exxos's Atari, Amiga & retro forum

stephen_usher wrote: 18 Jun 2024 10:25 Yay! We have life!

So, it's likely that the transceiver between the CPU and the RAM isn't working correctly. It may have had the D2 line on the CPU side pulled low but that's changed to not working at all.

@terriblefire do you agree?

If so then you'll need to find a replacement 74F245. The 'F' is important as it needs to be the "Fast" version of the chip.

There are two transceivers. Its unlikely both stopped working at the same time?

Does the DiagROM output hint at the top/bottom or both halfs of the data bus failing? If its both i'd trace down those enable lines as 2 chips failing at the same time is unlikely.

Kepler wrote: 18 Jun 2024 10:43 Are you not just slightly curious as to the reason why Diagrom is now being recognized?

If you've been handling the board during diagnostics, something could have shifted and made a bad connection work.

Best to ignore that for now (Unless it stops working again) and continue to investigate the original issue as it's best to investigate one thing at a time.

Indeed. And you can't really diagnose a fault if it's not there either...

And DiagROM not finding the RAM could be the same fault but with a different symptom.

e.g. if the transceivers are being enabled/disabled incorrectly and the bus behind them happens to start off with D2 low but the other lines high then that could explain the problem with D2 being low and DiagROM not starting. It's just that in this case the system initialised with all the data lines high (as chips ground the data lines but to make a zero but they don't feed voltage in to bring them high, that's done by the pull-up resistors) and hence the system could run.

OK. Next step I think will be to verify if Gary is seeing the address bus correctly and if he's outputting anything on the transceiver/latch control lines.

First of all, have a look at pins 33 to 39 on Gary. These are the address bus lines. Look to see if there's anything abnormal looking. It can be hard to see as after a reset not all the address lines may be actively high. If there's a pin "floating" then that would suggest a bad connection.

If these all look OK have a look at pins 3 and 25. These control the transceivers and the latches. Pins 3 and 25 control the latches. If pin three is low then the latches will be outputting data onto the CPU data bus, so it should be high until the system tries to read the memory, at which point they will have been programmed with the correct data. Before that point they'll have random data so if they're enabled they could pollute the data bus. Also check for continuity between pin 3 of Gary and Pin 1 of U106 and also Pin 25 of Gary to Pin 11 of U106.

Is Gary in a socket? It's not unknown for these sockets to be a problem on Amigas (there are a few YouTube videos where this turned out to be the case), so this could possibly be your problem too. If you have some contact cleaner it may be a good idea to remove Gary, put contact cleaner on the socket and reseat the chip and test.

P.S. I've just noticed that the transceivers are being used as uni-directional bus isolators only, a very strange choice, so unless they aren't getting 5V on Pin 1 they can't be a cause for the data bus pollution (unless they're really broken).

Yes, Gary is in a socket and that seems to be pristine.
Could I just clarify, by "transceivers", do you mean 103-106?

We're expecting one of U103 to U106 (as seen on the right here) to perhaps have an issue.

Kepler wrote: 18 Jun 2024 15:08 Yes, Gary is in a socket and that seems to be pristine.
Could I just clarify, by "transceivers", do you mean 103-106?

I'd still reseat Gary at the very least. Looks can be deceiving as there can still be enough oxidation of the surface of the pins and socket leaves to stop them making good contact. It's a relatively safe thing to do anyway.

P.S. With all retro equipment, if they stop working the first thing I do is reseat all the socketed chips. More than half the time it gets them working again. :-)

OK. I will start checking those pins...
If I don't respond quickly, I apologize in advance...I am at the same time trying to fix my A3000 that has started randomly freezing and my 1701 that no longer powers on...with these classic machines and monitors, it is only a matter of when and not if they are going to fail!

Reseat the chips. ;-)

It's working! I don't quite believe it... never has the Amiga ball looked quite so alluring!
All I did was thoroughly clean the legs of Gary and its socket - although I am certain I have done that before - and lo and behold... I want to thank you all for what I can only describe as the best advice and guidance I have ever seen in a forum - fortunately it has all been documented for everybody else to follow in the future!

... and now I must make a confession...yesterday evening whilst checking the logic state of various pins on the Kickstart ROM, I removed the Diagrom and inserted the Kickstart ROM and powered on...there was no activity on the data lines...puzzled I checked the Kickstart ROM and noticed it had become very hot! I then realized - to my horror and dismay - that I had inserted it the wrong way! I quickly turned off the power and removed the ROM believing that I had no doubt killed it and perhaps some of the connected components...I then inserted the Diagrom again and it was at that precise point where it started working...Make of that what you will, but did my action of inadvertently inserting the ROM the wrong way "fix" something that had been "broken" before or was that merely coincidental..?