Phil wrote [below...seems a bit angry again...]

Perhaps a simple illustration will help you, Phil.

Take an example 100R 1% resistor. Let's say the probability of it
being only 99R is P.

Take another from the same batch. The probability of it also being
99R is also P

The chance of them *both* being 99R is Psquared. Since P is always
less than one, Psquared is always smaller than P.

There is no other combination of values that adds up to 198R, so the
chance of the two in series being 1% is smaller than either on its
own.

OTOH, there are many possible combinations of values, the one larger
and the other smaller than the nominal value, or vice versa, that
could add up to the nominal value, making it more likely that two in
series would be near the nominal value.

In other words, even if the probability distribution of values in
the batch is flat with vertical limits at 1%, as you suggest, the
distribution of two in series will be an inverted U shape. With many
resistors in series, it will approach a normal distribution I think.

I have been trying to find out whether resistors are sampled, or if
each one is measured. The above works for either but, if they are
sampled then the compound resistor will have a better tolerance than
the single value.

So, does "1%" mean every one, or is there a confidence limit,
perhaps 99.9%within 1%?

Anyone know whether the nominal value is the mean, mode, median, or
none of those? Anyone know what the distribution is around the
nominal value?

I'm sure it's not actually flat. They aren't made by chance.

cheers, Ian


"Phil Allison" wrote in message
...

"Ian Iveson"

** A colossal, context REMOVING, ****ing pommy IDIOT !!!!


** WRONG !!!

Even if *both* resistors were 1% high, the series value is also
1% high.

Using resistors in series or parallel IMPROVES the % accuracy
of the final value.


Phil, those two statements appear contradictory, because your use
of "%" changes from the first to the second.


** Only seems that way to someone with autism.



Or you are wrong...


** Go get rooted - you autistic prick.


What improves is the *chance* that a resulting value will be
within tolerance.


** WRONG.

It will be within tolerance for all non faulty resistors.


Since quoted tolerance assumes a particular distribution of
values around the quoted value,


** No it does not.

The % tolerance of a resistor is the MAX amount that any sample
will differ from the nominal value INCLUDING the effects of a
normal service life.





........ Phil