Etching new pads for replacement MOSFET

[WaveFunction]

The subject of today's post is an unfortunate ASUS R510L notebook. This one was in pretty bad shape, as it was the victim of a liquid spill. The residue from the spill caused a short circuit which overheated and burned up one of the input MOSFETs. It also caused a pretty big crater at the site of the short and the failure of the EC and LAN chips, but today I want to concentrate on just the input MOSFET and how I repaired it.

Since the failed MOSFET also left a pretty good crater, that left me with nowhere to attach its replacement. It was also necessary to replace two nearby capacitors which burned up along with it. There are a couple ways I could have gone about this (a piece of FR4 to mount it on, for example), but I came up with a technique that worked well and was actually quite fun to do. I decided to etch some new pads by utilizing the existing copper on the board. I went with etching because cutting the copper runs the risk of exposing lower layers (e.g. a ground plane). That's a disaster waiting to happen once you add solder to the mix.

This is what I had to work with...



Okay let's clean this up a bit...



Yikes! Looks like the MOSFET split right in two. I removed it and also removed every bit of the charred PCB underneath. That stuff is conductive and had to go...



Welp, I guess we know where the replacement MOSFET isn't going to go. Time to make some new pads. I scraped off the solder mask over the copper I wished to remove. I needed the pads to be big enough to solder the wires and capacitors to later...



The pictures get a little distorted around the edges at this point since I had to resort to holding a magnifying glass up to my phone. Anyway.. now comes the really fun part. The etching...



I used a distilled white vinegar/hydrogen peroxide/salt mixture to do the etching and created a "bowl" for it to sit in using some hot glue. Hydrogen peroxide can come in various concentrations, so I found it easiest to just start with a 50/50 mixture and then tweak it until I got a nice reaction with lots of bubbles. Enough salt should be used so that some remains undissolved in the mixture. It's also a good idea to cover any metal other than the copper you are etching with some hot glue. The mixture may react very mildly with those other metals, but not nearly as much as it does with solid copper.

After the etching was complete (I believe it took about an hour), I soaked up the etching liquid and then used rubbing alcohol to easily remove the hot glue.

The next step was to expose the copper where the new MOSFET would go...



I also exposed a little bit of copper for one of the capacitors, but that's not shown in the picture. After soldering on two new MOSFETs (I wasn't going to trust the second one just because it wasn't blackened), the two capacitors, and wires to replace the missing traces, I covered everything with an MG Chemicals black overcoat pen. It's not the best thing in the world since it wipes away too easily, but it's better than nothing...




The entire image album can be found here: http://imgur.com/a/FJUz8

A video of the etching mixture in action can be found here:


I just wanted to document this experience in the hopes that it would give others some ideas to utilize in repairs of their own.

That's all for today. Take care everyone.

[rs20]

So, did it work?

[WaveFunction]

So, did it work?

Yes, the repair worked well. There are a couple other ICs that need to be replaced also, but at least the board gets power now.

[Richard Head]

Did you consider simply cutting a small piece of copper clad for the MOSFET and glueing it over the crater? Not pretty but then I don't think any solution would be.

[WaveFunction]

Did you consider simply cutting a small piece of copper clad

Yes, I considered using a piece of FR4. But that's no fun. Also, there could have been an issue with getting the whole contraption to fit back in the case properly.

[Ian.M]

Epoxy with low density phenolic microsphere filler (e.g. West Systems 407) to make good the board substrate after char removal.  Sand flush, skim coat with epoxy to seal then apply new lands cut from self-adhesive copper foil, soldered at the edges onto the existing tracks.

[WaveFunction]

Thanks for the filler suggestion. I could imagine a rare scenario where there is no other option than to fill the damaged area so components may be placed on top of it. In this case though, I found the overall aesthetics, function, and ease of this technique to be similar, but without the need to buy those expensive supplies.

[Ian.M]

It makes a *lot* more sense if you've already got a selection of Epoxies and fillers for various mechanical repairs. 

Otherwise, doing it on the cheap, pure unscented Talcum powder can be used to thicken epoxy (make sure its real Talc, not corn starch!) and make a readily sandable filler, though as it's somewhat hygroscopic, you must seal sanded surfaces with another coat of epoxy.

If a board is badly damaged, it is generally easier to cut and fit a 'dutchman' patch with stepped or bevelled edges out of the same type and thickness of PCB laminate rather than attempting an epoxy only reconstruction, however if there is local charring round the legs of a through hole part, reconstructing a 1/4" diameter area with epoxy loaded with cellulose microfibres then re-drilling the hole and patching the track is a lot easier than cutting and fitting small dutchmen.   If you don't have microfibres, try commercial epoxy putty, but make sure it isn't metal loaded.

[rs20]

If a board is badly damaged, it is generally easier to cut and fit a 'dutchman' patch with stepped or bevelled edges out of the same type and thickness of PCB laminate rather than attempting an epoxy only reconstruction...

...assuming it's not a multilayer board? You'd want to avoid cutting a big hole through the bottom and middle layers.

[ez24]

Nice work

[Ian.M]

Yes. Multi-layer boards reduce your options considerably.  There are specialists who reconstruct them by processes like controlled depth milling to expose the damage, then building it back up layer by layer and track by track, but that's way beyond the capabilities of the average amateur or small service shop. I believe that sort of repair work is pretty much limited to aerospace or military boards that are both irreplaceable and extremely high value.

For the rest of us - if its multi-layer and the board's carbonised, its FUBARed except in the case of localised damage to the top layer above a solid copper plane that has spread the heat enough to shield the other layers.

[WaveFunction]

...pure unscented Talcum powder can be used to thicken epoxy...

That's a nice tip. Thanks. There's also a certain satisfaction in having a selection of multipurpose supplies around that have a thousand and one uses, rather than needing a thousand and one different products. While the subject of epoxies is on my mind, do you (or does anyone else) have a favorite epoxy for general use on PCBs?

[Ian.M]

Both the trouble with, and the advantage of Talc as a filler for polymerising resins is its softness.  The result is weaker than resin with just about any other filler, but that makes it much easier to sand to profile.  Talc and Polyester resin are the main components of car body fillers like Bondo.