[volt-nuts] LCR meter calibration certificate

[Dr. David Kirkby (Kirkby Microwave Ltd)]

I just got my LCR meter back from Keysight today.

For anyone interested in looking at the cal certificate, which includes the
uncertainties, a copy may be found here,

http://www.kirkbymicrowave.co.uk/cal_certificates/Keysight-standard-calibration-with-uncertainties-for-4284A-precison-LCR-meter-18-08-2015.pdf

It's not entirely sure if it was adjusted or not, as the instrument was
upgraded, although I'm not convinced that the upgrade really would have
required any adjustments. I suspect only a verification it was working, as
it was purely a software upgrade. Though perhaps it allows other hardware
in the meter to be used.

It lists

As received condition: "Not applicable, as this calibration certificate
applies to the initial calibration of a new, refurbished or upgraded
equipment."
Action taken: "The equipment was upgraded."

So nothing specifically mentioned about it being adjusted, but nothing to
indicate if it was in spec when received either.

There's a few things rather puzzling about this.

1) Why did Keysight use two 3458A multimeters to calibrate it?

2) Why did Keysight use two 53132A frequency counters to calibrate it?

3) It's not clear to me what the actual values of the devices they used to
calibrate it with are. They used a range of resistors (100 Ohms and 1000
Ohms), and capacitors (10 pF to 1 uF in decade steps), but no inductors.
The nominal value is given (e.g. 10 pF, 100 pF), and the uncertainty. But I
assume those capacitors are not perfect, and so they would be calibrated
periodically. Would they not use the actual value to determine if an
instrument meets is spec, rather than the nominal value?

One line is this:

min=9.9605 pF measured=9.9941 pF max=10.0409 pF uncertainty = 0.0014 pF

If you average the minimum and maximum values, you don't get 10.0000 pF,
but 10.0007 pF. Does that mean the actual value of the capacitor they used
is 10.0007 pF, within their uncertainty of 0.0014 pF?

4) Strangest of all, and I'm sure some time-nuts might think is odd, is the
uncertainty on the measurement of the "Test Frequency Accuracy".  It is
measured at various frequency, and Keysight's uncertainty ranges from 7.6
mHz when measuring 1 kHz to 7.6 Hz when measuring the 1 MHz. I'm puzzled
why the uncertainty should be so high.A half-reasonable frequency counter
that has not been calibrated for a decade should do better than that!

Anyway, I thought it might interest some.


It does show, as someone else indicated earlier, that inductors are not
needed to calibrate LCR meters.