Cheap and fast way to measure inductance-frequency

[ctzof]

Hello this question was probably made before but I want to ask it from another perspective.

I want to build an inductive linear sensor. The inductance of the sensor changes during movement and this has as a result a change in frequency of an LC tank oscillator. So what I am doing at the moment is to measure the frequency of the oscillator and translate this value to a linear movement.

The method for measuring frequency is simple and has been used a lot. So basically I count how many pulses I have in a certain period of time (1 ms) using a microcontroller counter. My working frequency is 20 MHz at the moment, so so 20500 counts for example means 20.5 MHz.

I don't really care about the exact value of frequency or inductance but more in measuring the change of these values during movement.

My sensor is accurate enough (um) but I want to use alternative ways to measure the inductance-frequency which are faster and maybe less expensive. I should say at this point that the sensor should be able to measure more than one coils at the same time (2-4)

Any suggestions?

 

[D.A.(Tony)Stewart]

You could use a PLL with an LC tuned oscillator and Varicap possible and measure the voltage with a coarse adjustment to get the VCO in the sensitive range referenced against a fixed OSC at the highest practical resonant frequency.

Otherwise you can drive a fixed frequency and measure the phase shift with L in a tank circuit with high Q to improve the sensitivity using a precision logic phase detector which requires a precision comparator.

It depends on the range of drift and sensitivity you want. In the 70's we built an Eddy Current instrument with about 100ppm on inductance and automated self calibration using 100kHz and 200kHz for amplitude and phase to measure relative shift using PLL technology with quadrature mixers. Measurement time as 0.1us done every 1ms or 0.1mm of movement

 

[BradtheRad]

It sounds as though you want to plot movement versus frequency. It becomes very easy if you can watch an XY plot on an oscilloscope.

One tool that might help is a frequency to voltage converter. However it won't give you a linear correspondency to Henry value. The formula for LC resonance is based on the square root of LC.

 

[Warpspeed]

I want to build an inductive linear sensor.

The most usual way of using inductors to measure physical position is with a "linear variable differential transformer" (LVDT)

https://en.wikipedia.org/wiki/Linear_variable_differential_transformer

Its good because its produces a linear electrical output with displacement.

 

[D.A.(Tony)Stewart]

Since you have not spec'd L or the range, a solution is not possible. But my suggestions still apply for fast cheap & accurate.

 

[ctzof]

Sorry for my late response. My measurement range is about 20 mm of linear movement. I was thinking though maybe for one chip solution like a microcontroller. I mean I would need anyway a micro to linearize for example my output or to transmit my data under some protocol (for example I want to use CAN), so putting a pll I think would increase the total cost. Right now I am using only a micro and the oscillator circuit which costs under 0.8$. For the oscillator circuit I can further reduce the cost by using PWM with a low pass filter as an oscillator. LVDT is of course a solution but I want first to investigate eddy current based sensors.