Palette RAM Error 1st Gen AES*FIXED*

Got a nice green screen on a 1st gen AES.

Sony 5864s test good, so do the latches and gates.

The chip state is a bit weird:

1-H
2-H
3-P
4-P
5-P
6-P
7-P
8-P
9-P
10-P
11-P
12-P
13-P
14-L
15-P
16-P
17-P
18-P
19-H
20-L
21-L (A10)
22-L
23-L (A11)
24-L (A9)
25-L (A8)
26-H
27-H (WE) > LS32 (Good) > PRO-C0
28-H

Edit: All the lines that are low are coming from the PRO-B0.

Is this the state of both 5864s? It seems strange that pin2 (A12) and pin19 (IO8) are stuck high - especially pin19 if it is stuck high on both 5864s. Does your logic probe have a memory setting or something that can tell you if /WE goes low at least once? If /WE does not go low even once then the palette RAM is not even being tested by the BIOS. Perhaps the UniBIOS or smkdan's Diagnostic BIOS would be useful here?

mjs256 said:

Is this the state of both 5864s? It seems strange that pin2 (A12) and pin19 (IO8) are stuck high - especially pin19 if it is stuck high on both 5864s. Does your logic probe have a memory setting or something that can tell you if /WE goes low at least once? If /WE does not go low even once then the palette RAM is not even being tested by the BIOS. Perhaps the UniBIOS or smkdan's Diagnostic BIOS would be useful here?

Click to expand...

Both are the same, the palette ram has two banks but everything is connected together except the address lines I believe.

I tried the D-Bios and just got colored blocks, with the Unibios I just get a pink screen which is supposed to mean bios self test failure which isn't true since the stock bios gives a green screen and all bios connections are 100%.

I'll check the /WE and see if it starts low. I just hope the B0 isn't bad.

So there's no resetting? It just stays stuck at one of those screens? Not sure how much you can trust what you see on screen.

What do you hear when you put in the sound cart with diagnostic BIOS? Do you atleast hear the beeps at the very start? Any sound codes coming after that? If you hear the beeps at the start but nothing after, wait about 2 minutes since a certain error can cause it to take a while.

smkdan said:

So there's no resetting? It just stays stuck at one of those screens? Not sure how much you can trust what you see on screen.

What do you hear when you put in the sound cart with diagnostic BIOS? Do you atleast hear the beeps at the very start? Any sound codes coming after that? If you hear the beeps at the start but nothing after, wait about 2 minutes since a certain error can cause it to take a while.

Click to expand...

Ah, didn't even use the sound testing. I'll try that.

No resetting at all, just.. I'll take a pic.

ok so palette RAM /WE never goes low, even on reset.

On the bios itself I am getting low for:

17-Q2
33-A11
34-A12
35-A13
36-A14
37-A15

That's with the D-Bios in the socket. Don't know if that's normal.

As for sound beeps, they are VERY soft for some reason and the only thing I could make out after about 25 resets was either 11001 or 1001, sometimes it sounded like 0101 and sometimes 100111. Hard to say.

Now the screen is different as well. It starts of with just black and red blocks and now the red blocks dissolve and leave a pure black screen until I reset again.

I think the green screen you were seeing was the result of an "upstream" problem and not the system trying to tell you that a palette error was found. From what you said it seems like the system is not even getting far enough through the startup sequence to test the palette RAM. You might try lifting one of the pins on the SMD video RAM to see if the screen changes to the video RAM error color (though I am not sure which is tested first - palette or video RAM). Honestly I don't understand why they did not make the AES BIOS self-test display text like the MVS BIOS does. Have you checked the state of the main address and data bus lines at the 68k and also the status pins of the 68k? I am currently working on a USA model AES with a minor sound glitch (pretty sure it's 1st gen) so if you find anything suspicious I can check out any logic states on mine for comparison.

On the 68k itself the data lines are pulsing but the upper address lines are low and they are connected to the PRO-B0.

Xian Xi said:

Both are the same, the palette ram has two banks but everything is connected together except the address lines I believe.

I tried the D-Bios and just got colored blocks, with the Unibios I just get a pink screen which is supposed to mean bios self test failure which isn't true since the stock bios gives a green screen and all bios connections are 100%.

Click to expand...

The unibios does not do any tests on an AES system with no cart inserted. You need to insert a cart and use the button code to run the hardware tests.

Razoola said:

The unibios does not do any tests on an AES system with no cart inserted. You need to insert a cart and use the button code to run the hardware tests.

Click to expand...

That was with a cart installed. The command wouldn't even register.

Some progress. I found a very hard to see broken trace that apparently ran from the bios to the palette ram, I'm guessing that's why it wasn't in watch dog but still had problems.

So now it boots with the Neo Diagnostics and I get this error:

Z80 Dead/Error/Comm Issue
Exp: C3
Act: 50
Z80 Reported Error Code: 10
YM2610 I/O Error

Probably why I can barely hear anything.

With the UniBIOS can you bypass the Z80 error and see if it works (but with no sound of course)?

Correction. The bad trace was from the bios to a 245 for the palette ram.

As for the sound it has been fixed. The trace from YM2610(SDD7) to Z80 (D7) was broken near the cart slot.

It is so hard to see broken traces on these 1st gen board if it's a hairline break.

Thanks for the help, everything is now working properly. All tests passed.

Very strange, I also have a 1st gen AES I just fixed and again, one of the traces from one of the 245s attached to the palette RAM to the BIOS was severed.

RetroRepair said:

Very strange, I also have a 1st gen AES I just fixed and again, one of the traces from one of the 245s attached to the palette RAM to the BIOS was severed.

Click to expand...

Maybe the trace was weak from the beginning and using it just broke it? inserting and removing carts from the slot can bend the board a bit, it puts mechanical stress on it, maybe that did it. Post pics of the board/trace (we like pics in the tech section).

Regards.