How to calibrate VNA with high path loss between VNA port and Cal plane?
Port 1 and 2 have about 25dB path loss ( Cable + fixture )
SOLT and SOLR can not give good results.
A response calibration should be possible, if you only want a rough idea of S21 and S12. But otherwise I don't see how you are going to get any reflection calibration though 25 dB of loss.
Can you develop an EM model of your fixture? If so, you could determine the S-parameters of your fixture, calibrate to the reference planes of the coax connectors. I assume that the loss of your cables is not so bad to make even that possible. I assume the problem is mainly due to the fixture.
HFSS is one commercial option. openEMS is a free open-source modelling program. Expect to have a computer with a lot of RAM.
drkirkby is correct. A response calibration is the only viable option if you can't reduce the loss in your cables and fixtures. Usually, in high power scenarios, the high attenuation is only on the output side of the DUT, in which case, you could do an enhanced response cal and that would give you S11 and S21, but with 25 dB loss on the input, you can't even do that. Is the loss just from long cables and high frequencies or is there some kind of an external testset and switches involved here? if you can provide some more detail about your measurement setup, we might be able to offer other suggestions.
the loss from long cable,switches and fixture, 25dB loss at high frequencies.a picture is given below.I need to get the S11 and S22, would you like to give some suggestions?tks
mindhen & Carvalho - in both of your cases, the only possible option is to move the input and output reflectometers close to the DUT, if your VNAs have direct receiver access. Neither of you mentioned the model number and options on the VNA you are using, but if you are using a PNA and it has loops on the front panel, then you have direct receiver access. Essentially, you would need to place a dual directional coupler after your switch (at the closest point to the DUT that you still have connectorized access to and feed the forward and the reverse coupled signals directly back to the reference and measurement receiver inputs on the front panel with as short a cable as possible. the output of the coupler would be your calibration plane and you can then de-embed any extra fixture path that maybe between calibration plane and the DUT. As Carvalho already discovered, demebedding a high loss path doesn't work for the same reason that you can't do a full 2-port cal through that path, but by moving the reflectometer closer to the DUT, you are hopefully putting most of the path loss behind the reflectometer, where it won't impact the calibration.
Hi daras. I had a similar setup, which has coupler`s before and after my DUT. The losses "Cable plus fixture" are around 13 dB (before and after DUT). I tried to deembeding my S-par. but, only S21 seem me correctly (or close my Amplifier datasheet info). I`m reading some notes and the S-parameters deembed will work only to low-losses. others Sparam (s11, s22 and s12) give a terrible results. So, how can I fix my issue? any suggestions? Regards,
Carvalho wrote: Hi daras. I had a similar setup, which has coupler`s before and after my DUT. The losses "Cable plus fixture" are around 13 dB (before and after DUT). I tried to deembeding my S-par. but, only S21 seem me correctly (or close my Amplifier datasheet info). I`m reading some notes and the S-parameters deembed will work only to low-losses. others Sparam (s11, s22 and s12) give a terrible results. So, how can I fix my issue? any suggestions? Regards,
It is polite to start a new thread, and not highjack someone elses.
See my response above. While a separate thread would be the standard protocol, when there is high overlap in the questions, I personally don't mind dealing with both of them in one thread, but I can't speak for the original poster.
tks to everyone.
Our problems are indeed highly overlapping.
it's ok for me in one thread.
PNA-x is used in our lab.
I still have a doubt.
by moving the reflectometer closer to the DUT, the cable loss between the dual directional coupler and receivers may be high.
how to deal with ?
If you simply can't shorten the length of cables (are you in some kind of TVAC or antenna range scenario?) then you either have to live with just the response correction or you would have to move the entire receiver closer to the DUT. we have a PNA-X based solution that works up to 120 GHz, using remote mixing heads (N5291A) that might work, but depending on your frequency range, it might be over-kill (it is not an inexpensive system). the question you have to answer is how critical are the S11 and S22 measurements for my device. With 25 dB of loss presented to the input and output ports of your DUT, you will be able to get a very reasonable S21 and S12 measurements with a response calibration. Many of our Satellite customers use response or enhanced response corrections when testing their integrated payloads, for precisely the same reason as you and are satisficed with the results. Putting together a full 2-port correction capable VNA with remote heads will be a costly solution if those reflection measurements are not absolutely required.
Retrieving data ...