exxos's Atari, Amiga & retro forum

stephen_usher wrote: 19 Jun 2021 20:54 I s'pose if the only issue is that one pin on the ULA and it's controlled by two inputs to the ULA that you could possibly put a NAND gate on A14 and A15 with the output on the ROM output enable to replicate the functionality. Just a thought.

That's a pretty good idea, I'll think about that for sure.

Those ULAs are pretty strong pieces, not easily destroyed.
Saying that, if PSU died in a bad way...

There's actually a complication in the memory map, see: http://www.acornelectron.co.uk/ugs/elec ... 3_eng.html

There's some memory mapped I/O between 0xFC00 and 0xFF00 so it's not a simple "ROM above 0xC000" address decoding. So there's a bit of thought needed. Maybe you'll need a GAL rather than simple TTL logic.

arkadiusz.makarenko wrote: 19 Jun 2021 21:29 Those ULAs are pretty strong pieces, not easily destroyed.
Saying that, if PSU died in a bad way...

Yeah... Maybe look at the CPU first. They're the easiest to socket and replace.

(And maybe just see ig A14 and A15 toggle at all with your nasty 'scope as an initial test)

arkadiusz.makarenko wrote: 19 Jun 2021 21:29 Those ULAs are pretty strong pieces, not easily destroyed.
Saying that, if PSU died in a bad way...

I can't say when the caps in the PSU board died, but the caps that failed mean no power, so I think it unlikely that would have killed the ULA.

stephen_usher wrote: 19 Jun 2021 21:35 There's actually a complication in the memory map, see: http://www.acornelectron.co.uk/ugs/elec ... 3_eng.html

There's some memory mapped I/O between 0xFC00 and 0xFF00 so it's not a simple "ROM above 0xC000" address decoding. So there's a bit of thought needed. Maybe you'll need a GAL rather than simple TTL logic.

That's starting to get a bit beyond my capabilities with a GAL. Not sure if you noticed on the *. thread, but the schematic is wrong and shows pins 20/22 of the ROM connected to the wrong places.

What occurs to me are the following possibilities:

1. The ULA is dead and the ROM select is stuck high.
2. The ROM has an internal short causing pin 20 to be high, and by association, pin 61 of the ULA.
3. There is another direct short to 5v on the pin 20/pin 61 track that I haven't found yet. I had this on my BBC B.
4. The CPU is bad and triggering the ULA to go high on pin 61. My BBC B had a dead CPU, so it's not entirely impossible.

I think next I'll go on a short hunt and have a look at the A14/15 lines and see what they do. If they aren't behaving as expected, then I think that is potentially cause to replace the CPU, and as you say, that's a relatively easy job. My Nasty-o-scope should at least be able to see A14/15 toggling.

It's unfortunate that I'd need to cut out the CPU to safely desolder it with the equipment I have to hand, but at least I would avoid damaging the board and I do have a socket to hand and the CPU from one of the Beebs.

Ok, so A14 stays low no matter what I do.

I've tried the Nasty-o-Scope, which clearly can catch a state change as it does it just fine for reset, which goes low on pressing break. I can set a suitable trigger too so I don't have to see it happen it on screen.

Again, what is happening isn't clear, but an indication of a change.

I can see activity on A15, it's difficult to catch with the N-o-S.

But, why would A14 stay low. There is nothing at all happening, which would mean the ROM line of the ULA would never go low, and the ROM would never be selected.

I know the N-o-S is not great for this, but I cross checked with my logic probe and the multimeter and A14 is permanently low and never changes. I am starting to suspect the CPU.

Assuming the BBC behaves similarly, I can see activity on A14/A15 and there is a blip when I hit break. I am starting to suspect the CPU, especially given what was said on*. about that image being consistent with no CPU.

Also, hooking up the frequency counter to A14/A15 I get nothing at all on A14, but get a count on A15. On the BBC B I get a count off both.

That is something I can do something about, and I don't really have anything to lose.

@stephen_usher Any thoughts?

On a side note I cannot find any shorts to ROM pin 20 or ULA pin 61, so that's one thing ruled out.

It does sound like the CPU doesn't it? If you dare you could piggyback a good CPU on the current one or just chop it out and fit a socket and try a known good. That seems to be the only way to rule the CPU out entirely and 6502s are still available.

At this point it sounds like the least risky option.

P.S. Hmm..if A14 is being pulled low then a piggyback won't help. Just read hoglet's post on *. and checking the resistance from A14 to ground and _5V would probably be a good idea. I'm guessing that there should be something pulling A14 up, such as a resistor network in the ULA. It's tricky to diagnose with the CPU in circuit.

stephen_usher wrote: 20 Jun 2021 13:03 It does sound like the CPU doesn't it? If you dare you could piggyback a good CPU on the current one or just chop it out and fit a socket and try a known good. That seems to be the only way to rule the CPU out entirely and 6502s are still available.

At this point it sounds like the least risky option.

P.S. Hmm..if A14 is being pulled low then a piggyback won't help. Just read hoglet's post on *. and checking the resistance from A14 to ground and _5V would probably be a good idea. I'm guessing that there should be something pulling A14 up, such as a resistor network in the ULA. It's tricky to diagnose with the CPU in circuit.

I did already try piggy backing a while back, but no difference, so I'll hop on over to *. and have a read of what hoglet said. and get on it once I've called my Dad for Fathers Day.

Will report back later.

:dualthumbup: