HOME / KC901 serial VNA

S21 isolation calibration

By Ivan_Ionov •  2017-11-09 17:25:42 •  324 clicks
I've got recently KC901V and playing with it, it's a very good and promising product, but it has some weak points.

I see, in S21 measurement it has poor isolation on 5-7GHz range, real measurement range is only 40-50dB. But if I need to add 3-10dB attenuators to improve ports 50om matching and remove measurements ripple, the useful range becomes even smaller, down to 30-40dB. It's even worse, than s21 measurement on some cheap products, like NWT6000 I have.

I think, most leakage comes from internal RF switch, and this leakage should not change phase depending on DUT load, so this leakage can be compensated by calibration.

I think, as you have a vector S21 measurement, you can record S21 vector response (with disconnected output and input ports), remember it in internal memory and subtract it from real S21 measurements. With this calibration, you will have 60-80 range on 5-7GHz, and this will be a great improvement.

Regards, Ivan
3 Reply | until2018-12-13 12:45Add a reply
    VK3NH  2017-11-14#1 -Reply

    I am not an expert in the software capability of the KC901V but I will offer the following comments based on my understanding of the process and use of the KC901V.

    1/ Corrections as part of the SOL calibration occur at a number of points dictated by the selected start and stop frequencies. (User calibration is voided if either of these is changed)
    2/ You are correct in your observation that phase of the leakage should be constant but that is only at a set frequency.
    3/ As the S21 function is a frequency swept process, compensating for the leakage based on phase I don't believe is possible.
    4/ If the software for the field strength function at a set frequency with its associated loss calculation option was modified to be applied to the calculation then that might work in achieving a greater dynamic range.

    I happy for you to challenge the reasoning of any of the points above.

    regards Noel

    Ivan_Ionov  2017-11-10#2 -Reply

    Hello Noel,

    I think, ripple in S21 measurement is a different problem (not ideal 50om matching of the output port reflections between DUT and ouput port), it can be solved as you say, by adding 6 or 10dB attenuator between output port and the measured device. Idealy, we can build 10dB wideband power amplifier 10dB attenuator, this will make both thing together - it makes better 50om matching and doesn't drop signal amplitude.

    Poor isolation is a different thing, I beleve it comes inside the KC901V, in the RF switch, which directs signal to either port1 or port2.
    This signal leakage is stable, and it won't change its phase. That's why I think it can be compensated in software.

    VK3NH  2017-11-10#3 -Reply


    I agree poor isolation in the upper range is an issue.

    Whilst I suggested the use of the extra fixed attenuation as indicated above you are right about the loss of dynamic range. But doing what you suggest will not compensate for the phase change that happens when you are measuring say a HP stepped attenuator set (0 to 110dB) as I have done by first calibrating the unit at 10dB step because the steps in switching in and out attenuation actually change the signal path length.

    However if you just use the KC901V as a signal source for substitution transfer comparisons with a reference attenuator set or for tuning a filter notch with any suitable spectrum analyzer or measurement receiver, then the isolation problems go away.

    regards Noel


LoginPublish or Still no account?Register


Theme List

KC901 serial VNA