Neo Geo AES Console buying , are all able to be RGB fixed ?

s8n · Post by **s8n** » Thu Aug 10, 2017 7:32 am

hi guys , this is the 3rd forum i've hit needing some help with this issue and haven't got any replies. I need some help with Neo Geo AES Console purchasing i own 1 atm but plan on more. I have read that some have not so great out of the box RGB output but if i'm correct you can fix this. Below is what i found over at Jamma X , so if i read it correctly all model revisions RGB output can be fixed ?

if they can be RGB fixed i can buy any model and modify it without trying to hunt down specific models , i need some confirms before moving forward and purchasing AES's

Ratings are for out of the box RGB not fixed RGB.

1st Version (~3-1)
Rating - B

RGB Fix - Replace 100uf caps with 470uf value (10v is best fit). You do not need to replace the 68ohm
resistors with 75ohm since they aren't 1% resistors but you can replace them for good measure.

1st Revision (~3-2) w/daughterboard
Rating - B

RGB Fix - Same as 3-1.

2nd Revision (3-3)
Rating - C

RGB Fix - Replace 100uf caps with 470uf value (10v is best fit)*, The 3-3 has a low output RGB. If you plan
to mount an internal encoder for Component video, you will need to perform an RGB bypass if you plan to use
the RGB output from the AV socket, if not the RGB picture will be extremely dim.

3rd Revision (3-4)
Rating - C

RGB Fix - Same as 3-3.

4th Revision (3-5)*
Rating - A (Non-Checkerboard) B (Checkerboard)

RGB Fix - Board version 3-5 is an odd one. There is a known checkerboard problem via RGB BUT not all
3-5 consoles suffer from the checkerboard pattern. If yours has the checkerboard pattern then an RGB bypass
is needed.

5th Revision (3-6)*
Rating - B-

RGB Fix - Board version 3-6 suffers from vertical lines in picture. This is caused by the RGB lines running
under the clock. A simple cutting of the 3 extra traces fixes the problem. If not for the vertical line
problem, the 3-6 would have an A rating for it's RGB.

*Revision 3-5 and 3-6 are the two revisions where composite video is on both Pin 3 and 7 of the AV socket
whereas 3-1 to 3-4 have composite video on Pin 3 and Composite Sync on Pin 7.

LDigital · Post by **LDigital** » Thu Aug 10, 2017 9:17 am

I'm not an expert but I believe that you can buy an specific neo geo rgb amp to improve all models.

FBX · Post by **FBX** » Thu Aug 10, 2017 5:10 pm

I sort of got a little burned by some general incomplete information about the "good" RGB Neo Geos versus "bad" ones. The web page I referenced went by serial numbers and only had a very small sample to work with. This ended up causing me to get literally the WORST version for both audio and RGB video: AES3-3.

Turns out the ones to get are the much later AES3-6 consoles, because getting flawless RGB out of them is a simple matter of cutting a few traces to the CXA encoder chip (similar to the jail bars issue with some Model 1 Sega Genesis consoles, where you would sever the video trace to pin #6 to get crystal clear RGB).

The AES3-6 revision also has clear audio, as SNK messed up the stereo audio circuits starting with 3-3 and on into 3-4. By 3-5 they realized their mistake and started adding the missing caps back into the circuit, and by 3-6 everything was kosher again.

So yeah, do yourself a favor and find a 3-6 revision, then do the trace cuts and PRESTO! You've got a perfect Neo Geo. I only wish I had known this before I dumped $250 on the wrong revision.

-FBX

DirkSwizzler · Post by **DirkSwizzler** » Thu Aug 10, 2017 5:14 pm

s8n wrote:*Revision 3-5 and 3-6 are the two revisions where composite video is on both Pin 3 and 7 of the AV socket
whereas 3-1 to 3-4 have composite video on Pin 3 and Composite Sync on Pin 7.

OMG I bet that explains why my 3-6 tends to have the most sync issues of any console I own. I've tried a HDRetrovision NG->Genesis2 adapter with a genesis 2 CSYNC cable that would constantly have issues. And later a CSYNC cable made just for NeoGeo from retrogamingcables.co.uk would occasionally have issues running through some transcoders.

Thanks so much for that info.

FBX wrote:So yeah, do yourself a favor and find a 3-6 revision, then do the trace cuts and PRESTO! You've got a perfect Neo Geo. I only wish I had known this before I dumped $250 on the wrong revision.

The trace fix is indeed super easy. Except for the occasional sync issue, the 3-6 is super good.

s8n · Post by **s8n** » Fri Aug 11, 2017 6:17 am

thanks guys for the info and tips much appreciated , i have emailed 2 Japanese Stores on ebay and haven't got a response. I mentioned looking inside the Cart Slot with a flash light to see the Motherboard Revision number , i am searching for 3-6's based on FBX's tips. Is it a language barrier problem to why i'm not getting any replies with info ? or maybe i should be contacting US sellers ?. I plan on owning multiple AES's and this is a bit frustrating.

DirkSwizzler · Post by **DirkSwizzler** » Fri Aug 11, 2017 8:05 am

s8n wrote:thanks guys for the info and tips much appreciated , i have emailed 2 Japanese Stores on ebay and haven't got a response. I mentioned looking inside the Cart Slot with a flash light to see the Motherboard Revision number , i am searching for 3-6's based on FBX's tips. Is it a language barrier problem to why i'm not getting any replies with info ? or maybe i should be contacting US sellers ?. I plan on owning multiple AES's and this is a bit frustrating.

I could be wrong, but isn't the serial number a pretty good indicator of which model it is? With the 3-6 being the last and most common revision.

s8n · Post by **s8n** » Fri Aug 11, 2017 8:15 am

yes from what i have read , but i have also read that motherboards can be swapped out so yeah i need the seller to check. I did get a message from one of the Japanese sellers about 2 hours ago and they said they don't know what Revision they have and to please buy from someone else. So now i'm looking to buy from US sellers , it's a shame as most of what's on ebay are from Japan but i just have to keep hunting for US sellers.

FBX · Post by **FBX** » Fri Aug 11, 2017 8:19 am

FYI the 3-6 revision started at 157,000 on the serial number. So if you see anything in the 200k or 300k range, it's a 95% chance of being a 3-6. Now I say 95% because there is the rare chance you get a previously serviced model. My own database showed 3 consoles out of 50 above 200k were 3-5 revision due to servicing.

s8n · Post by **s8n** » Fri Aug 11, 2017 8:29 am

that's a good tip thanks for the added info FBX , i have contacted a US seller about an hour ago with a AES with a serial of slightly over 200,000. Hopefully they get back to me saying it's a 3-6 , i even gave a tip and photo where to look inside the Cart Slot to find the Revision with a flash light.

FBX · Post by **FBX** » Fri Aug 11, 2017 8:35 am

Also here's a link on how to get crystal clear RGB on the 3-6 (provided you don't mind losing composite video in the process). It's a very easy mod:

http://www.mmmonkey.co.uk/snk-neo-geo-a ... cal-lines/

Note that he cut both sides of each trace, but you only need to cut one side.

-FBX

s8n · Post by **s8n** » Fri Aug 11, 2017 8:49 am

i do own one 3-6 AES and need to do that , thanks for the link. Btw are you the Firebrand X that does the Framemeister Profiles ?

FBX · Post by **FBX** » Fri Aug 11, 2017 8:51 am

s8n wrote: Btw are you the Firebrand X that does the Framemeister Profiles ?

Me? Nah, I just link to them in my sig out of weirdness.

Gunstar · Post by **Gunstar** » Fri Aug 11, 2017 11:24 am

If you end up with a non 3-6 can you bring it up to par with various amps? Thanks for serial number info, good to know.

FBX · Post by **FBX** » Fri Aug 11, 2017 2:05 pm

So this is sort of strange, but I have a theory on what happened:

The past few days I've been annoyed by a checkerboarding issue on the RGB image of my 3-3 AES. I unplug all the cables and test them with my multimeter. Everything checks out fine. I plug everything back in today and now the RGB is crystal-freakin-clear! I mean beautiful! So my theory is somehow either my power cord or my RGB cable weren't fully seated properly, and re-plugging everything back in fixed the issue. Only thing I can think of anyway. But I'm definitely not going look a gift horse in the mouth on this one!

So now I'm just back to the only issue being the faint 60-cycle hum on the stereo audio up front, but that's a known issue that requires adding in a few caps SNK left out by mistake.

Much happier now that my RGB image is magically clean. I'm running it into my Framemeister right now and it looks sweet!

leonk · Post by **leonk** » Fri Aug 11, 2017 7:06 pm

Jamma X instructions are correct.

I've done a few 3-6 with the simple trace cut with excellent results.

I also had a person give me a neo geo to fix that was dark / loosing sync on his BVM. Replacing the capacitors as per Sony's PDF for their CX chip fixed the problem.

No need for external rgb amp.

RGBSource · Post by **RGBSource** » Sun Aug 13, 2017 6:17 pm

All of the AES revisions require modification to get optimal RGB output. How much depends upon the revision. From the revs I've updated, the results are consistent between revisions.

One thing not yet mentioned in this thread is that the earlier revs actually use CXA1145P pin 11 for CSYNC output (8DIN pin 7). It's easy enough to restore CSYNC on the later revs like 3-6 by disconnecting the ferrite bead jumper to composite video and connecting a 470uF + 75 ohm resistor to CXA1145P pin 11. Using CSYNC out from the CXA1145P should resolve sync issues (in most cases) if you have them.

GamesX/nfggames posted a guide many years ago for how to update any AES version - http://nfggames.com/games/neorgb2/

The two resistors (one pair for each channel) before the Sony encoder are a voltage divider circuit that drops the voltage level from the resistor array. Replacing the existing resistors with 1% tolerance will help ensure level matching between R/G/B channels.

Here's a before/after on a pro-model (early rev that doesn't have 3-1 marking):

Lastly, as leonk mentioned, there's no need to bypass the CXA1145P with another amp or circuit, just update the existing components, and cut a few traces on the 3-6! That said, I don't think that will stop anyone from selling an AES bypass amp circuit.

DirkSwizzler · Post by **DirkSwizzler** » Sun Aug 13, 2017 6:49 pm

RGBSource wrote:All of the AES revisions require modification to get optimal RGB output. How much depends upon the revision. From the revs I've updated, the results are consistent between revisions.

One thing not yet mentioned in this thread is that the earlier revs actually use CXA1145P pin 11 for CSYNC output (8DIN pin 7). It's easy enough to restore CSYNC on the later revs like 3-6 by disconnecting the ferrite bead jumper to composite video and connecting a 470uF + 75 ohm resistor to CXA1145P pin 11. Using CSYNC out from the CXA1145P should resolve sync issues (in most cases) if you have them.

GamesX/nfggames posted a guide many years ago for how to update any AES version - http://nfggames.com/games/neorgb2/

I'm interested in performing the csync mod on my 3-6. I'm still not great at reading circuit diagrams. But I believe the link you posted shows that the CSYNC output also needs a 75ohm pull up resistor. Is that correct? Or is that supposed to be in the cable already? Or am I just reading it wrong?

RGBSource · Post by **RGBSource** » Sun Aug 13, 2017 7:09 pm

DirkSwizzler wrote:
RGBSource wrote:All of the AES revisions require modification to get optimal RGB output. How much depends upon the revision. From the revs I've updated, the results are consistent between revisions.

One thing not yet mentioned in this thread is that the earlier revs actually use CXA1145P pin 11 for CSYNC output (8DIN pin 7). It's easy enough to restore CSYNC on the later revs like 3-6 by disconnecting the ferrite bead jumper to composite video and connecting a 470uF + 75 ohm resistor to CXA1145P pin 11. Using CSYNC out from the CXA1145P should resolve sync issues (in most cases) if you have them.

GamesX/nfggames posted a guide many years ago for how to update any AES version - http://nfggames.com/games/neorgb2/
I'm interested in performing the csync mod on my 3-6. I'm still not great at reading circuit diagrams. But I believe the link you posted shows that the CSYNC output also needs a 75ohm pull up resistor. Is that correct? Or is that supposed to be in the cable already? Or am I just reading it wrong?

While the CXA1145 datasheet does indeed show a pull-up resistor before output in their generic/test example, the early model AES consoles do not include it, so I do not advise adding it if you want "75 ohm"/video level CSYNC. Also, just like the RGB outputs have a resistor to ground in the diagram, these resistors exist on the destination side (TV/video processor).

DirkSwizzler · Post by **DirkSwizzler** » Sun Aug 13, 2017 7:28 pm

RGBSource wrote:
DirkSwizzler wrote:
RGBSource wrote:All of the AES revisions require modification to get optimal RGB output. How much depends upon the revision. From the revs I've updated, the results are consistent between revisions.

One thing not yet mentioned in this thread is that the earlier revs actually use CXA1145P pin 11 for CSYNC output (8DIN pin 7). It's easy enough to restore CSYNC on the later revs like 3-6 by disconnecting the ferrite bead jumper to composite video and connecting a 470uF + 75 ohm resistor to CXA1145P pin 11. Using CSYNC out from the CXA1145P should resolve sync issues (in most cases) if you have them.

GamesX/nfggames posted a guide many years ago for how to update any AES version - http://nfggames.com/games/neorgb2/
I'm interested in performing the csync mod on my 3-6. I'm still not great at reading circuit diagrams. But I believe the link you posted shows that the CSYNC output also needs a 75ohm pull up resistor. Is that correct? Or is that supposed to be in the cable already? Or am I just reading it wrong?
While the CXA1145 datasheet does indeed show a pull-up resistor before output in their generic/test example, the early model AES consoles do not include it, so I do not advise adding it if you want "75 ohm"/video level CSYNC.

Ah, makes sense. I can't quite figure out how to interpret the "red marks show differences" notation. Thanks!

RGBSource · Post by **RGBSource** » Sun Aug 13, 2017 7:31 pm

No problem! The GamesX modified image for the "Old system encoder circuit" even has pin 11 with a red X! ¯\_(ツ)_/¯

Also, here's the application circuit from the datasheet. GamesX page used the "Electrical Characteristics Test Circuit" instead of the Application Circuit diagram, which is why the extra/external components are present.

FBX · Post by **FBX** » Sun Aug 13, 2017 10:24 pm

RGBSource wrote:
Lastly, as leonk mentioned, there's no need to bypass the CXA1145P with another amp or circuit, just update the existing components, and cut a few traces on the 3-6!

Note he glossed right over my posts that even linked the trace cut method, and pretended Leonk first mentions it.

DirkSwizzler · Post by **DirkSwizzler** » Mon Aug 14, 2017 12:01 am

FBX wrote:
RGBSource wrote:
Lastly, as leonk mentioned, there's no need to bypass the CXA1145P with another amp or circuit, just update the existing components, and cut a few traces on the 3-6!
Note he glossed right over my posts that even linked the trace cut method, and pretended Leonk first mentions it.

Clearly because you're an impostor FirebrandX

s8n · Post by **s8n** » Mon Aug 14, 2017 9:50 am

All of the AES revisions require modification to get optimal RGB output. How much depends upon the revision. From the revs I've updated, the results are consistent between revisions.

RGBSource , does this mean the guides i posted at the start of this thread which i got from Jamma X are the remedy to get good RGB from AES's ?

RGBSource · Post by **RGBSource** » Mon Aug 14, 2017 4:23 pm

s8n wrote:
All of the AES revisions require modification to get optimal RGB output. How much depends upon the revision. From the revs I've updated, the results are consistent between revisions.

RGBSource , does this mean the guides i posted at the start of this thread which i got from Jamma X are the remedy to get good RGB from AES's ?

They are ok, just incomplete. Here's whats missing (applies to all revisions):
1. Update the input side resistors with 1% resistors, and 1uF electrolytics (negative side away from Sony encoder)
// Example P/Ns 270-6.8K-RC, 270-3.0K-RC, Nichicon UTT1H010MDD

2. Ensure RGBS output (Sync from CXA pin 11, not 10) are all running through new 470uF caps (220uF or 330uF are "probably ok") and 1% 75 ohm resistors to their respective output connector pins.
// Example P/Ns 10ZLQ470MEFC6.3X11, 270-75-RC

shmups.system11.org