Log in

No account? Create an account
Bigger hammer


We had some pretty extreme rain here last night - the heaviest I can remember since Sheffield flooded last year. Unfortunately some of it leaked through Fenchurch's windscreen wiper spindles and dripped onto my stereo. It worked for a few minutes this morning before the front panel crashed and stopped responding to controls or updating the display. Upon reaching my destination it didn't stop playing when I turned the ignition off, so I had to pull the fuse (it has both a permanently live high-current supply and an ignition sense input). After putting the fuse back in it wouldn't power up again and the display remained blank.

Hopefully drying it out will fix it, but the last time I killed a car stereo this way (due to a leaky Land Rover roof), when I took it apart I found that the PCB and parts of the chassis were badly corroded. This is a bit unfortunate because it's quite a decent unit with a built in MP3 CD player. The stereo in my other Minor is a useless old tape deck that I'm thinking about simply removing until I have the time and money to fit something better to it.


Took it apart this evening and it was bone dry inside, though with some slight water marks and corrosion on the legs of the biggest chip on the board (probably an ASIC with an embedded DSP in it). I'll give it a clean and try refitting it this weekend. It doesn't look anywhere near as bad as the previous water-damaged stereo - I think that one had been repeatedly soaked for months before it finally stopped working.
It's an 80 pin TQFP, gull wing.

I know what you mean; I did a summer job once in the board test department of a professional mixing desk manufacturer. It was our job to test piles of newly manufactured boards and repair any faults on them. I used to drag the tip of a sharp craft knife blade down each row of pins to feel for any loose ones.

Annoyingly the company had a policy of never scrapping any boards, no matter how bad the fault or how long they took to fix, and we didn't have any useful documentation describing how they worked, so we were often left groping in the dark. There was an older engineer there who had developed The Knack - he could often pinpoint a fault intuitively because he had repaired so many boards over the years, even though he couldn't explain how they worked or how he knew a particular component was to blame.
I often found bad vias on a particular batch of four layer boards - lots of fine patch wire needed to fix those.

They had just developed a new product that included a rack containing a stack of large boards (about 16" square) each with a dense array of high-powered BGA DSPs on it. The power consumption/heat output was frightening. They were still debugging the code on the system we built up on the test room floor. It was the first time they had used BGA and the in-house assembly facility couldn't yet deal with them, so they had to be sent outside. I was very glad they didn't ask me to test/fix any of those boards!
IIRC the DSP boards were horizontal. They plugged into a backplane in a 19" rackmount case with lots of fans in the sides, about 8 boards per case. The smaller boards (eg. input and output channels) were vertical - about 30 per case, hot pluggable I think.

There was a PC running Windows NT in the rack too. This was also their first desk that ran Windows - the previous one ran some obscure embedded RTOS with their own GUI. It tended to bluescreen rather a lot during testing...