Friday 12 April 2013

No, not the comfy chair

Yesterday one of the devs at work tweeted that when one of his teammates stood up, his hardware crashed.

This reminded me of a test report I'd read and sure enough a bit of Googling brought me this story from Jonathan Kohl about tapping the desk with his knee.

Today the devs investigated the problem - see video below


The chair was the culprit - but only one particular chair that generated some static electricity.

Further investigation found the problem was this cable




the voltage spike produced when getting up from the chair


the fix



So now I have to get myself a polyester tracksuit and jump around a bit when doing my testing
 - but dont worry readers, I wont be posting a video of that.


( title of the post? Thanks to Monty Python of course )

2 comments:

Jonathan Kohl said...

Cool, thanks for sharing Phil.

With pervasive computing and mobile devices, this kind of physical interaction is our new normal. It puts an enormous stress on blackbox testing, because you end up with a lot of intermittent bugs when you test without the knowledge of the hardware and other systems, and the design and usage of tech in the product you are testing.

I find it fun. :-)

Steve Poling said...

Years back I was working with a line scanner fed into a digitizing card on a PC. The software worked like a charm, most of the time. But it started going hairy. The images came out with hash all through them. I didn't understand. I was chasing the problem one afternoon and my favorite radio program was on. So, I moved my radio to the lab and noticed that every time my software had a bug that I heard hash on the radio. Oh, there was RF interference that was being propagating through the power line. I moved the outlet my computer was plugged into and the problem magically cured itself.