exxos's Atari, Amiga & retro forum

PhilC wrote: 02 Dec 2019 18:32 What version of Tos are you trying to boot with Derkom?

That's the fun part in all this. I don't know. Back when the Alpha board was all the rage, Atarian Computing donated his board to me because he didn't have time to continue working on it. I never got it booting before H4 was on the horizon, so I set it aside. So the TOS chip I'm using is whatever he had on that board.

Since that variable is an unknown, I've got an exxos 102/104 chip in the mail, hopefully arriving sometime this week, and then at least we'll know.

You say TOS 'chip'... Surely you mean 'chips' as there should be at least two, unless you have one of the replacement boards created recently.

stephen_usher wrote: 02 Dec 2019 19:01 You say TOS 'chip'... Surely you mean 'chips' as there should be at least two, unless you have one of the replacement boards created recently.

Exxos designed the board with a 27c4096 which has enough room for two versions of Tos, instead of the two 27c256.

Oh, I see!

derkom wrote: 02 Dec 2019 18:40 Okay. The answer is simple, but perhaps not helpful. Everything is low except A22, every bus error.

So 01000000000000000000000 which is 2MB mark, That is in the middle of ST-RAM.. Could be the MMU is trying to size RAM, but I don't get why it would get stuck repeating the same address over and over. Really it should give up and continue booting anyway.. well assuming the MMU found any RAM at all.

I would suggest fluxing all the MMU & RAM pins and just re-heat them all up just to be sure. Also check for bent pins on the MMU itself.

Do you get the same behaviour with the diagnostic cart plugged in ?

exxos wrote: 02 Dec 2019 20:04 I would suggest fluxing all the MMU & RAM pins and just re-heat them all up just to be sure. Also check for bent pins on the MMU itself.

Yeah, I'll redo them all tomorrow. The MMU is definitely fine. I've inspected it more than once, and booted a real ST with it, so it's functional.

Do you get the same behaviour with the diagnostic cart plugged in ?

Looks the same, yeah.

EDIT: It occurred to me that the ST I booted this MMU in has only 1 MB, so if the 01000... address in the middle of ST RAM is actually relevant, it could be a problem that simply doesn't show on a 1 MB ST, so I swapped MMUs with that one, but no change. I don't have any 4 MB STs to test with, but it doesn't seem likely I'd have two MMUs that fail in the same way, so I'll put that thought to bed.

@derkom So what happens with a SIMM instead of the falcon ram board ? any change on bus oddness ?

exxos wrote: 02 Dec 2019 21:22 @derkom So what happens with a SIMM instead of the falcon ram board ? any change on bus oddness ?

Actually I just tried that, since you confirmed that a larger SIMM should be okay. No noticeable difference in behaviour. But it also acts the same way with no RAM at all. It just occurred to me that my earlier tests without RAM were not watching the address lines, so I tried that again. So that kind of suggests that it's not finding any RAM. I don't know how TOS acts in the absence of RAM. Maybe it's just continually looking for it, and continually bus erroring when it finds none?

derkom wrote: 02 Dec 2019 21:27 Actually I just tried that, since you confirmed that a larger SIMM should be okay. No noticeable difference in behaviour. But it also acts the same way with no RAM at all. It just occurred to me that my earlier tests without RAM were not watching the address lines, so I tried that again. So that kind of suggests that it's not finding any RAM.

Can you get your meter probes on resistance and check the MMU DTACK pin (on the IC pin itself in the socket) to a 68K expansion socket and make sure you see a connection there.

exxos wrote: 02 Dec 2019 21:29 Can you get your meter probes on resistance and check the MMU DTACK pin (on the IC pin itself in the socket) to a 68K expansion socket and make sure you see a connection there.

Depending on which expansion port I choose, 1.2 - 1.5 Ω, which I would guess is good enough.