How do you measure the value of an unknown inductor? If you have an LCR bridge or meter, you are probably going to use that. If not, there are many different techniques you can use. All of them rely on the same thing my Algebra teacher Mr. Harder used to say back in the 1970’s: you have to use what you know to get what you don’t know.
[Ronald Dekker] must think the same way. He took a 50-ohm signal generator and a scope. He puts the signal output to about 20kHz and adjusts for 1V peak-to-peak on the scope. Then he puts the unknown inductor across the signal and adjusts the frequency (and only the frequency) for an output of 1/2 volt peak-to-peak.
The idea is that the magnitude of the inductive reactance at the half-way part must be 50 ohms (forming a 50/50 voltage divider with the source impedance). [Ronald] does the math derivation in detail, but it works out that the inductor (in uH) is 4570/f where f is the frequency in kHz. In reality, the setting of the 1V reference is not completely necessary, but it simplifies the way he does the measurement if you read the full post.
Of course, there is probably some stray resistance in the circuit, but not enough to make much difference in most inductors. If you have reason to suspect otherwise, [Karen Orton] contributed the math to get to a slightly more complex expression that lets you factor the DC resistance of the coil in your calculations.
This isn’t the only game in town, of course. We like measuring inductors with a grid dip meter. On the other hand, if it’s input or output impedance that you’re interested in, go talk to Elliot.
I love reading about alternate methods and tricks for determining unknowns like this… Great stuff!
Nice.
I use to take the 1khz reference output of a standard oscilloscope, put it through a known capacitor then to a inductor to ground. Measure across the inductor with the same oscilloscope and look at the ringing frequency. Calculate from there.
The only tool required is a standard oscilloscope a capacitor and some maths.
Any arduino-ey boards fast enough with good enough ADC to build a black box that does this?
STM32F103 has dual 12-bit 1Msps ADC and lots of goodies. You can get those $2 boards from China. Use an external DAC+driver via the SPI for the signal generator.
You can make an LC oscillator and use *duino to measure frequency with known capacitance and then calculate inductance. There are few microcontroller-based LC meters online. Some of them use external or internal comparator to form LC oscillator…
There is also AD9833/AD9834 that can be used to measure complex impedance, which then can be used to calculate parameters of component or network.
Instead of trying to build the AC meter by sampling with the ADC, you might be better off with an active rectifier and measure the DC value. You can hit the bandwidth and sampling limits on the frequency on the ADC vs buying faster opamps.
Check out this simple circuit for use with multimeter (PDF in Polish):
http://serwis.avt.pl/manuals/AVT2813.pdf
Inductors are funny things.
If they are very small (physically or just low value) getting an accurate measure requires compensating for the test leads, etc.
If you are running at a high enough frequency you get capacitance between coil turns
Not to mention stripline effects, where a PCB track with an open circuit at the end becomes inductive.
Inductors can be quite big, out of a Prius:
https://youtu.be/CKiQwzH4n-E
Also, video shows the same measurement technique. ;)
The part that seems a little odd to me is owning such a nice signal generator but not owning a simple LCR meter. I guess one could rig something up with a DDS chip. You would need an adjustable amplifier to get the output to a known level. Even then the output level would probably change as you move the frequency. I guess you could split it into the second trace of a dual-trace oscilloscope and keep adjusting the output as it moves.
There’s the Poor Ham’s Scalar Network Analyzer (PHSNA). Uses an Arduino, driving a DDS chip and getting amplitude from a Log amp/detector.
DDS chips tend to change output with frequency, so calibrate by sweeping frequency without inductor. Repeat with inductor across terminals.
You could also set the frequency to about 10 Hz and get a good approximation of the DC resistance.
I played with this a while back, using a square wave generator using a PIC16 instead of the smitt trigger used in the video:
https://youtu.be/74fz9iwZ_sM
Works well except my edge transitions aren’t fast enough (probably) and dividing the resulting inductance by 3. I used an excel sheet to incorporate the error and give the correct values. I checked with known inductors and it seems to work. I made the frequency adjustable using a potmeter from 1khz to about 20khz. I replaced the 10pF capacitor with a 560pF capacitor, it gave me a stronger signal. I used a breadboard but it still worked ok down to 4.7uH.
I don’t trust my super-cheap function generator to put out a constant voltage as the frequency changes, nor do I trust my voltmeters to measure voltage accurately at high frequencies. So I’ve been using a ratiometric approach, measuring two voltages, across the inductor and across a series resistor, at several frequencies, then fitting a model to the data. See https://gasstationwithoutpumps.wordpress.com/2013/06/24/fitting-l-and-r-values/ for example. (This method can handle more complicated devices than just inductors—I usually use it for characterizing loudspeakers, which may have multiple mechanical resonances.)