What's new
By:
Filters

AES plays blind - sound, no video

kazuo

kazuo

Champion
Joined
Sep 25, 2015
Messages
1,252
Reaction score
1,264
Location
United States / Thailand
Taking a break from another AES I've been working on to work on yet another dead AES I somehow inherited... long story. ;)

When I got it, it just showed a black screen. Opening it revealed corrosion in the solder mask across various areas, mainly around the LSPC2 and the G0 customs. I scraped away the mask, confirmed continuity, then covered everything up with solder mask. I ended up reflowing most of the LSPC2 since some of the pins looked dodgy (but they were all solid when probed) and I did find a few broken traces between the cartridge slot and G0, so once I was done with that, I fired it up and got sound, but still a black screen.

I've pretty much gone over most of the board and nothing stands out to me as suspect, and I'm a little hesitant to do anything else since it does seem to work, it's just playing blind. There does appear to be some occasional white noise at the bottom edge of the screen, as well as some faint black jailbars that come and go, but other than that, no image whatsoever. Pressing various areas of the PCB has no impact on what's going on. The obvious stuff I already tried:
  • Cart slot blasted with deoxit; no change
  • Tried a different cart; no change (both are known to work)
  • Booted with no cart; no change (i.e. black screen)
  • Cables and Power are known to work on another AES
  • Visual inspection of the resistor array, video encoder and AV jack; nothing suspicious found
Does anyone have any suggestions on what to do next? Video encoder? Bad LSPC2?

Thanks!
 
Last edited:
I haven't (yet) had to debug a broken AES, but my first thought would be to work logically back from the A/V connector. Test either side of the video encoder as it seems to sit across all of the video signals, if it doesn't look like its getting fed anything, work back from there; it looks like there's a DAC circuit with some 74xx logic involved.

Encoder: https://wiki.neogeodev.org/index.php?title=File:Neogeo_aes_schematics_pal_2-page-007.jpg
Video DAC: https://wiki.neogeodev.org/index.php?title=File:Neogeo_aes_schematics_pal_2-page-002.jpg
 
Thanks for the pointers. Unfortunately I am limited to a meter; I don't have a scope or a logic probe at the moment (and I definitely cannot get a scope lol, no space). I did order a logic probe which should be here.. eventually though.

Could I just probe around with the meter and check resistance around the encoder? My ability to parse these diagrams is admittedly a little weak.
 
A hail-mary test would be to check it on a different display if you haven't yet. Sometimes the AES will have capacitors & resistors on the output lines leading to the video cable connector that are out of spec for NTSC. I ran into some televisions that would not display properly until I swapped in 470uf and 75ohm on the console.

A scope would help, even a cheap Chinese kit you could get for $40-$50 (or maybe even less.)

Would suggest getting a read of the waveform for the pin of the LSPC2 that is supplying CSYNC to the input pin of the CXA video chip if you verify the CXA is good.
 
Do you know which caps & resistors I would need to swap out & where? I can probably do that pretty quickly and I can get the parts easily from a local shop. I suspect that this probably isn't the issue since I have another 3-6 board that syncs perfectly fine to this display, but I'm down to try it since it isn't a lot of work and doesn't cost much. Unfortunately I do not have another display I can try and it is pretty much not gonna happen for various reasons.

If you have any suggestions on a cheap and tiny scope, I'm all ears. If I don't figure anything out in the near future and I find myself with more time, I may just swap the encoder over from the other AES to quickly rule it out as the issue.
 
It's the 100 uf/68 ohm set laid out in the encoder diagram Bodgit posted: C36, C37, C38, C39, C40, R65, R64, R63, R62, R61

Note that this circuit is probably from the earliest AES model. It would be best to compare the values in your other 3-6 board; If that one works on your lone display and has the same caps/resistors as your faulty board then I would rule that out being a problem.

You NEED to verify the CXA encoder's working (or not) before you just yank it out. The vias on AES PCBs are extremely fragile and I don't recommend just 'swapping' the encoder from one to the other or you're going to have a bad time. Leave the one in your working board alone and buy new ones on eBay, replacement CXA1145 chips are cheap. If you remove it I would suggest clipping the legs of the chip with a flush cutter and individually heating the remaining soldered-in pins with an iron tip to pull them out gently with tweezers before solder-sucking or cleaning the vias with copper braid.

I have no experience with this scope but this would be a start:

https://www.amazon.com/FNIRSI-DSO152-Oscilloscope-Automotive-Bandwidth/dp/B0C6XPVLPZ/

For an older but more verified brand/model you could check out a used Velleman HPS10. Those average $30.
 
I guess I misunderstood what you meant about the component swap, I thought you were saying that you used values other than factory. Both boards look identical so that’s probably not the issue.

I’m painfully familiar with how fragile AES boards are, this isn’t my first rodeo. I’ve gotten pretty comfortable ‘swapping’ components on these using low melt solder and low temp settings on my iron, so not really concerned about causing damage. Ordering a new encoder is tricky for a variety of reasons, otherwise I’d just do that. Trust me, ‘swapping’ encoders around is not my first choice, but I have to make do given my current situation.
 
If possible, try the system on another display or upscaler. Some perfectly working systems give no image at all on my Framemeister because there's something slightly off with the sync.
 
Any idea if it plays nice with a Tink 2X?

I think I can probably get one cheap to see what's up, but I've seen conflicting reports about whether or not it syncs correctly. I'd be using composite and not RGB for this, but I do wonder if it'd be worth picking up a cheap RGB cable and trying it that way as well (watch it be wired for composite, because of course it would be :)).
 
Magic Trashman said:
A scope would help, even a cheap Chinese kit you could get for $40-$50 (or maybe even less.)

Would suggest getting a read of the waveform for the pin of the LSPC2 that is supplying CSYNC to the input pin of the CXA video chip if you verify the CXA is good.
Click to expand...
Your CXA is at least partly working if you have working audio as the encoder diagram shows it has an audio in/out pair so it's passing audio at least.

Might be that you're just missing SYNC, with only a multimeter you can try checking continuity between the LSPC2 and the SYNC IN pin on the CXA. I couldn't see where SYNC was being generated at first glance but I see now it's coming off pin 150 on the LSPC2 and should go to pin 10 on the CXA.

If you do start swapping chips, and they're DIP package, (which I think the CXA is on the AES), try and fit a socket so you only have to (de)solder once.

A scope will make things a lot easier if you can get one, a SYNC signal is a pretty well-defined waveform.
 
Thanks for the insight! Definitely planning to socket it; I picked up a few 24 pin sockets to use if I do get around to trying a swap at some point. I'll probably reflow all the customs and buzz out the encoder before I do anything else.

I'll circle back once I've had a look!
 
UPDATE: Confirmed continuity between the LSPC2 (pin 150) and the CXA (pin 10). I also managed to get my hands on the scope that Trashman recommended (thanks, BTW!) and here's what I ended up with:

aes_scope.jpg


Got the same thing on both ends. The other AES I have looks exactly the same on the scope as well. I tried to see if I could snoop on the CVIDEO out (pin 20 on the CXA) but didn't get anything - do I need to put a resistor in-line, or perhaps just check right off a composite cable?

I tried a Tink 2X that I was able to borrow from a friend - no video in 2X (line double), no video in pass-through. The other AES works perfectly in either mode. I don't think the display is the issue.
 
As you also have a working AES, try probing all of the pins on the CXA on both systems and compare them.

Ideally if you can display some sort of static test pattern it makes comparisons easier.
 
Unfortunately I don't think I have any way of doing that. All I can do is a drop a cart in there and hope for the best.

I did try probing around the encoder and didn't notice anything unusual, so I'm back to zero at this point.
 
You must log in or register to reply here.
Back
Top