Trying to fix Mikro's wrecked TT board

[dml]

Thanks everyone for the PSU info. Might be useful later.

I have an ATX here with the correct voltages/current and the wrong connectors. I'll probably just adapt that for now since it is free.

[Steve]

It looks to me to be deliberately cut, perhaps it was once housed in some kind of a custom rack-mount or tower. If it was deliberate, then it is probably fine.

[dml]

Interesting

I started checking the ESR of a few caps around the board and while the first few were ok - I found one which measured a dead short. Then I found another one, and another...

Checked the ACIA sockets and VCC-GND are shorted.

Checked the area I reworked next to the power socket - the new axial ceramic caps - most read normally but.... C104 reads a dead short.

This is the cap Mikro had removed before I got the board.

Hmmmmmmmmmmmmmm.....

...on closer checking, not a dead short but only 28 ohms across that cap, or 10mV drop. It's across VCC/GND in parallel with a 470uF. Still seems off as the cap is not charging with the probe. Tempted to remove both and check those pads again.

Lifted one side of the 470uF and no change. Looks like it had been lifted the same way already.

Maybe Mikro has something to add since it feels like I'm repeating an old story here

Thinking about looking out my most accurate DMM and narrowing down whatever is providing that 28 ohm bridge.

[Badwolf]

Thinking about looking out my most accurate DMM and narrowing down whatever is providing that 28 ohm bridge.

It's not that corner, is it?

Whilst I kind of agree with Steve that it looks a bit neat to be a break (especially if it's a 4-layer board!), it's likely still exposing a ground and 5V plane very close to each other.

If it were deliberate, you'd think someone would have had to have taken the time to polish that edge so that there were no little dags that could be shorting, but who knows what's happened to it since?

These shorts (or near shorts) sound like they come and go -- potentially some tiny little bit of copper contacting and disconnecting on that break as you move the board around?

BW

[dml]

It's not that corner, is it?

Beginning to wonder.

I put a Keithley DMM on the case but 28 ohms is just big enough to make narrowing it down by distance pretty much a lost cause.

Whilst I kind of agree with Steve that it looks a bit neat to be a break (especially if it's a 4-layer board!), it's likely still exposing a ground and 5V plane very close to each other.

It could be. It is very suspicious though re: replacing C104 brings a short back - which Mikro apparently removed while eliminating a short.

It might be there is some crosstalk inside the the thru-hole. I put a torch under it and there's a star configuration on both pads on an inner layer - one will be GND, one VCC +5V.

If it were deliberate, you'd think someone would have had to have taken the time to polish that edge so that there were no little dags that could be shorting, but who knows what's happened to it since?

Yep, no idea what happened with that corner. I might just have to grind/polish it a bit to see whats going on.

These shorts (or near shorts) sound like they come and go -- potentially some tiny little bit of copper contacting and disconnecting on that break as you move the board around?

I may start by removing C104 again but I have the feeling the short will remain and now I'm wishing I kept a record of those 2 pads. I'm sure I checked them too.

[mikro]

@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)

viewtopic.php?p=66980#p66980. Jookie made his own from wood, I purchased one from BEST for $30.

[mikro]

Maybe Mikro has something to add since it feels like I'm repeating an old story here

I'd love to but I was really just shooting in the dark. So I started with checking those caps near the PSU connector (and yes, the one died while desoldering), they were all fine, then I did "something" which I can't remember anymore (perhaps cleaned up some part of the PCB?) and voila, the short circuit was gone, the computer started at least to the level so i could detect clocks on the CPU etc.

Seeing how terribly inconvenient debugging TT's ICs (let alone the PGA CPU...) is, I'd decided to let it go.... to you.

[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)

viewtopic.php?p=66980#p66980. Jookie made his own from wood, I purchased one from BEST for $30.

Interesting - sorry, don't want to go off topic, but the PSU Jookie refers to (MeanWell RQ-65B) is 5, 15, -5, -15. According to this spec link. I wonder if 15v is within spec for the 12V/-12V rail.

PS @dml notice here: viewtopic.php?p=67067#p67067 Apparently the TT requires an FPU to boot ... worth taking note I suppose, when you get to that stage.

[dml]

Seeing how terribly inconvenient debugging TT's ICs (let alone the PGA CPU...) is, I'd decided to let it go.... to you.

So I don't need to feel so bad now at being confused also!

I removed C104 and no change, as expected. Still the mysterious 28 ohms across those pads.

I did the steel wool thing at the cut corner and confirmed there are both GND and VCC planes exposed at that edge so I'm going to follow BW's direction next I think.

Failing that, I might try zeroing the DMM at 28 ohms to see if I can get a vague idea of which quadrant of the board is shorted, or move to the Datron with 7 digits

[dml]

PS @dml notice here: viewtopic.php?p=67067#p67067 Apparently the TT requires an FPU to boot ... worth taking note I suppose, when you get to that stage.

Yep have taken note of that one.

I have been unable to narrow the short with the DMM - I think mainly because there is decoupling everywhere so it always settles to the same 27.48 ohms and pretty much noise after that.

I could try applying 5V, raise the current and look for a temperature rise somewhere unexpected...

...or maybe spend more time on that cut corner but I did try wire wool on it and the resistance has not moved at all.