> Your last question about overclocking a PIC was just dumb. =A0Sorry, but =
there
> is no nice sugar coated way to say that,

No.
Experts have to be wary of reinterpreting a beginner's question to
mean what they take it to mean, rather than accepting it at face
value.

Here "what would happen if ..." is far far different from "Can I with
100% success and certainty...".

[[Some but not ll following materials is loosely modified from another post=
:]]

'There is the danger to be avoided of being so experienced that one
sees the answer to a question as being so obvious that it is too
trivial to have been asked about. It's all too easy to feel that
because the subject is discussed or quantified in a datasheet that
this counts as adequate  "explanation" as well.

The sometimes great gap between "what" and "why" for a beginner can
become inobvious to
the widely experienced.

> and you need to sit down and listen  for a change.

Agree. That's seldom bad advice for most of us :-).

> The reason it was so dumb was because at had nothing to do
> with the level of electronics or PIC or firmware knowledge or being a
> newbie. =A0The manufacturer rates something as xxx and you ask what happe=
ns
> when you subject it to 5 xxx.

Straw man. Because, ...

=A0This is not a question of MHz or clocking a
> PIC, just the attitude that you are somehow entitled to get around
> something, or the laws of physics doesn't apply to you.

... he never claimed this or even suggested it. One can read-in a
whiff of adumbration if one wishes, but it may or may not have been
his intention. It may reasonably well have meant " ... eg does it
burst into flames, die permanently, not work for now, run terribly,
actually work but ...?".

> If a bridge is rated for 4 trucks, what happens when you put 20 on it? =
=A0Can your 1 quart
> water bottle hold 5 quarts? =A0If the speed limit is 20 MPH, what happens=
 if
> you drive at 100 MPH? =A0Red line on my engine is 6000 RPM. =A0What happe=
ns when
> I run it at 30000 RPM? =A0If you think those questions are stupid, then
> consider how they're really no different than what you asked.

No. The comparisons are fair enough as a good guide to the probable
answer BUT you are trying to insist that HE sets himself mental
filters that YOU would never set for your self. In areas like this
where you are indeed Grandmaster with Oakleaves and Diamonds the
equivalence of the comparison is so clear as to be not worth asking.
But in areas where you personally have less certainty you will
certainly explore experiment and question. You may well choose to do
so privately so that nobody sees your ignorance, rather than ask
'aloud' in public, but nature offers enough "still works at 500%
overload, fancy that!" situations that you (and I) will still look for
them.

Eye logarithmic response to light is an example.
You can view things at 100,000 lux in daylight, read at 20 lux in
color, at 10 lux in good detail in partial monochrome, and find your
way around at under 0.5 lux in moonlight. About 200,000:1 or more.
That much is partially due to iris effects. BUT it keeps on going from
there.

The utterly utterly stunning ability of the dark adapted human eye to
detect a single photon !!!!!!!!!!!!!!!!! is far beyond moonlight
vision. Such a dynamic range is beyond incredible. Microchip may have
managed some lesser miracle. (They didn't).

____  from here is again lightly edited prior material which is
however wholly relevant _____________

More useful probably (IMHO as ever) would have been a comment that the
result "will not be marvellously good" followed by a discussion of how
un-marvellous and why. A discussion of the process of production
sampling to achieve high and low speed parts, how far you
may be able to push such things, why you probably don't want to,
production yields, distribution tail sizes and extents, guaranteed
specs, maximum specs, contributions to the journal of irreproducible
results and more might have ensued. It didn't. More's the shame.

Modest 'overclocking' may well work perfectly well.
If it doesn't you have no comeback.
If it does you can't be sure that it will not glitch occasionally -
free intermittent bugs at any place in any program are only for
masochists.
As you push the speed up results will get worse. Some  ICs will fault
sooner than others and some will be much much better. Even the best
will not achieve a 5X clock speed gain and the one you have was
probably made on a Friday before a long weekend,. Where several speed
classes are offered with otherwise identical specs (rarer now than in
the dim past) the fats ones are usually found by cherry picking from
the general production. Nowadays consistency is often so good that
"slow" parts may be fast parts badged differently. YMMV however.
If you do reliably achieve a 10% or even 50% gain in speed there is a
risk that next time you will be less lucky and it may be that  a
perfectly functioning overclocked part will conspire with Murphy to
ruin your day at the worst possible moment. The gains that you can
achieve this way can be bought by using a faster processor and this is
usually a far superior approach.

So, while more reading would have resulted in less appearance of
ignorance, I think (IMHO etc) that the charge of 'just dumb" falls far
short of reasonable objectivity. No?




              R

-- =

http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist