Lower resource linux
in [Linux Help]
|
From: VWWall on 26 Jan 2010 18:04 Moe Trin wrote: > On Mon, 25 Jan 2010, in the Usenet newsgroup alt.os.linux, in article > <nPSdnWeJaZvPhcPWnZ2dnUVZ_gydnZ2d(a)earthlink.com>, VWWall wrote: >> I used to tell my reliability engineers that there was no such thing >> as an "electronic" failure. When analyzed, all failures are >> mechanical; they just show up as symptoms in the "electronics". > > Might be a bit of a stretch, because you can get some very interesting > failures with voltage or thermal (ambient) changes causing differences > in speed, which gets into a race condition. Electronic noise can also > give you heartburn. But I do recall one that drove me nuts for a > while - it was finally isolated as sound (high speed mechanical > vibration) causing frequency changes - minute, but enough to change > the heating produced in a transistor. I've seen microphonics before, > by this was ridiculous. > Isn't this a case of *mechanical* vibration causing the failure? Even thermal failures usually result in mechanical changes that eventually cause the failure. Latent ESD might be arguable, but even that is probably caused by mechanical defects created in the semiconductor's crystal structure. I remember one where a "light gun", (hand held sensor used to identify planes on a PPI radar screen), caused various problems when the operator squeezed the "trigger" to activate the sensor. It turned out the microswitch was producing a mechanical pulse which was turned into electronic noise by an adjacent ceramic capacitor that was acting as a transducer. The difference with many software failures is that they involve a human interface which is not always what the programmer imagined or tested for. For every fool proof design, hardware or software, evolution seems to have produced a better fool to disprove it. -- Virg Wall
From: Mike Jones on 26 Jan 2010 19:12 Responding to VWWall: [...] > For every fool proof design, hardware or software, evolution seems to > have produced a better fool to disprove it. Good quotable there! Somebody give that guy a prize. :) -- *=( http://www.thedailymash.co.uk/ *=( For all your UK news needs.
From: Moe Trin on 26 Jan 2010 22:09
On Tue, 26 Jan 2010, in the Usenet newsgroup alt.os.linux, in article <IZedndZQcot36cLWnZ2dnUVZ_t6dnZ2d(a)earthlink.com>, VWWall wrote: >Moe Trin wrote: >> Might be a bit of a stretch, because you can get some very interesting >> failures with voltage or thermal (ambient) changes causing differences >> in speed, which gets into a race condition. Electronic noise can also >> give you heartburn. But I do recall one that drove me nuts for a >> while - it was finally isolated as sound (high speed mechanical >> vibration) causing frequency changes - minute, but enough to change >> the heating produced in a transistor. I've seen microphonics before, >> by this was ridiculous. > >Isn't this a case of *mechanical* vibration causing the failure? Bit of a stretch, but yes. What wasn't understood was the mechanism that connected the two. There was also a thermal lag that masked the connection. >Even thermal failures usually result in mechanical changes that >eventually cause the failure. Thermal or voltage changes causing propagation velocity changes might be visible down at the atomic level, but wouldn't be mechanically measurable. They usually would be quite apparent looking at the shape of the electrical waveform. >For every fool proof design, hardware or software, evolution seems >to have produced a better fool to disprove it. I suspect that law was proven before humans existed. It's just that some people are slow to get the word. Old guy |