I Inductance Measurement: Error Due To Series Resistance

[BlackMelon]

Hi!

Please refer to the word file.
I am thinking of measuring inductance of an inductor by applying a DC voltage onto it. Then, calculate L = V/(di/dt).
I wonder if the series resistance (R) of the voltage source plus of the inductor under test would cause much error.

So, I derive the transient responses of the cases with and without R. I calculated the error between the two cases and got the indeterminate form.

I would like to know how to properly analyze this problem?

Thanks!
BlackMelon

 

[berkeman]

Can you attach a PDF version of the document please? That's safer for folks to open than a Word file (which can contain macros). Thanks.

 

[osilmag]

It would be easier/less error to apply an AC voltage with a known frequency with the resistor and inductor in series. That way you can find the impedance of the inductor Z= jwL by measuring the current and voltage normally instead of measuring the change of the current.

 

[DaveE]

I didn't look at your word file.*

But, you can use this method at any point in time on those transient response curves. But you have to know the voltage across the inductance at that time. Since the series resistor will drop some of that voltage, the best time is likely at the start when the inductor current is zero.

You might also assume an exponential response and fit a model to any two measurements (two different times) to resolve R and L.

Most impedance measuring instruments will use a steady state ac excitation at a single frequency then measure the gain and phase response for more accuracy. You could also just use two measurements of magnitude at two different frequencies.

Often the best method depends on your application.

*You'll get more and better answers from us if you make it easy to understand your question. I'm happy to help if it's easy. Otherwise I charge $140/hour.

 

[Baluncore]

I would like to know how to properly analyze this problem?

The way you analyze the problem, will depend on the measurement(s) you make with your instruments.
What range of inductance do you intend to measure?
What voltage or current parameter will you measure, over what period of time?

 

[BlackMelon]

Thank you all. Next time, I will make the problem easier to understand.

The way you analyze the problem, will depend on the measurement(s) you make with your instruments.
What range of inductance do you intend to measure?
What voltage or current parameter will you measure, over what period of time?

Inductance range 10uH to 10mH
Voltage = 12Vdc
Current and period of time are < 3 A and 10 ms (may be)

I have just got an idea. I will just apply the 12Vdc until the inductor reaches its steady state to measure the series resistance (Rseries = 12Vdc/I_steady).

Disconnect and reconnect the 12Vdc again. This time, monitor i
Vseries = i*Rseries ----> v_L = 12 - Vseries -----> L = v_L*delta t/delta i

 

[Baluncore]

The inductor should always be paralleled with a resistor during the test. When the supply is switched off, the negative voltage inductive spike is then attenuated, preventing damage to the inductor insulation. For the same reason, the switch should not be in series with the inductor.

Avoid diodes if possible. They introduce more problems than they solve. I have used, and suggest a circuit like this, for the meter.
Note that when the battery is switched off, after the long "charge" period, Vout produces a negative step, that then decays to zero. The initial positive and then negative voltage of the step, give the internal resistance of the inductor. The inductance can then be computed from dVout/dt as the current begins to decay in the Rref loop.