Topband: Measuring Common Mode Chokes

[Michael Tope]

On 12/18/2019 3:46 PM, Chuck Hutton wrote:
> In the past, I have simply used my N2PK VNA to measure impedance of the choke by connecting the shield of the coax to the VNA ports.
>
> Recently I've been discussing common mode chokes with others who have a different methodology.
> They prefer to do a transmission test through the choke and report the "common mode rejection".
> This is done by placimg a crossover cable between the VNA output and the choke. The choke output is connected in a normal fashion (center to center, shield to shield) to the VNA input.
>
> This does not seem ideal to me.
> First, the choke is being driven in differential mode rather than common mode.
> Second, the measurement depends on (varying) isolation between the coax center and shield. So it's not truly common mode rejection.
>
> Am I on thr right track?
> A handful of Googles has not netted me any clear summary of test methodology for reportimg CMRR. I fimd a small number of tests reportimg impedance.
>
> Chuck
>
Chuck,

I agree with your assessment of the "crossover cable" method. If 
understand your description correctly, that puts the choke into a 
transformer mode where the shield is the primary and the inner conductor 
is the secondary. IMO, that is not a good way to measure common-mode 
impedance since the common mode impedance will appear in parallel with 
the 50 ohm impedance provided by port 2  of the VNA. In principle, I 
suppose you could de-imbed the common mode impedance from the measured 
S11. In practice, I can't imagine that being very accurate since you are 
trying to de-imbed a very high impedance from a very low impedance. A 
small measurement error, would lead to very large errors in the estimate 
of the commode mode impedance.

Connecting the respective ends of the choke's coax shield to the 
respective center conductors of the two VNA ports is a much better 
method (i.e. VNA_P1--<<--|Zshield|-->>--VNA_P2). In that case the 
magnitude of the S21, (i.e. MS21dB) should equal 20*log[|50/(Z+50)|]. 
For large Z, this simplifies to ~20*log(50/|Z|). Thus, if the magnitude 
of S21 at some frequency is -40dB, then the magnitude of Z at that 
frequency is ~5000 ohms. If you do the math using the phase of S21, you 
should be able resolve the resistive and reactive components of |Z|. I 
think a method similar to this is what K9YC recommends in his app notes 
on ferrite chokes.

73, Mike W4EF.........