DaytimeDreamer
Southern Pounce.,
- Joined
- Jul 22, 2005
- Posts
- 747
Awesome stuff! Many thanks for this.
Neo Geo AES 3-3 revision: Mods & enchancements - Help needed
- Thread starter DaytimeDreamer
- Start date
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
BTW I spoke at length about my 3-3 bypass mod with an experienced EE on the subject. He gave my mod a thumbs up as being definitely viable because I made sure to attenuate the brightness. He says the only thing is it won't be entirely 'perfect' on very technical and minor details, but that this is far outweighed by the improved visual clarity over the CXA's jail bars. The only thing he suggested adding is some 220uf caps purely for protection against idiocy, but that technically it doesn't need them otherwise. Here are some snippets of his comments:
Driving straight off the DAC should also be within spec, as long as your attenuation is correct.
The Neo Geo DAC output is a bunch of LS gates connected to each other so they are already sourcing/sinking into themselves. So using or not using a series cap would be negligible to performance. It's safer to use a cap, so I would still leave it in. 120uF minimum, but I would use 220uF. Can use 470uF if you want to go crazy. Can also leave out the capacitors like you have it now. They are not necessary. It's just for added protection in case some dum-dum heads hook things up wrong on the other end.
The trade off for simplicity in driving the video line directly off the DAC gates is huge. You still have impedance mismatches and non-linearity, but it's still better overall than the junk noise you get stock.
And with that, I'm betting the 3-5 bypass mod is subject to the same issues about impedance and non-linearity. However, the resulting picture looks so much better than stock encoder that as I said, the benefits far outweigh the negatives.
So here's more eye candy from my bypass mod optimally timed by the OSSC:
Driving straight off the DAC should also be within spec, as long as your attenuation is correct.
The Neo Geo DAC output is a bunch of LS gates connected to each other so they are already sourcing/sinking into themselves. So using or not using a series cap would be negligible to performance. It's safer to use a cap, so I would still leave it in. 120uF minimum, but I would use 220uF. Can use 470uF if you want to go crazy. Can also leave out the capacitors like you have it now. They are not necessary. It's just for added protection in case some dum-dum heads hook things up wrong on the other end.
The trade off for simplicity in driving the video line directly off the DAC gates is huge. You still have impedance mismatches and non-linearity, but it's still better overall than the junk noise you get stock.
And with that, I'm betting the 3-5 bypass mod is subject to the same issues about impedance and non-linearity. However, the resulting picture looks so much better than stock encoder that as I said, the benefits far outweigh the negatives.
So here's more eye candy from my bypass mod optimally timed by the OSSC:
Last edited:
Thanks so much for all this info FBX I have an AES 3-4 board that is stock and I am thinking of doing both the audio and RGB mod. One question I have is it seems that you are using an OSSC for testing the RGB result. Have you also tried it out on a PVM? Just curious if the results are still good. My aes gives off a dim picture now in RGB. I was fixing a friends neo geo cd and that system with the same scart cable gave off a beautiful RGB picture.
DaytimeDreamer
Southern Pounce.,
- Joined
- Jul 22, 2005
- Posts
- 747
BTW I spoke at length about my 3-3 bypass mod with an experienced EE on the subject. He gave my mod a thumbs up as being definitely viable because I made sure to attenuate the brightness. He says the only thing is it won't be entirely 'perfect' on very technical and minor details, but that this is far outweighed by the improved visual clarity over the CXA's jail bars. The only thing he suggested adding is some 220uf caps purely for protection against idiocy, but that technically it doesn't need them otherwise. Here are some snippets of his comments:
Driving straight off the DAC should also be within spec, as long as your attenuation is correct.
The Neo Geo DAC output is a bunch of LS gates connected to each other so they are already sourcing/sinking into themselves. So using or not using a series cap would be negligible to performance. It's safer to use a cap, so I would still leave it in. 120uF minimum, but I would use 220uF. Can use 470uF if you want to go crazy. Can also leave out the capacitors like you have it now. They are not necessary. It's just for added protection in case some dum-dum heads hook things up wrong on the other end.
The trade off for simplicity in driving the video line directly off the DAC gates is huge. You still have impedance mismatches and non-linearity, but it's still better overall than the junk noise you get stock.
And with that, I'm betting the 3-5 bypass mod is subject to the same issues about impedance and non-linearity. However, the resulting picture looks so much better than stock encoder that as I said, the benefits far outweigh the negatives.
So here's more eye candy from my bypass mod optimally timed by the OSSC:
Again, great info right there. What are your optimal OSSC timings for your Neo Geo 3-3 by the way?
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
When you do this bypass mod, all issues of dim RGB picture go right out the window. The 143 Ohm resistance gives a fantastic picture, and if you decide it's too bright, simply add a little more resistance. Too dark, add a little less. Really though, I found the 143 total resistance (the 75 Ohm plus the 68 Ohm stock combined) landed superbly on a good brightness level for my 3-3. I cross-referenced this by checking for color crushing on the Framemeister's A/D menu feature, and the default setting was already perfect. That said, I've not bothered to try it on my PVM just yet.
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
They actually got wiped out when I upgraded the firmware, so I had to start all over again. Going forward, I'm starting a new text file format download from my web site for all OSSC profiles. Neo Geo is now the first under the new format:
OSSC - Neo Geo Optimal Timing
And of course new eye candy from the freshly calibrated OSSC timings:
Cheers!
DaytimeDreamer
Southern Pounce.,
- Joined
- Jul 22, 2005
- Posts
- 747
Great, thanks for sharing those settings. Don't you like the scanline look? I'll experiment a bit with scanlines but first I must update my OSSC firmware. I'm still on 0.76 
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
Great, thanks for sharing those settings. Don't you like the scanline look? I'll experiment a bit with scanlines but first I must update my OSSC firmware. I'm still on 0.76
Not fan of artificial scanlines. Just not my cup of tea. Scanlines are for PVMs. Modern displays, I want eye-cutting sharp pixels.
Just a heads up in that you may have to slightly adjust the resistance on the RGB lines for the 3-4, The voltage dividers were measured at slightly different levels than the 3-3. Here's a chart by "arcadeTV":
R1 (in series), R2 (to GND)
AES3-3: R1=6.65K, R2=2.35K
AES3-4: R1=7.1K, R2=1.9K
AES3-5: R1=7.51K, R2=2.29K
AES3-6: R1=7.94K, R2=2.06K
Edit: I realize my mod removes this part of the dividing line, so it may not matter and 143 Ohms will still work as best even on the 3-4. However, just keep in mind you might need to change the resistance if it turns out too bright or too dark. Too bright = not enough resistance. Too dark = too much resistance.
Last edited:
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
Finally got around to hooking it up to my 20M2U and it looks frikken beautiful! Bright and vivid colors on default dial settings.
That said, remember again that the brightness attenuation is simply a matter of changing those resistor values. So if your 3-4 ends up looking too bright or too dark, you know what to do. :-)
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
I updated my web page article (hit refresh on the browser to make sure it's updated) to show my installation of 220uf caps in the SCART head of my RGB cable. This is to provide protection against voltage backflow from failing equipment, and does not change the picture quality or brightness at all:
Neo Geo AES3-3 RGB Bypass Mod
Neo Geo AES3-3 RGB Bypass Mod
Followed the rgb bypass and the sound fix on my 3-4 AES and the result were amazing. I did not have to change the resistors 75ohm worked great. Before I did the mod the rgb I was getting was pretty dim.
My system is perfect now and I'm enjoying cyber lip on the neosd in beautiful rgb. Just for reference I am using a Sony crt that has been modded for rgb.
Thanks again for sharing your findings.
Last edited:
FBX
Quiz Detective
- Joined
- Aug 7, 2017
- Posts
- 89
Just bumping this with actual scope readings now that I have a Rigol 4-channel scope. So first up is the csync measurement of about 325mV (tolerance range for SCART is 220mV to 950mV):
And then I showed all three RGB channels at the same time under peak output with a reading close to 690mV (anything in the 600mV to 700mV range is considered safe peak output ):
So now my bypass method for the 3-3 and 3-4 is actually backed up by scope measurements. And speaking of revisions, I Googled photographs of an actual 3-2 board somebody found, and it looked identical to the 3-3 and 3-4 layout as far as I could tell, so likely the RGB and Audio bypasses would be the same process on the 3-2.
And then I showed all three RGB channels at the same time under peak output with a reading close to 690mV (anything in the 600mV to 700mV range is considered safe peak output ):
So now my bypass method for the 3-3 and 3-4 is actually backed up by scope measurements. And speaking of revisions, I Googled photographs of an actual 3-2 board somebody found, and it looked identical to the 3-3 and 3-4 layout as far as I could tell, so likely the RGB and Audio bypasses would be the same process on the 3-2.
Last edited: