[volt-nuts] users of Fluke 750A's beware

[Stephen Grady]

Fellow Volt-Nuts,

 

I notice that in recent discussions regarding the replacement of the
batteries in Fluke 750A’s seems to indicate that they are still being used
to make precision voltage measurements.

Now I haven’t used one in about 30 years and when I did it was part of a
Fluke 7105A with a Fluke 335D, Fluke 720A, 721A, 750A and 845AB. The voltage
reference was a bank of four standard cells in an enclosure.

 

Now I am going to describe an issue I had with my Fluke 750A I cannot say
that every Fluke 750A was like this but I feel it is important enough to at
least make everyone who uses these to ensure that theirs is not the same.

 

The problem comes down to current flow in the divider in my Fluke 750A . In
my divider current flowed between the output low terminal and the input low
terminal and this developed a potential difference of approximately 20 µV
between the input and output low terminals and depending how the divider was
guarded or earthed could lead to an error of approximately 20 µV/V (ppm). 

 

For those who are reading this out of interest and not familiar with the
750A it is a single unit rack mounted box containing input terminals on the
left hand side and output terminals on the right hand side;  there are also
an input switch (10-1000V) and an output switch (0.1V-1000V); there are also
terminals to connect a standard cell, switches to select the standard cell
voltage, terminals to connect a null meter and a toggle switch to make the
divider – standard cell – null meter circuit. The idea of this divider is
you supply an input voltage to the input terminals and set the input switch
to 1000V and supply approximately 1000V to the input terminals, this
supplies 1000V to the 1000V tap of the divider and if a meter is connected
to the output terminals and the output switch selects a voltage on the
switch then this voltage is (approximately) presented to the meter. To
increase the accuracy of the output voltage a standard cell of known voltage
is applied to the standard cell terminals with a null meter connected to the
null meter terminals and the standard cell voltage is dialled on the 750A’s
standard cell terminals, the input voltage is then adjusted for a null on
the null meter. At this point the reference divider (750A) is “standardised”
and the output from the output terminals (provided the 750A’s ratios are
corrected calibrated) are then supposed to be within the stated accuracy of
the divider (this is another issue in that Fluke rated the accuracy as 10
ppm although I believe is can be greater but this is not the issue of this
discussion and I will leave that for another time if there is sufficient
interest).

 

Normally using a divider no current is drawn from the output terminals
otherwise this loads the lower section of the divider and changes the
ratio/output voltage. For this reason a second null meter (or the one used
for the reference adjustment is moved) is connected between the output high
terminal and the voltage being measured. In this way the reference divider
is always used in a null bridge arrangement and zero current is being drawn
from the bridge.

 

Getting back to my original problem of the difference in potential between
the output and input low terminals, I found in my divider that the bottom
resistor of the divider was connected to the output low terminal and then to
the input low. I believe this was a mistake as the bottom of the resistor of
the divider should have been connected to the input low and then to the
output low, so no current (potential difference) flows in the low circuit.

 

Now Fluke actually calibrated this voltage drop into the calibration
build-up of the 750A against the 720A as the low was grounded at the low
output and the link on the 335D low to ground was removed.

 

This is why I believe all 750A’s are like this as the voltage drop between
the lows is accounted for in the calibration build-up. So when using the
750A it is critical to always ground the low at the output low and to insure
that the input low is not grounded. If it is and you are standardising to a
standard cell then there is an approximate 20 µV/V error although if you are
standardising to a 10V reference then this error will drop to 2 µV/V which
is not significant compared to the calibration error of the 750A.

 

I hope this was of some assistance to users of Fluke 750A’s

 

Kind Regards,

 

Stephen Grady

Sydney Australia