Dear John, LF Groups,
I have been trying to measure the resistance and reactance of various
resistors at this frequency and have had difficulty getting a good
resistive balance....
...The bridge is a GR1606A...
...For resistances of the order of 100 ohms and higher I do not seem to
have a problem. For resistances of the order of 10 ohms I do not seem able
to get a decent balance.
I found the manual and schematic for this bridge at
http://www.k7jrl.com/pub/manuals/gr/1606/GR%201606-A%20RF%20Bridge.pdf.
Looking at the circuit, the "Unknown" arm of the bridge consists of the
impedance being measured plus a 390 ohm resistor, 220pF "initial balance"
variable capacitor and 220pF "reactance" capacitor all in series. Assuming
these capacitors are near mid-range when balancing your 10ohm resistor, the
reactance of the resulting 55pF series capacitance at 137kHz is about 21
kohms. So the reactive component of the unknown arm impedance is about 2000
times larger than the resistive component, and the null will be
correspondingly about 2000 times more sensitive to a fractional change in
capacitance as it is to a fractional change in resistance. This all means
that, in order to get a sharp null when adjusting the resistance control,
you will also need to adjust the reactance control very precisely, and you
will need to get the null depth down to say 80dB or more to get a good
resistance reading. I doubt it is possible to adjust the reactance control
that precisely - it will go out of balance if you give it as much as a funny
look! If you are measuring 100 ohms, the problem is made 10 times easier, of
course.
A related problem is that the resistive component measurement will be very
sensitive to dielectric losses in the capacitors, which have very high
impedances at low frequencies - this will lead to frequency-dependent errors
that become larger as the frequency decreases.
I guess that, excellent though bridges of this sort are, their makers had to
optimise the design for the particular range of frequencies and impedances
they thought their customers would most like to use it for. The GR1606A
seems to be aimed mainly at the MF/HF range, so I suppose compromises must
be accepted at LF. If you just want to measure resistance, you might be
better off adapting a simple "noise bridge" circuit as often appears in the
amateur handbooks, provided the parasitic reactance of your resistors is
fairly small.
Cheers, Jim Moritz
73 de M0BMU
|
