From: email@example.com (Winfield Hill firstname.lastname@example.org)
Subject: Re: Inductance to Voltage Converter
Date: 28 Sep 2002 07:46:13 -0700
NNTP-Posting-Date: 28 Sep 2002 14:46:13 GMT
Robert Baer wrote...
> Bruce Kingsley wrote:
>> I'm looking for an accurate, stable circuit for converting Inductance to
>> Voltage for monitoring an Inductive sensor position. Here's the facts:
>> 1) The inductor sensor (2-wire) varies 7 mH to 9 mH based on target's
>> (ferrous) distance.
>> 2) Excitation must be 1kHz @ 400mV.
>> 3) DC Output desired is a clean 0 to 10 VDC.
>> 4) The sensor is known as an Inductance Proximity Sensor.
> Well, it sounds like that the inductance of the sensor changes when
> iron or other ferrous material gets near it. Putting the inductor in a
> bridge or as a part of an oscillator would be the ways to get an AC
> signal that would represent the inductance. [snip]
A bridge circuit has substantial advantages, for example it measures only
the inductive portion of the sensor reactance. Also, very long cables to
the sensor are not a problem.
The simple circuit below, adapted from my highly-sensitive capacitance
bridge, has further nice properties. It is very stable, depending only
on the reference inductor. It's highly robust and isn't at all sensitive
to interference and noise, because it measures in a narrow frequency band
about the excitation frequency; the bandwidth is set by the response time.
It can be very fast, e.g. 20ms, and still be insensitive to noise.
7 to 9mH
sensor Lx ,...............,
______ ____ ____ : shielded box :
| coax : | transresistance
o (to 20m) ,...........' | 10k amp :
| : / ____ +-- SJ ----->-,
+---> AD734 -----\/\/----OOOO--' : |
| MPY : / 10mH ref Lr : |
| | '...........................' |
| inverting |
| DC summer |
| | | error AD734 |
| | '-----<--+-- integrator --<- MPY -<--' 10V at 7mH
| offset | | 0V at 9mH
| '-------> filter --- | ---->> OUT
+--------->---- integrator ---->-------'
| -90 deg phase Vo = 80 (Vs Lr/Lx -0.44)
1kHz, 400mVac Vo = 0.32/Lx - 35.6
My bridge circuit has a DC voltage output. As drawn, the 0 to 10v output
is proportional to distance, or 1/L rather than L. The circuit can work
over temperature extremes and doesn't need much power. Of course, some
work is necessary to complete it, analyzing the noise and drift, loop gain
and stability, etc. But that takes some serious engineering, which would
come at a price. :-)