Trying to fix Mikro's wrecked TT board

[dml]

Had a look at the schematic again and it appears each resistor pack isn't a normal resistor network with a common pin.

It looks more like a set of 220R/330R dividers across +5V/GND. If that's the case, there are 6 dividers x 3 packs x (220R+330R) all in parallel across the 5V rail - which should mean..... about 30 ohms :)

I'll get my coat!

[dml]

The trouble with thermal cameras is interpreting the results. I'm presuming that the 29C is indicating the max temperature? I would have thought that to be normal for the resistors. Does your power supply give you a current draw?

Yeah I wasn't really trying to get absolute temperatures, I was more interested in seeing where the current was routing to, at such a low voltage (1V - 1.5V I used in those tests earlier) and how things change as I raised the voltage.

The resistor networks I think are the answer to the low resistance, now that I see how they are actually configured.

[dml]

The resistors have been installed the correct way around haven't they? Just a thought.

That is also worth checking, although I think these divider thingies are symmetrical at pins 1 & 8 - so only the middle pin voltages would end up incorrect and those go to the DB25 socket. Still... it would be sneaky-bad if they really are flipped.

[dml]

Some uneven heating of some other ICs as the voltage is brought up to 3v, 4v, 5V.

Screenshot 2025-04-13 at 20.10.40.png

Screenshot 2025-04-13 at 20.10.15.png

Nothing blew up or particularly warm, except for the resistors. I think I'll start putting pieces back soon.


I forgot how terrible the SeekThermal app is and the fact it doesn't do anything about different thermal drift on every pixel.

I wrote my own MacOS desktop app for this camera in 2018 to fix those problems when I was looking for a water leak. Just can't find the project now... just lazy to dig.

[Badwolf]

It looks more like a set of 220R/330R dividers across +5V/GND. If that's the case, there are 6 dividers x 3 packs x (220R+330R) all in parallel across the 5V rail - which should mean..... about 30 ohms :)

*headslap*

:lol:

That's a hell of a base load, though, isn't it?!

BW

[dml]

Quick update on this...

I'm planning to repopulate the board next and test with the 5V rail only.

However I want to use known-good ICs where possible, minimising the number of unknowns to the main custom chips. There are a lot of ICs on the board so putting potentially bad jellybean(ish) chips back in will just complicate debugging.

I don't *think* there are any/many bad chips TBH but it is easily possible - I don't think more than a white screen has been seen from this board in the tests which were done (?).

I also want to avoid plugging chips in and out unnecessarily.

The main thing that's bugging me at the moment is the 16 DRAMs - I'm waiting for a DIP DRAM tester to arrive and while the ICs are probably all ok, I prefer to give them a cursory check before putting them back in the board. I can test them more extensively when the machine is working.

I'm also unsure if the machine can boot without the SCSI controller present (something I don't have - the same IC in the Falcon is different, I think the CMOS variant, unsure of compatibility). If TOS polls the device on startup it could be an issue. This is likely bypassed by a diagnostic ROM but again, complicates interpretation of white-screen result meantime :)

[stephen_usher]

TOS 3.06 doesn't care about the SCSI chip not being there from what I remember. It certainly doesn't care if the SCC serial chip is there or not. Even if it does care about the SCSI chip, it will be later on in the boot process after it's put the logo on the screen and done the RAM test.

As for the SCSI chip, it's the same one as used in early Macintoshes and Sun workstations so should be easy to get as a second hand part.

[frank.lukas]

You can not swap the Atari TT 5380 SCSI-Controller der Fa. Logic (UA02) for a Atari Falcon 53C80 ...!

[dml]

Thanks both re: the SCSI/SCC info... very helpful to know.

Yes I was not planning on just trying my luck with a Falcon SCSI IC on the TT board, at least not without careful reading of the datasheets. It is typical for CMOS parts to be non-interchangeable with TTL etc., with burny toasty EOL results.

[rubber_jonnie]

Had a look at the schematic again and it appears each resistor pack isn't a normal resistor network with a common pin.

It looks more like a set of 220R/330R dividers across +5V/GND. If that's the case, there are 6 dividers x 3 packs x (220R+330R) all in parallel across the 5V rail - which should mean..... about 30 ohms :)

I'll get my coat!

The Amiga does the same thing with resistor packs, it has 10 pin items that break down into 5 individual resistors rather than having 9 resistors all connecting to a common pin.