My best fix was soldering a test wire to a BGA ball, that was not connected to a trace. I used a 30 gauge wire, but had to shave it flat with an Exacto knife, to fit under the package housing. With the soldering iron on max temp, I simultaneously heated and pushed the wire in the general direction of the ball in question. After passing a gentle pull test, connection was confirmed and I could measure a signal to prove that the software was working or not. My luck was that the desired IO was on the outside row.

I think the best thing that I did that literally saved the day was when my family and I had returned home from visiting family to finish our Christmas break and found our house very cold.  It was late in the evening on Christmas day, everyone was tired and now exceptionally cold.  Of course with it being the Holiday no one was available to assist and we were looking at having to try and find a hotel room and hope that we wouldn't reach the point of pipes bursting before we could get a resolution for the furnace.  Having done nothing with furnaces before this was all a scary Dragon looking thing with gas and multiple burners/flames when it is working.  Now it was just a cold shell with a small blinking light trying to desperately communicate it's problems to me.  Counting the flashes and finally finding a source online that covered my ancient furnace took awhile, but showed that the flame sensor was not finding flame and kept shutting the gas off.  Even though I could see that there was flame and gas the sensor said I was wrong.  Luckily the source also mentioned these furnaces would build up carbon on the sensor quite often and prior to replacing items and being charged extensive fees for a repairman (I shudder to imagine the Holiday/Emergency add on if anyone would even come out), you should take a brillo pad and after being sure everything was powered down and safe, remove the sensor and clean it.

Sure enough that was the fix and after powering everything back up the furnace lit and stayed lit as the clean sensor finally allowed the furnace to start heating the home again!

My electronics fix was simple but gave me access to wealth of musical resources that were hard to come by.

After purchasing a RCA connector to banana jack patch cord, I routed audio from my Satellite dish AUX audio output to my computer sound card. Using Audacity an open-source digital audio editor and recording application software I recorded continuous music channels and then cut individual tunes from the recording to mp3 files.

With the music content available today, the effort invested to garner some tunes seems extreme. When you live above the 64 latitude you soon discover that what happens in the south stays in the south. I was making illegal copies of music long before the record companies were taking individual to court.

It was a busy day and means a lot to the company, so we saved the day and our jobs.

We had a small electronic sensor design that I was called in to troubleshoot at work. This unit had just returned from contract manufacturing and had a short on the board. Without a thermal imager and with some very basic test equipment, I decided to give up on using electronic test equipment and just eyeball the board. After a few minutes, I spotted the LIS2DH "square" SMD accelerometer package was oriented the wrong way around, with the Pin 1 mark facing completely opposite to where it was printed on the board. We didn't have the time to send it all the way back to China for a touch-up, so I decided to use my untrained hands to solder it back on using a tin-the-pads, add flux, blast with a heat-gun approach. Fortunately, this worked and bought the accelerometer online and the board was now free of shorts.

But then a second problem arose - one of our digital ADC chips was spewing out garbage data at higher rates and the waveforms on the digital side looked crook. I suspected an issue with power - a colleague confirmed it was a DGND pad which hadn't been connected since someone rotated it on the schematic and that somehow "disconnected" it. Great to know the problem, but now I'm the one that has to fix it because nobody else would attempt soldering a 30AWG wire to the "stump" of land at 0.5mm pitch.

Not only was it successful, I repeated it a few times to bring up several more units. Not a pretty fix, but having not been trained to do SMD work, I'm glad I could still get it done and help the department deliver the devices "on time" for a media appearance demonstration.

- Gough

Many years ago when I was a graduate student at the University of Hawaii I worked on a project that involved transmitting a seismograph signal across campus from the Geophysics building to our lab in the Engineering building using a leased twisted pair telephone line.  I took the system offline for about a week while I was updating the transceiver circuit design and when I reconnected everything we no longer had any signal received.  Because we could not measure any of the termination resistances we knew there was a break in the line somewhere (the Engineering building was new and still under construction).  So, I used a splitter with an oscilloscope and a pulse generator and with my first practical experience using TDR I was able to determine that the disconnect was several hundred meters away.  With the help of Hawaiian Tel we determined that the line had been disconnected at a distribution panel.  As Murphy would have it, a technician had scanned the panel for active lines during the time I was offline and because he saw no termination impedances he had disconnected the line!  We got them to agree not to do that anymore.

I've since used TDR many times in my career with much faster risetimes and precision using the capability that is built into many of the higher end scopes and network analyzers.  Of course, in the case of the telephone line, I didn't need much resolution .

Ralph

Apologies if I have written about this before.....

Back in the mid 1970's, my college department was loaned a mini-computer, whose make escapes me (similar, I think, to DEC PDP8 or maybe a PDP11 or 12). This machine included an ADC, and ran an extended version of BASIC. There were 2 commands to control the ADC - a simple one, which captured a single, instantaneous value and a more complex one intended to capture a specified number of samples at a specified sampling interval.

I banged my head over the fact that a trivial program using the complex command would NOT work, whereas the simple one, called inside a loop worked fine. After several hours I phoned the company and after a day or so the UK support arm had run out of ideas and told me that they were flying an expert over from Palo Alto to investigate.

The expert duly arrived, looked at my programs, ran them, then ran his own and finally agreed with me that there was indeed a problem. He opened up the back of the cabinet and after a few minutes of inspection announced

     "The clock isn't jumpered to the ADC".

He opened up his briefcase, pulled out a jumper cable and fitted it. We then ran the program and lo and behold, it worked! I expressed gratitude and also great interest in what he had done. He was happy to show me exactly what he had jumpered and gave me some spare jumper cables, "just in case".

Some months later, I got a call from the sales rep. at the computer company, to say that a prestigious government establishment ( I'm not allowed to say any more) had the same problem on their brand new computer, which was exactly the same model that I had worked on.

Did I know what their US expert had done to fix my problem?.......Fortunately, in those days I had an excellent memory!

I went down to the government establishment, all expenses paid: first class rail, collected by chauffeur driven limo and put up in a 5 star hotel.

The next day I was taken to the computer and shown the problem. I made my fix - ie attaching the jumper -  take at least 5 minutes...they were suitably impressed and very happy with the result. I was chauffered back to the railway station and went home.

My department continued to receive preferential treatment from the computer company until the sales rep moved on.

Neil

I think my best fix, which still astonishes me that I did it, was while still being a fairly new newbie at the University where I worked, we were always being exhorted to get out into industry and get support (money, placements, grants and so on). I was at a company trying to get a grant with them from the government when we ended up discussing a problem they had. They made some sort of equipment (I cannot remember what it was now) that needed to measure something, it might have been temperatur, and the system had started to produce apparently randomly varying measurements, only one or two degrees but it was enough to worry them. I gaily volunteered to go along and see if I could fix it - and fixing things isn't my strong point!. I borrowed the only DSO in the University, worth £4000 (the equivalent now would be about £20,000), with no permission, no insurance, no special case. After a bit of questioning it turned out the fault had appeared when they had changed from a transformer based power supply to a switched mode power supply. Eventually I was able to capture a signal from the ground line on the DSO showing a 20 mV ground bounce spike caused by the new power supply. Not being an analogue electronics person I was unable to suggest a suitable filter circuit, which would obviously have cost a fair bit to implement: more components and a new PCB. Instead, taking a much simpler approach, as the system was controlled by a PIC microcontroller I suggested that they just averaged two or three successive measurements instead. Solved the problem and didn't cost any money. I didn't even charge them for my time - it just wasn't worth the effort. Sadly, I didn't get the company to apply for the government grant with me either.

Dubbie

My best electronics repair was a CNG boiler.

Long story short: just before the freezing winter, my friend had a control board of a CNG boiler failure. It heated water and all the rooms. He was offered a €400 new board (not including the work) or a new boiler. He mentioned it to me, I suspected bad relay contacts, so we ordered a new relay (already obsolete, so someone in Australia had a stock of these). The repair went OK (€20 relay, not including the work) and the friend thinks of me every winter since then

I also wrote a blog post about it: Repair of a CNG boiler control board

I recently was asked to authorize procurement of a new sound pressure level meter. When I inquired what happened to the old one, I was informed that the display wasn't working. I went over to see the meter and discovered the LCD was missing a bunch of segments. I picked it up and handed it back to the technician - fully functional, explaining it just takes a magic touch.

Of course I went on to explain that the elastomeric connectors connecting some LCDs don't have uniform pressure and can get a little intermittent, but if you push in the right spot they will reconnect. Which is what I did unobtrusively when I picked it up.

It may eventually need a more permanent fix, but this trick allowed the instrument to be used immediately without procurement or repair delays. I have seen this simple fix last for over a year before it needs to be re-seated.

Previously I did a quick blog on element14 of a disk drive recovery that others might want to try (at their own risk):

https://youtu.be/F8EDkqBL4zE