Hi!

I`m trying to measure frequency response of DUT with NA and power meter, using NA as source. Lets say that I have measured source power at one frequency with power meter (-0.60 dBm). This signal goes to DUT. I measured SWR on DUT at the same same frequency, that would be 1.3, so the Reflection coefficient would be (SWR-1)/(SWR+1)=0.1304. Mismatch loss would then be -10*log10(1-0.01304^2)=0.0745 dBm. So due this, the actual power what goes into DUT is -0.6dBm-0.0745dBm=-0.6745dBm. Then I measure power output from DUT with the same power meter and the result would be -0.95dBm. So is the total attenuation due the SWR -0.95dBm-(-0.6745dBm)=-0.2755dBm. Am I using right calculations? I tried to do this procedure, and i still get about 0.1dBm error at those frequencys that have high SWR, so what goes wrong? Is the mismatch loss calculated right, and is there yet something else that I should measure/calculate? Im measuring with NA and power meter because I have measured S21 of the DUT with 2-3 different analyzers but somehow everyone gives me a slightly different response, so I can`t trust this S21 measurement. This project that I`m working on is actually a power measurement, but I added a switch into transmission line. So if I do the same measurement first with correct cables, then using these cables and switch, the measurement error can be calculated. This is how I got that 0.1dBm error. This error is on these frequencys that have high SWR (up to 1.3). I think in this measurement the desirable error would be something like 0.02dBm.

EDIT:

I askes actually same question a while ago, and I got advice to use NA source & receiver Power meter calibration. I have Rhode & Schwarz NRVS power meter. Can I do this calibration with this meter, or does it have to be HP meter?

Thanks!

I`m trying to measure frequency response of DUT with NA and power meter, using NA as source. Lets say that I have measured source power at one frequency with power meter (-0.60 dBm). This signal goes to DUT. I measured SWR on DUT at the same same frequency, that would be 1.3, so the Reflection coefficient would be (SWR-1)/(SWR+1)=0.1304. Mismatch loss would then be -10*log10(1-0.01304^2)=0.0745 dBm. So due this, the actual power what goes into DUT is -0.6dBm-0.0745dBm=-0.6745dBm. Then I measure power output from DUT with the same power meter and the result would be -0.95dBm. So is the total attenuation due the SWR -0.95dBm-(-0.6745dBm)=-0.2755dBm. Am I using right calculations? I tried to do this procedure, and i still get about 0.1dBm error at those frequencys that have high SWR, so what goes wrong? Is the mismatch loss calculated right, and is there yet something else that I should measure/calculate? Im measuring with NA and power meter because I have measured S21 of the DUT with 2-3 different analyzers but somehow everyone gives me a slightly different response, so I can`t trust this S21 measurement. This project that I`m working on is actually a power measurement, but I added a switch into transmission line. So if I do the same measurement first with correct cables, then using these cables and switch, the measurement error can be calculated. This is how I got that 0.1dBm error. This error is on these frequencys that have high SWR (up to 1.3). I think in this measurement the desirable error would be something like 0.02dBm.

EDIT:

I askes actually same question a while ago, and I got advice to use NA source & receiver Power meter calibration. I have Rhode & Schwarz NRVS power meter. Can I do this calibration with this meter, or does it have to be HP meter?

Thanks!

Your best bet would be to scrap the power meter and just use a fully calibrated

vectorNA (did you use a scalar NA?) You haven't mentioned if you did a full open/short/load cal of the NA before measuring your DUT. If you did, then this should give you the best results since it will take into account most of the errors.You mentioned the input swr of the DUT, but you failed to mention the output swr of the source, the output swr of the DUT, and the input swr of the power meter (or NA), all of which will contribute to errors. This is why one should use a fully corrected vector NA...and it should be used well below any specified gain compression level (-10dBm is generally a good starting point.)

Also, a power meter is a scalar device and ignores phase. Without phase information, and assuming some typical values for match, you can easily see +/- .2dB of variations.

The mismatch error between a source with a 1.3 output swr and DUT with an input swr of 1.3 will be about +/- .15dB.

If you don't have a trusty cardboard HP Reflectometer calculator handy, here is a link to an electronic one:

http://www.agilent.com/metrology/download1.shtml

BTW, this is a 3-day weekend, so don't expect any responses until Tuesday.