U25 and U31 (the Sharp branded chips) right under the Yamaha chip are the DRAMs for video stuff. There should be plenty of suitable parts to use although I cannot give any exact part numbers. 64Kbit sized and 4bits wide, 120ns or faster access time.

But the stuff on the images suggests some kind of an actual access fault rather than just bad memory, things look shifted instead of everything in correct places but with garbage all over.

Tiido wrote on 2023-10-16, 16:32:

U25 and U31 (the Sharp branded chips) right under the Yamaha chip are the DRAMs for video stuff. There should be plenty of suitable parts to use although I cannot give any exact part numbers. 64Kbit sized and 4bits wide, 120ns or faster access time.

But the stuff on the images suggests some kind of an actual access fault rather than just bad memory, things look shifted instead of everything in correct places but with garbage all over.

Would you still go with socketing/replacing those chips first and seeing what happens?

Socket won't do harm, any replacement parts might be having bad bits in them too so it is good to be ready for that.

Tiido wrote on 2023-10-17, 04:42:

Socket won't do harm, any replacement parts might be having bad bits in them too so it is good to be ready for that.

Welp you were right. Replacement chips came from eBay faster than I expected. I socketed the old ones and put the new ones in, and the behavior is exactly the same. Even if the new chips had faults, I would expect there to be some kind of variation. Same errors in checkit diagnostics too.

Any idea where I go from here? I don’t know how to narrow things down at this point, and there sure is a lot of logic on the board.

I did figure out that this card has an MDA mode that can be activated with DIP switches and it seems to be Hercules compatible, but still has the same behavior.

In that case you would want an oscilloscope and start looking at the individual signals, perhaps some output from a logic chip is bad. Perhaps finger test can help too, since bad signals usually will be affected by presence/contact with a finger and it can help to zone in on where the fault may be. Temperature can also help, faulty patrs are often warmer or even hot compared to others (sometimes they can stay cold too).

Tiido wrote on 2023-10-18, 05:19:

In that case you would want an oscilloscope and start looking at the individual signals, perhaps some output from a logic chip is bad. Perhaps finger test can help too, since bad signals usually will be affected by presence/contact with a finger and it can help to zone in on where the fault may be. Temperature can also help, faulty patrs are often warmer or even hot compared to others (sometimes they can stay cold too).

I felt around the board after leaving it on for a while and nothing felt particularly hot, nor did wiggling the components change anything. I did verify with an external cga monitor the same behavior.

Unfortunately getting to the point of using an oscilloscope for this is beyond me.

You can still slide fingers over the pins of every chip except the pins where the 6MHz crystal connect to, since the finger will disrupt clock signal generation and the card will definitely misbehave when that is happening 🤣.
If you find something that causes the garbage to be affected predictably you have something to analyze further.

Tiido wrote on 2023-10-18, 14:37:

You can still slide fingers over the pins of every chip except the pins where the 6MHz crystal connect to, since the finger will disrupt clock signal generation and the card will definitely misbehave when that is happening 🤣.
If you find something that causes the garbage to be affected predictably you have something to analyze further.

Not so lucky. I’ve gone over every component while running, tapped on everything I could, and don’t see any differences in real time.

It is possible the biggest chips are at fault internally but without some proper measuring capability it will be difficult to get anywhere further unfortunately.

You can glean some insight from the datasheet ~ https://z3d9b7u8.stackpathcdn.com/pdf-down/V/ … 6366_Yamaha.pdf