCART hood, and I burned my finger a little, but it ended up looking pretty neat in there.

Thanks again for the suggestions on trying to get direct sync from the NGCD, but it looks like the LM1881 is the only solution for NGCD on an XM29.

[RGB32E]

I built the LM1881 circuit and it works fine.  Games looks great.  Took me about an hour to wedge all the parts into a SCART hood, and I burned my finger a little, but it ended up looking pretty neat in there.

Thanks again for the suggestions on trying to get direct sync from the NGCD, but it looks like the LM1881 is the only solution for NGCD on an XM29.

I'm glad the LM1881 route worked for you!  Also, if you ever get a RGB modded NEC system (Duo, TG16, ect) you'll have to do the same thing (LM1881 circuit / sync separator) to get a stable image. 

[tomwaits]

I'd really like to RGB mod my Duo, but I think it's beyond my soldering/modding ability.  And it would be expensive to replace if anything went wrong.  Don't want to spend the $$ having someone else mod it either.


SNES is up next... I've got an SNES SCART cable but there's no csync wire available to swap onto the cvideo line so I'm hoping I can just pop the console open and somehow re-wire csync onto the cvideo pin at the connector.  Gamebit is on the way.  I've also got a couple extra LM1881's in case I need to go that route.

I might try the N64 RGB mod after that.  I've got a US launch N64 so the mod looks pretty straightforward.  Even if I mess it up, a replacement N64 would be cheap.

Maybe try the new Colecovision component mod that's coming out soon too (into my RGB converter) but that might be too tough for me.  I'm going to wait for the install instructions before deciding on that one.  Don't play Colecovision much anyway so it's not really a priority.