So, I've been intrigued for awhile about the issues that there are with window switchpacks - randomly failing, or switches becoming intermittent.

Sometimes, yes a blast of contact cleaner works, or it works for awhile and then it starts playing up again.

We've got a P38 in the workshop that we are fixing up to sell, rather than break it for parts - and the other day I noticed that the window switches were playing up. They had all been working fine, but now they were intermittent at best.

So, being that I like to pull things apart, and because I wanted to see if it was fixable rather than chuck it out and buy a replacement - which will be second hand anyway, I got it on the work bench this afternoon and stripped it right back. I did a test run on one of them to see if it was a case of being able to repair, before breaking out the camera and taking photos... Pix or it didn't happen, right?

What I found (before I get into the pictures and the steps I took) was that the little metal disc that is actually the 'switch' looks to get tarnished over time, and on one of them I saw there was a tiny spot of corrosion build up too - which was causing the connection to be intermittent.

So I'll start from the point where you've got the window switchpack out of the vehicle, on a bench - and the switch module removed from the trim and the board unclipped from this. On my one there were a bunch of Torx head screws holding the switch module to the trim. I think they were either a T10 or T15. I just grabbed the screwdriver off the shelf and didn't check what it was!

So, you've got this board in front of you, with a whole load of dust and crud on it, and a bunch of switches that don't work... The first step is to take the switch caps off. The one I was working on I could unclip them with my fingers but pulling up one side, and then following with the other. Technically you don't NEED to remove them, but I found it useful at the end to make sure everything was still sitting in place after the clear light pipe octopus was put back in place.

(the Rear Inhibit switch was my test one so was put back together and tested by this point!)

So next, turn the board upside down and you get a view of the bottom of the switches like this:

Take a sharp knife and then cut off the black plastic tips of the switch housings. Be careful removing them as you want to put them to one side for later... We'll use them to help hold the switches back on after the fixing is done.

You should end up with something looking a bit like this:

You should then be able to remove the switch from the board. Sometimes they need a bit of persuasion from a thin spudger (though I was able to pull most of mine off with a bit of a wiggle)

And you end up with this:

And the board looking like this:

Then take your sharp knife again (I use a scalpel for these smaller bits) and cut the tape around the switch pads so they can be removed:

Then remove the switch pads and you'll have a board that looks like this:

And then you get the switch pads separate. as you can see, there is some tarnishing on one side of these.

I then used a small screwdriver just to scratch the surface and scrub the tarnishing off of it.

After removing the tarnishing from the underside of the button, I decided to apply a VERY thin layer of solder on them - but don't get too excited with it - if you turn it over and can't press it down for it to 'click' then you've put too much on!

After doing this, then I lined them back up one at a time, and put some clear tape on them to keep them back in place:

With that bit back on it's possible to press the buttons and test the connections, to make sure they are solid and not intermittent anymore. I have done a pin-out of the pins on the main chip on the board and will put that at the bottom of this post, or in the next one once I've finished doing a draw up of the chip. To test the buttons, use a multimeter on continuity, and put the -ve probe on the ground point for the switch pack (the bolt on the voltage regulator by the connector is a good place) and then the +ve probe onto the pin of the main chip that you are testing. You will get a reading of 000 or 001 ohms on the side of the switch you are testing (or should do!) and a reading of about 900ohms when the opposite side of the switch is pressed. If you get a reading without pressing the switch at all, then the button is shorting out, so will be active all the time!

Once you have tested it an happy to begin reassembly, then use your sharp scalpel/knife etc to prick holes back in the clear tape so the legs of the switch housing and light pipe can poke back through. Then push the switch back in and then use a soldering iron to melt the plastic of the legs back to hold it in place. This is where I used the cut-offs from disassembly and melted them back in aswell and used a pair of tweezers to flatten it out to make sure that it's not going to come apart again.

And you should end up with something like this:

And a whole board of them:

Once you've got the whole thing repaired and the switches are all tested, then you can put the switch caps back on, the light pipe 'octopus' back on, you can then put it back in the casing and reassemble it all... Put it back in the vehicle and test it!

Image below is of the chip that's on the switch pack - looking at it, the black dot is the semi-circle towards the edge of the board.

Hope others find this useful, and can keep a few switchpacks in use for a bit longer - since they seem to be 'unfixable'

I am also looking into the possibility of replacing the little metal 'pad' switches (if they've worn out completely for example) with the small 'tactile' surface mount switches (like the buttons in the remote fob) - but since these seem to be repairable, that's a bit further down the list again!

This won't fix the problems where the switch pack completely fails (like some people find after putting a new battery in)... I am not sure what causes this, but if someone has a spare one about somewhere which has failed in this manner, then I'm happy to take it off your hands and take a look into what has actually failed on it and see if it is repairable too...

Marty