For those with Car ABS and ECU problems;!..

[geoff]

Here we go - back to unrealistic assumptions

[geoff]

Can't be, its electronics, it lasts forever

[Dave Liquorice]

Another who reads what they want to read not what is written. Who said anything about "extensive"?

[Dave Plowman (News)]

I'd guess Dave is talking about pro broadcast gear and in that respect I'd agree with him. It's generally not engineered by bean counters. ;-)

[Dave Plowman (News)]

I mention it because it has happened to me. On an aftermarket ECU I know quite well. Offered to have a go at repairing one which had had a fairly major fault at the ignition output and had burned some tracks on the board

- including an (some) inner one(s). Circuit diagram etc readily available. But not the PC board design. It proved cheaper to rebuild it with a new PCB, junking all low cost components.

[geoff]

In message , "Dave Plowman (News)" writes

Somewhat left field of real world then

[<me9]

I've still got a radio upstairs dating from the early 30s. It's a year or two since I last used it but it was still working ok, despite having the original electrolytics in the smoothing. Some others have been changed, and modifications were done in the late 60s. I now wish I hadn't tried to improve its performance then.

[<me9]

Depends on the spec. Generally under-running helps longevity, but not always with electrolytics. Some (most) specifications are for a limited life.

[Tim Watts]

On Mon, 14 Jun 2010 14:19:32 +0100, tony sayer wibbled:

And germanium wasn't immune either. My old dad had a germanium "receiver" (tuner + amp) and it was forever, after about 10 years of service) frying the power output bipolars (I guess heat related doping migration - failure was usually gradual). They were a bugger to get too - but just about possible. The circuit diagram did have DC test voltages at various points which made fault finding fairly simple with nothing more than a multimeter. It got serviced 2 or 3 times by us and was still working (sort of) when about 35 years old. Only a 25W stereo amp too - into 15W speakers and turning it up to half max would blow your eardrums off. The 1/10 button remained depressed for most of the time.

[The Natural Philosopher]

The failure modes in semiconductors are - apart from gross overload - either temperature cycling that fatigues joints ( simple failure) or cracks seals leading to poisoning of the silicon, or high temperature operation that leads to doping migration, and eventual out of spec drift

- which can lead to either failure to operate at ends of temp range, or gross failure if the component is stressed.

Mil spec semiconductors have better seals and pay attention to differential thermal coefficients in the packages.

However its rare to see medium term failures like these: either it goes quickly, because its really over stressed, or it lasts many many years.

Some forms of flash and EPROM may fail sooner due to other issues: they depend on charge retention.

That can leak away, leading to bit corruption.

Other effects that can lead to failure that looks like time related failure of the chips, can be traced to time related failures in other components, leading to increased stress on the chips. E.g. drying electrolytics that lead to instability and/or increased current flow in semiconductors.

Or simple one time overloads, due to voltage spikes etc. These can damage without destroying, but lead to shortened life.

[Frank Erskine]

Is that the KW2000a?

:-)

[AlanD]

I had a similar issue with the ABS on a 1995 850 T5 a few years back. I bought the security screwdriver to get into the ECU (about £3), took it apart, re-soldered everything and re-fitted. ABS cured, no further issues.

I think over time that the joints to the PCB-mounted solenoids that operate the hydraulic valves may crack and re-soldering sorts them out.

[Man at B&Q]

ve

Not until you have designed as much electronics as I have.

MBQ

[Man at B&Q]

We were talking car ECUs.

MBQ

[Man at B&Q]

In the world of dodgy boiler PCBs, maybe.

MBQ

[<me9]

That was brand spanking new in the '60s Now well past it's best. What I was referring to was a National NC100 circa 1932.

[Andy Champ]

Bumped into a chip once (NEC V40 IIRC) where the extended range one was done on a bigger technology - 8 micron, not 5, or something like that. Can't check, the book is at work.

Andy

[Man at B&Q]

If it was mil spec that may have had more to do with a mil qualified production line not being available for the smaller geometry, or it may be realated to radiation hardening.

MBQ

[geoff]

Were we f*ck

"Let me show you what it followed on from

Absolutely.

In our electronics lab where we maintain *very* expensive electronic and electro/hydraulic equipment, whacking things with a rubber hammer is a standard fault-finding technique.

The vast majority of faults are not fancy electronic failures.

90% are mechanical, connector, chaffed wire, etc. Of the remaining 10% electronic problems, 90% of those are PSU-related. Only a very small percentage is actual gubbins-failure."

Nice attempt at a swerve there

[geoff]

I've designed plenty from spacework to military in my time

But keep up trying to deflect ...