Trying to fix Mikro's wrecked TT board

[don_apple]

According to https://www.atariancomputing.com/blog/t ... on-guide-4 the purpose of the daughterboard is to enable the processor to run at 32MHz, because originally the mainboard was only designed for a 16MHz processor.

[dml]

It takes an external 32MHz clock input so my guess is the onboard one is 16MHz and the control signal timings are perhaps still transitioning when the 'middle' edge of the 32MHz clock might sample them, so make sure they all occur on a common clock edge instead?

So long as there is actually a clock coming into the PGA - as opposed to supplying one out - I should be good to go. I might probe the clock pin with the IC missing to see whats going on first though.

In a new mini drama, Q402 is open at one junction (2n3904 NPN). Not sure what that's for but it is near the ROMs, onboard DRAM, J400 and what is probably the SCSI area.

There are some other 3904s near the back ports, all of them reading strangely (almost a short but not quite). I don't think all 3 are faulty but maybe something they are connected to. It's not capacitors though, something with a fixed, very low resistance.

Need to look into these areas before doing anything else.

[edit]

Something definitely wrong - Q201/202/203 all read 0.16v drop at both junctions - but the schematic says they are connected base to TTVIDEO RGB out (pulled 360R to GND), collector to VCC, emitter to video connector via 75R. The TTVIDEO chip is missing and the rest doesn't make sense. Only 3 shorted caps could pull the emitters low.... should probably extract those transistors and see whats going on.

[edit]

...removed one of the transistors, checks out ok. Measured the pads on the board and 0.16v drop still there from base to emitter or collector. One of the vias doesn't conduct all the way through without pressing quite hard with the probes. More flux and wick and cleaning.... 0.16v drop is gone from all 3 transistors, measuring open to VCC and 360R to GND.

o_o

I'm going to pretend that didn't happen, put the transistor back, check things look normal and move along.

[dml]

It's sunny this week and that's always distracting this far north - but did manage to get the ACIA sockets in.

Screenshot 2025-04-07 at 19.52.38.png

Also found my ancient ATX supply that actually does -5v so I should be able to cobble power together soon.

Now, which chips should I explode first? -_o

[Badwolf]

One of the vias doesn't conduct all the way through without pressing quite hard with the probes.

You mean there was no connection from one side of the board to the other through the via? :shock:

Any evidence of board delamination? From impact, perhaps?

BW

[dml]

Any evidence of board delamination? From impact, perhaps?

Nothing visible. I'm not sure what actually happened to the board though. Given a whole corner is missing from the PCB it's difficult to be sure :p

The rest of it looks quite clean and only a few tiny scratches on the underside.

[Steve]

Also found my ancient ATX supply that actually does -5v so I should be able to cobble power together soon.

If you fancy using a more modern PSU, feel free to use this board: https://exxosforum.co.uk/forum/viewtopic.php?t=3551

Just FYI make sure you choose a PSU with at least 0.3A (perhaps 0.5A) on the -12v. PicoPSU's usually only have 0.1A which isn't enough to generate the -5V required for ECL.

Or alternatively you can buy a new design TT PSU here: https://www.16-32bit.eu/mste-tt-psu
You'd just need to find a way to 3D print a PSU base bracket.

[frank.lukas]

Given a whole corner is missing from the PCB it's difficult to be sure ...

It is a Multilayer ...!

Maybe you can see something if you look from above and hold a strong light source up against it from below?

[mikro]

If you fancy using a more modern PSU, feel free to use this board: https://exxosforum.co.uk/forum/viewtopic.php?t=3551

Just FYI make sure you choose a PSU with at least 0.3A (perhaps 0.5A) on the -12v. PicoPSU's usually only have 0.1A which isn't enough to generate the -5V required for ECL.

Or alternatively you can buy a new design TT PSU here: https://www.16-32bit.eu/mste-tt-psu
You'd just need to find a way to 3D print a PSU base bracket.

Considering that the former requires custom PCB and the latter is expensive and out of stock, I'd recommend Jookie's way, i.e. using the Meanwell PSU. It has all the amperage and voltage needed (and of course its own case), I just bought a new molex connector and some wires.

[Steve]

@mikro Could you please link to the Meanwell model Jookie used? I haven't seen this. By the way, do you know if anyone has designed an STL 3D print for the PSU base? (So you can mount a custom PSU)

[dml]

It is a Multilayer ...!
Maybe you can see something if you look from above and hold a strong light source up against it from below?

We think there are no traces here:

Screenshot 2025-04-08 at 09.47.31.png

Screenshot 2025-04-08 at 09.47.14.png

...hoping there are no traces here :-)


I did try a powerful LED torch under that corner but the groundplane stops any light getting through.

The only way I could check for traces I think would be to polish it flat and inspect with a microscope - or maybe use the steel wool trick with the multimeter, rubbing the wool it on the open edge while poking another lump of steel wool around the underside of the board (and groundplane will make that very error-prone).

I think for now I'm going to assume there are no valuable traces here (and regret it later).