If the build is simple then what is the issue in diagnosing the specific problem (or problems) with the board?

I mean, sure you *could* de-solder and replace everything. It just seems like a lot of extra work compared to isolating and replacing the most likely point(s) of failure.

And if the problem ends up being something like a broken trace then replacing everything wouldn't fix it anyway.

Shponglefan wrote on 2023-07-02, 06:16:

If the build is simple then what is the issue in diagnosing the specific problem (or problems) with the board?

I mean, sure you *could* de-solder and replace everything. It just seems like a lot of extra work compared to isolating and replacing the most likely point(s) of failure.

And if the problem ends up being something like a broken trace then replacing everything wouldn't fix it anyway.

Unfortunately the little hardware diagnostics I know was from the 10th edition of the book "Upgrading and Repairing PCs". That was about 10 years ago and I already forgot 90% of it. I still have the book, I can technically re-read it.

So I can either learn electronics repair in my spare time, or brute force the repair. As for trace damage, that's the only possible failure point I see so far. There is no visible damage, and I believe I didn't see any ICs not receiving VCC input.

Unless anyone else finds a possible issue with the brute force method, I think I would be willing to spend a day and money on this project.

I just recently fixed one of those FIC 4386-VC-HD motherboards. The problem with that board was that the previous owner tried to fix it (replace Dallas RTC).

He had managed to install socket but it was a mess. Unfortunately I didn't take a picture how it looked at first but everything he had soldered was black. I knew right away that there was likely a burnt trace. I removed the socket and cleaned the mess.

Take a look of all those burned traces

To make things worse, he had first started to accidentally remove BIOS socket but luckily noticed after desoldering only 3 pins. Luckily because one of those three pins now had broken trace as well.

Eventually I managed to fix the board. Be careful. It is very much possible to cause more damage.

Motherboards are multilayer PCBs and sometimes there are shorts where you can't get to them, and the board is toast.

I have seen videos of techs grinding a hole in a GPU ground plane to get rid of an internal short and it was not pretty, but they fixed it.

10 percent is not repaired even with a complete set of documentation and test equipment.

If reference levels or switching times floated away in some logic element.
You can find this either by accident (brute force) or with a thorough knowledge of the scheme at the developer level.

Parts cannon approach. Same success rate as going to the dealer when your 30 year old exotic car doesnt run, proportionally same cost too 😀

>So I can either learn electronics repair in my spare time, or brute force the repair.

success rate just went down to a level of a teenager off the street working at the car dealer as a mechanic 😀 All the parts, all the diagnostic logs, no skill in using the tools.

Why not buy another board?

clownwolf wrote on 2023-07-02, 06:48:

Unfortunately the little hardware diagnostics I know was from the 10th edition of the book "Upgrading and Repairing PCs". That was about 10 years ago and I already forgot 90% of it. I still have the book, I can technically re-read it.

I would start with a more basic approach such as testing for continuity, shorts, and voltage levels.

Do you have a multimeter?

Unless anyone else finds a possible issue with the brute force method, I think I would be willing to spend a day and money on this project.

There is a risk is that you inadvertently create more issues. This could include damaging a trace on the PCB or creating a short.

I've done repairs where I've accidently caused issues (e.g. solder bridges) and being able to diagnose those and repair them is important.

If you're already willing to spend an entire day replacing everything, why just not spend that time to do some basic diagnosis? This includes testing for shorts, continuity, power levels, etc. It could end up saving you a lot of unnecessary work and the risk of damaging things further.

Ok so the success chance is down from 99.9 to a more realistic 90% due to possible internal trace damage/shorts.

As for introducing soldering damage, I am good enough to able to solder QFPs even in crowded modern motherboards, so I am hoping not to stress the board by repositioning/resoldering parts too much. I literally practiced on QFPs years ago because I couldn't believe some people were able to solder those millipede legs by hand.

For diagnosis, I already spent time doing what I can. I already tried looking for shorts, tried looking for missing VCC inputs. But I will do another hour or so to double check before I desolder everything.

rasz_pl wrote on 2023-07-02, 09:43:

Parts cannon approach. Same success rate as going to the dealer when your 30 year old exotic car doesnt run, proportionally same […]
Show full quote

Parts cannon approach. Same success rate as going to the dealer when your 30 year old exotic car doesnt run, proportionally same cost too 😀

>So I can either learn electronics repair in my spare time, or brute force the repair.

success rate just went down to a level of a teenager off the street working at the car dealer as a mechanic 😀 All the parts, all the diagnostic logs, no skill in using the tools.

Why not buy another board?

You have to be visualizing a complex model, because my motherboard is as simple as possible. Below is an approximate picture (same ICs).

As for buying another one, I already have a working SX-40. This project is specific to this particular board, because the components for this model are so cheap and numerous, and the PCB is visibly undamaged.