[rescue] Slightly OT: Bad Cap Saga

[Curious George]

--- On Mon, 8/18/08, Geoffrey S. Mendelson <gsm at mendelson.com> wrote:

> On Sun, Aug 17, 2008 at 03:29:30PM -0700, Curious George
> wrote:
> 
> > I *think* most of the capacitor failures that I have seen are
> > brought about by *use*.  I.e., sitting idle (unpowered) doesn't
> > harm the devices.
> 
> Obivously you are a modern design engineer with no experience in older
> electronics.

You're right.  I started my career in the mid 70's (with the i4004)
and have *no* experience with "valves", etc.  :>

> Ask on any of the older radio mailing lists about what
> to do with a radio that has sat unused for many years. Many
> of them have capacitor failure due to aging, and they were made
> before any of this happened.

What *relative* timescales are you talking about?
The poster was talking about some "spare PC's" that are
sitting around waiting to replace units that are in service.
Are you claiming that those *sitting* unpowered will exhibit the
same failure rate as those that have been *running* for the
past couple of years (note the model he cited is not a 30 year
old tube radio!  :> )

> > But, I can't be sure of that!  E.g., if the problem was caused
> > by something in the manufacturing process that *contaminates*
> > the components (e.g., a bad wash), then it's possible that
> > their actual service life *is* shortened despite being
> > in storage for that time...
> 
> No, it wasn't a bad wash, it was a missing preservative. 

***IT*** wasn't a bad wash.  The point of my post was "are there
any *other* things going on, here?"  Stop fixating on this one
incident.  50 years hence, will you still be blaming all e-cap
failurews on some botched industrial espionage attempt in
Taiwan?  When do you start wondering if, perhaps, something
*else* has come into play?

I make it a point to look at "failed devices".  I want to
understand *why* they failed.  And, if there is anything that I,
as a designer, can do to reduce the likelihood of failure in
the future (and what the tradeoffs to do so might be).

I'm not only interested in how the electronics failed but how
mechanically something may have died.  How long that item
*should* last is a separate issue to me.  Should a Cadillac
last longer than a Chevy or a Yugo?  <shrug>  Understanding
how they fail is a skillset I actively develop as it lets me
add value to the things I work on.

Likewise, I routinely evaluate the firmware and software in
products.  Especially products that I don't use.  It helps me
figure out what's good about a design vs. bad.  E.g., this
past week I played with a dozen LCD monitors.  While the
"front panel controls" on most were similar, some were
designed in a clearly inferior manner that made using them
very unintuitive (and, since I suspect very *few* people
read the user manual for their monitor, having an intuitive
interface is *essential*).

The point of this thread was to ellicit any observations people
had over what types of failure modes they might be experiencing
in various bits of kit.

> As I said in a previous reply which you should have seen by
> now, they probably replaced it with something that extended
> the life of their product, but not as long as the original one.

So, your summary contribution is "(still) bad components"
akin to making 0.245W resistors to save a few micropennies
per unit (resistor).  OK, then I just avoid Taiwanese
components and vendors and all is well!  :>