Question
I used SDF to plug in EVAL-AD527XSDZ and power the system by one USB port of
the laptop, then using multimeter to test the resistance and compared with the
target value shown in the software.
1) If using Agilent 34401A digital multimeter, I got several different
results
a) When A is connected to 3.3V, “+”of 34401A is connected to A and “-“ is
connected to “W”, the resistance from 34401A is correct.
b) When A is connected to 3.3V, “-”of 34401A is connected to A and “+“ is
connected to “W”, the resistance from 34401A is wrong in some cases.
c) When A and W are floating, the results seem correct, which are
independent of the connection between 34401A and resistor terminals.
d) When A is connected to AGND, “-“ of 34401A is connected to A and “+” is
connected to “W”, the results are correct.
2) If using handhold multimeter, the results are always correct, which are
independent of the correction between handhold multimeter and resistor
terminals.
From the above phenomenons, I guess the reason of the wrong results may be the
voltage of terminal A or W is above 3.3V. In the datasheet, the voltage of
terminal A and W should be between VSS and VDD.But, why the results are always
correct when using handhold multimeter?
Answer
The internal multimeter has a current source, reference to LO input. Depending
on the impedance range, this current source forces more or less current. As you
pointed before the voltage across the terminals cannot exceed 3.3 +0.3V, in
your case. In your first configuration 1a, Agilent force current from A to W,
the measurements in this case are correct bc your current source reference is
floating.
In your second configuration 1b, LO terminal is connected to 3V3 and indirectly
you grounded your current reference, using the LDO.
Consequently, the voltage generated in the other extreme of the resistor, W,
can exceed the maximum allowable voltage, 3.6V and the internal diodes will
clamp this voltage.
In your third and fourth configurations, you are not exceeding the maximum
allowable voltage so no error should be measured.
Answering your second question, if the current used by the multimeter is small,
the voltage generated by the resistor should not exceed 0.3V.This is the case
of a handhold DMM.
In general I’d recommend measure the resistance without any connection to the
circuit, mainly bc the impedance that you are measuring could be wrong as the
DMM always measure the lower impedance.