Quality factor capacitor

[Darkcrusher]

Hi,

Anyone has an idea about the accuracy of the Q-factor specified by manufacturers in their data sheet?

Maybe the Q of a specific capacitor should be measured ? Is this easy without advanced equipment (eg boonton resonant line) ?


Thanks!

 

[Phytech]

The errs range from 1 to 100 times. You can measure the Q, but you need equipment to do it. If you have a network analyzer, it would easy to measure the Q at the frequency of your interest.

 

[Darkcrusher]

Yes I have acces to a network analyzer, what setup do you recommend ?

 

[Phytech]

What I do is the following:

You need an inductor, such as a coil, and a capacitor with known unloaded Q, Qck (the higher the value, the more accurate the final result, which can be ATC capacitor or PCB capacitor cut from Rogers material, also with about the same capacitance value.

First, form a parallel resonance from the coil and the know capacitor (e.g. you can cut open a 50 ohm transmission line on a PCB and short the gap with the know capacitor and coil). Connect it to a network analyzer, you will see a notch at a frequency. Measure the 3-dB bandwith of the notch and the center frequency, you will get an quality factor Q1=fc/(df3-dB). Network analyzer measure the Q directly for you if you can find the function key.

Then replace the know capacitor with the capacitor you want to measure, and measure the Q you will get Q2.

Now

1/Q1=1/Qcoil+1/Qck
1/Q2=1/Qcoil+1/Qcx

You get 1/Qcx=1/Q2-1/Q1+1/Qck, The higher the Qck, the more accurate the final results.

Good luck!

 

[Darkcrusher]

Thanks,

But how can you be sure about the reference capacitor Q ?

 

[Phytech]

There are different ways to find that out.

1. If you know the Q factor of the coil (Qcoil), you can directly find out Qck through the above measurement. It is better to find that out through several coils with different Q, which may also give you an err margin. Save this capacitor for future reference.

2. Use capacitors with very high Q factors, such as air filled capacitors, or single layer capacitor with Qck>3000 at say 1 GHz. When the Qck is very high, you can verify that when you use Qck=3000, or Qck=6000, there is only small influence on your final results. If you use air filled capacitor, you can easily assume Qck=5000, or above, which has little influence to your final results.

3. When you do more experiments, you will know the approximate range of Qs of some capacitors, and use these capacitors as your reference.

My advice is try several type of capacitors, and parts from different vendors, as your reference capacitor.

Good luck!

 

[Darkcrusher]

Thanks,

I will experiment with different capacitors within the same vendor and other venders, seems like the only way. Measurement and statistical analysis