15 Apr
2011
15 Apr
'11
6:18 p.m.
On 2011 Apr 15, at 5:39 PM, Dave McGuire wrote:
On 4/15/11 6:55 PM, Brent Hilpert wrote:
The metrology world typically uses calculable capacitors for C. For
voltage, Josephson junctions are the top dog, driven by a primary
frequency standard. This is a primary standard because the only other
parameters are e and Planck's constant. I believe quantum Hall effect
devices are still used for R standards but I haven't kept up with
that. SET (single-electron tunneling) devices are en vogue as current
standards but I don't know much about those either.
In electronic metrology, "measurements" are carried out by
comparison to these standards or derivatives thereof. Voltage is
measured potentiometrically via either a solid-state null detector or
a SQUID (superconducting quantum interference device). For
capacitance, as in Tony's situation, I'd expect a passive AC bridge
would be how it's done, comparing against (absolute best case) a
calculable capacitor. That gets you down to first principles, but the
"measurement" is still a comparison. As you stated above, measuring
the RC time constant with a precision timer and a precision voltage
source (both driven by primary standards, cesium fountain and
Josephson junction array) is still comparing it, by my interpretation,
mathematically against numbers calculated from those primary
standards. I'm not sure Tony's instructor actually had a point.
We all agree there are always comparisons to some reference or other,
but the point I was attempting to make of the distinction between
approaches remains. The calculable capacitor situation relies on
another definition of C to calculably take the measurement back to that
of the more fundamental measurement of distance, as opposed to simply
calibrating it as another C (more turtles).
I think
the wordst was when I was aksed to descrtivbe a method of
measuring something (the capacitance of a capacitor). I did so (an AC
bridge) and was told 'That is a compariston and not a measurement',
I am
still waiting for a definition of 'measurement' (or an esample of
one)
which
does not involve a comaprison to a standard.
I expect that what your instructor was getting at was that the use of
an
AC bridge typically involves comparison with another C, which also
needs
to be measured. It becomes a 'turtles all the way down' problem. This
would be in contrast to a solution which examines Cx in terms of the
definition of capacitance (time/charge) and breaks it down to more
fundamental/axiomatic measurement units (i.e.,time,mass), even though
this does involve comparison to some other standard.
"That is a comparison and not a measurement" may not be a full
expression of the distinction, but I think your instructor had a
point.
I'm not up to scratch on my measurement theory, what are the
fundamental
standards these days?, there's oscillation of the cesium atom for T,
there was the Meter bar in Paris for D, but hasn't that been
redefined?,
etc. Or, how many fundamental standards do we need to derive
everything
else? 3?.