mel
New Member
Posts: 9
|
Post by mel on Aug 26, 2016 12:56:00 GMT -5
While I have calibrated the VNA and tested a lot of devices, I seem to be limiting out in the S21 max reading. Let me explain, when I put in a coax toroid choke ala W1JR and read the S21 I get a good solid -30 dB at the frequency of choice, when I series two of them it only goes up to ~39-40 dB. Not the ~>60 dB I get with open connectors on the VNA. Am I getting "leak through" in my testing? If so, how do you mitigate it?
Thanks,
Mel, K6KBE
|
|
g3txq
New Member
Posts: 34
|
Post by g3txq on Sept 10, 2016 17:54:34 GMT -5
Mel,
You're getting the expected result!
If you place a high impedance between the A & B ports, and then double it, you will halve the current into Port B and you will see a 6dB increase in the measured attenuation.
Steve G3TXQ
|
|
w4dnr
New Member
Posts: 27
|
Post by w4dnr on Sept 11, 2016 13:33:28 GMT -5
Mel, You're getting the expected result! If you place a high impedance between the A & B ports, and then double it, you will halve the current into Port B and you will see a 6dB increase in the measured attenuation. Steve G3TXQ Steve, If there is s 6db *increase* wouldn't the reading be 30db + 30db + 6db = 66db rather than 40db ? Don W4DNR
|
|
g3txq
New Member
Posts: 34
|
Post by g3txq on Sept 11, 2016 14:20:19 GMT -5
Steve, If there is s 6db *increase* wouldn't the reading be 30db + 30db + 6db = 66db rather than 40db ? Don W4DNR No - the choke is not an attenuator with a characteristic impedance of 50 Ohms; so you can't add attenuation figures when you cascade chokes. Let's take the simple example of connecting a 10k resistor directly between Port A and Port B of the VNA: Picture Port A as a Thevenin source with an Open-Circuit voltage V, and a source resistance of 50 Ohms. During calibration we connect Port A directly to Port B, which has an input impedance of 50 Ohms. So a current of V/100 flows into Port B, and that is taken as the calibration reference. Now insert the 10k resistor between Port A and Port B; a current of V/10100 will now flow into Port B and the VNA will report S21 as -40dB because the current has changed by the factor 100/10100, and 20.Log(100/10100)=-40dB (close enough) If we now add another 10k resistor in series with the first, the current into Port B will drop to V/20100 - a change of 100/20100; so the VNA will report S21 as 20.Log(100/20100)=-46dB. Doubling the resistance between Port A and Port B has produced a 6dB change in S21. Measuring the choke's S21 tells us nothing about the CM attenuation it will provide in a practical application where it is operating in a system with impedances very different from 50 Ohms - it's simply a convenient way to determine its complex impedance. Steve G3TXQ
|
|
g3txq
New Member
Posts: 34
|
Post by g3txq on Sept 12, 2016 3:51:39 GMT -5
To confirm that calculation I made these measurements this morning on my VNA2180:
10k resistor between Port A & Port B: S21 = -40.06dB 2x10k resistors in series between Port A & Port B: S21 = -46.07dB
Just as you would expect from the theory.
It has nothing to do with "leak through"; I measured S21 = -90.87dB with Port A and Port B isolated.
Steve G3TXQ
|
|
w4dnr
New Member
Posts: 27
|
Post by w4dnr on Oct 19, 2016 21:45:06 GMT -5
Thank You Steve
It makes sense the way you explained it.
You are a good teacher.
Don W4DNR
|
|