How to precisely measure the Q inductor?

[biff44]

Lets say I am winding an inductor, probably a small toroid on powdered iron, in the 200 to 50 nH range. How can I precisely measure the inductor Q in the 100 to 300 MHz range?

I have a complex filter design, and the simulator says I need an inductor Q of at least 200 for me to meet the spec. If the inductor Q is not at least 200, then the filter loss will be 0.1 dB over the specification. I do not want to make up the entire filter just to make the Q measurement! So how can I measure the inductor accurately all by itself.

I looked at various ways on the simulator:
1)Measuring reflection coefficient of a one port series connected L-C network, but the difference between a Q of 150 and a Q of 200 is barely noticeable on a smith chart. I am not sure the Network Analyzer cal will be accurate enought to truly see the difference.

2)Set up a two port, with an series connected L-C placed in shunt. The depth of the transmission null is related to Q, but once again the null depth difference for a Q of 150 vs 200 goes from -36 to -37 dB, so if the test fixture has any through leakage it swamp out the effect.

3) make up a simple 1 pole bandpass structure, with a parallel L-C. But the difference in insertion loss resonance is, once again, very small when the Q is varied from 150 to 200. I do not think I can calibrate out the fixture losses accurately enough.


Any ideas on how to measure?

Conversely, anyone know where I can get a Q=250 inductor? Digikey is fresh out. Has to be pretty small, like a T30-6 core size.

 

[Rainleo]

Re: Measuring inductor Q

Setup a 2-port measurement, use VNA to measure the 2-port S-parameter. Converter it into Y-parameters.
Q=-Imag(Y11)/Real(Y11)

 

[radiohead]

Measuring inductor Q

I think you're along the wrong path. Coils can have a Q of up to 150 in that frequency range, but there is some tolerance which you have to take into account.

1. Why don't you use another technology if the Q-factor is that important? SAW filter would be a good fit for you.
2. It is really that important that your passband loss is very limited? In a system that can be compensated for using an amplifier. What really matters with a filter is its rejection and the steepness of the passband to stopband transition

Measuring the Q-factor can be done easily with the VNA, as said before.

 

[Borber]

Re: Measuring inductor Q

Make the best inductors that can be made. Use 1.5-2mm silver coated wire and make helical air core inductors.

 

[biff44]

Re: Measuring inductor Q

Helical air coil? Do you mean just an airwound inductor, or is there something special about a "helical" wind?

Can not use Saw because of very high power.

If you know of a technique to measure such high Q on a Network analyzer accurately, I am all ears. I do not think it is possible, given various net. analyzer errors stackups, to see such a high Q. Being off a 0.01 in ρ would give you a huge error.

 

[Borber]

Re: Measuring inductor Q

I ment airwound. Here is a test circuit used to measure Q at 60MHz:

 

[icpinu]

Q factor measurements.

Hi,

I am measuring the Q factor of a coil with a VNA using two methods.
1 Transmission
2. Reflection

Do I need to have an impedance match for the weekly coupled small coils connected to the VNA. According to what I understand, no impedance matching is needed.

Is this correct?

Thanks

 

[mtwieg]

When I need to measure high Q inductors and coils, I never use a S11 measurement. As as been said, it just isn't accurate enough once your Q goes much above 50 or so. Instead I put the inductor in parallel with a capacitor that is known to have a very good Q (better than the coil). Then I bring a pair of mutually decoupled pickup loops connected to the NA near the inductor and couple weakly to it. A S21 measurement will show the absorption of the LC, and you can measure its Q from that.

Just earlier today I used this approach to make a 560nH inductor out of magnet wire. Its Q ended up around 180 at 25MHz.

 

[Mazz]

Biff

VNA has not the required accuracy for this kind of measurements. Inductor's companies use impedance analizer (Agilent/HP 42876A) to do the job.

If you need such a high Q, why don't you buy inductor already characterized (such as Coilcraft 132 series, in the range you are interested they are guaranteed up to 3A rms nad Q>200)?

Mazz

 

[icpinu]

Thanks for your replies. I cant' buy the coil that I need.

 

[Phytech]

In my experience, the most accurate way to measure the unloaded Q is to parallel a high Q capacitor with your coil, and then short this with an interrupted 50 ohm transmission line. You will get a notch resonance and measure the Q of this notch resonance.

It's difficult to get Q above 200 in 100-300 MHz range. You can easily get 150-160, or even 180. It might be easier to change your filter design, such as using elliptic design, wider band width, or different number of poles, you may surprise yourself you actually can meet the insertion loss requirement with lower coil Q.

Good luck!