High precision calibration for the poor man
Hi guys,
I love your level of knowledge and I would like to get your thought about a
way to achieve a calibration of a high precision meter without a high
precision calibrator. Let me explain myself. I have many 6.5 & 7.5 digits
meter and only two calibrated and certified tools: a Fluke 5500A calibrator
and an HP 3458A meter. The 5500A is very stable but can only achieve 5.5
digits precision. The idea is to source a known stable voltage with the
calibrator, then use the 3458A to correct (lower) the offset in order to
feed a 6.5 or 7.5 digits meter in calibration mode.
Example : I need 10.00000V +/- 25µV. I set the 5500A (or any other very
stable voltage source) to 10.0000V. I get 10.000073V, according to the
known-good 3458A. Im now looking for a very precise way to step-down the
voltage (-70 µV) in order to achieve the rated accuracy needed to calibrate
the meter. That accuracy doesnt need to be maintained for more than one or
two minutes.
Is it a stupid idea? What can I use to add a very small and very stable,
variable negative offset to a known voltage? A variable resistor should not
be able to meet the required accuracy without a drift. So, perhaps a high
precision aop ?
Any idea greatly appreciated :-)
Samuel DEMEULEMEESTER
Presse Non Stop - Canard PC
http://www.canardpc.com
Tel : +33.6.13.73.4003
MSN : sam@x86.fr
Samuel DEMEULEMEESTER wrote:
Hi guys,
I love your level of knowledge and I would like to get your thought about a
way to achieve a calibration of a high precision meter without a high
precision calibrator. Let me explain myself. I have many 6.5& 7.5 digits
meter and only two calibrated and certified tools: a Fluke 5500A calibrator
and an HP 3458A meter. The 5500A is very stable but can only achieve 5.5
digits precision. The idea is to source a known stable voltage with the
calibrator, then use the 3458A to correct (lower) the offset in order to
feed a 6.5 or 7.5 digits meter in calibration mode.
Example : I need 10.00000V +/- 25µV. I set the 5500A (or any other very
stable voltage source) to 10.0000V. I get 10.000073V, according to the
known-good 3458A. I’m now looking for a very precise way to step-down the
voltage (-70 µV) in order to achieve the rated accuracy needed to calibrate
the meter. That accuracy doesn’t need to be maintained for more than one or
two minutes.
Is it a stupid idea? What can I use to add a very small and very stable,
variable negative offset to a known voltage? A variable resistor should not
be able to meet the required accuracy without a drift. So, perhaps a high
precision aop ?
Any idea greatly appreciated :-)
Samuel DEMEULEMEESTER
Presse Non Stop - Canard PC
http://www.canardpc.com
Tel : +33.6.13.73.4003
MSN : sam@x86.fr
volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
You are only making an adjustment of around 7.3ppm an offset stability
of 1% will contribute an instability of 0.073ppm.
This is readily achievable with a resistive attenuator without using
precision components (eg 1 ohm + 140k).
However the thermoemfs in the attenuator will need to be low and the
added output resistance (1 ohm) of the attenuator may be unacceptable.
One could also use a low drift chopper stabilised amplifier or
equivalent plus a low resolution DAC and a few 1% or ).1% resistors to
produce a lower output impedance source.
Bruce
Bruce
In message 01e601cb70a1$58aa9150$09ffb3f0$@canardpc.com, "Samuel DEMEULEMEEST
ER" writes:
I'm now looking for a very precise way to step-down the
voltage (-70 µV) in order to achieve the rated accuracy needed to calibrate
the meter. That accuracy doesn't need to be maintained for more than one
or two minutes.
Have you tried simply grabbing one of the plugs with your hand to
change the temperature of the metal-metal thermoelectric effect ?
70 microvolt may require more than a warm hand, but the principle
can be trivially extended with more sensitive junctions.
Poul-Henning
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
IMHO if you are not calibrating these devices officially for an
organization [and thus need strict procedures to pass] you needn't
really adjust the 5500a to output 10.000 000 Vdc.
What is most critical is a output voltage stable enough to complete the
test procedure, sufficient for the range under calibration such as a 1nV
if required, and an equally good calibrated meter, such as 3458a, to
confirm that. You adjust your DUT to match the readout of the 3458a.
Your DUT is effectively calibrated against the 3458a.
A good DMM is linear throughout its ranges, and a measurement taken
beyond the recommended calibration voltage can be extrapolated anywhere
in the range under test, and should measure 10.000 000V when given such
a voltage even if adjusted at 10.000 0073 V. Luckily calibration rarely
involves adjustment but just comparison of readings against a reference,
in this case the 3458a.
I fear that if you made some circuit to reduce the voltage without
confirmed testable precision, that you'd risk more inaccuracy that if
you simply take the stable output of the 5500a confirmed against the 3458a.
On 10/20/2010 5:54 PM, Samuel DEMEULEMEESTER wrote:
Hi guys,
I love your level of knowledge and I would like to get your thought about a
way to achieve a calibration of a high precision meter without a high
precision calibrator. Let me explain myself. I have many 6.5& 7.5 digits
meter and only two calibrated and certified tools: a Fluke 5500A calibrator
and an HP 3458A meter. The 5500A is very stable but can only achieve 5.5
digits precision. The idea is to source a known stable voltage with the
calibrator, then use the 3458A to correct (lower) the offset in order to
feed a 6.5 or 7.5 digits meter in calibration mode.
Example : I need 10.00000V +/- 25µV. I set the 5500A (or any other very
stable voltage source) to 10.0000V. I get 10.000073V, according to the
known-good 3458A. I’m now looking for a very precise way to step-down the
voltage (-70 µV) in order to achieve the rated accuracy needed to calibrate
the meter. That accuracy doesn’t need to be maintained for more than one or
two minutes.
Is it a stupid idea? What can I use to add a very small and very stable,
variable negative offset to a known voltage? A variable resistor should not
be able to meet the required accuracy without a drift. So, perhaps a high
precision aop ?
Any idea greatly appreciated :-)
Samuel DEMEULEMEESTER
Presse Non Stop - Canard PC
http://www.canardpc.com
Tel : +33.6.13.73.4003
MSN : sam@x86.fr
volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.