VNA-X, measure IMD3 with the LibreVNA

[satfan52]

Professional VNAs of the VectorStar line of Anritsu can be used to characterize non-linear devices such as power amplifiers. In addition to measuring S parameters, they can also generate a two tone signal centered on a given frequency and of a given signal power; then measure the corresponding IMD3. This feature is unavailable for hobbiest so far on low cost equipment, adding it to the LibreVNA would make it the first true VNA-X for hobbiest. LibreVNA can already do power sweeps and be used as a basic spectrum analyzer (if my understanding is right), so adding the IMD3 measurement would be a great asset and the last one missing to offer basic VNA-X capabilities to hobbiest.

[jankae]

This is not possible with the LibreVNA. As you mentioned it requires the simultaneous generation of two signals and the LibreVNA only has one source. You would either need a second source or a way to FM modulate the existing source. Both options require additional hardware and are not very cheap. I don't really want to include that by default in any new hardware versions. Maybe I can come up with some extension module to cover this functionality.

[satfan52]

If only one source is available and if it is not possible to modulate a DSB signal, maybe an approach with two LibreVNA used in T|Rmode using a combiner on port 1 to generate the two tone and a spliter at the ouput of the DUT to send the test signal to port 2 of each Libre VNA would work ?

If I as suspect it only one VNA receiver is needed to measure the IMD, maybe the second receiver of the second VNA can be connected to the input combiner 2 too, to measure the input signal and make adequate corrections in the calculation of IMD (if needed)

The PC app would have to be modified to be able connect to both LibreVNAs and to calculate the IMD products.

I know it may seem an overeach to use two librevnas but on the other hand, if IMD measurements can be automated it is worth the investment of a second LibreVNA in my view (considering the LibreVNA is not so expensive)

Finally, for optimal return on investment, the second librevna could also be used to create a virtual 4 port VNA provided the PC app is modified adequately too.

[jankae]

maybe an approach with two LibreVNA used in T|Rmode using a combiner on port 1 to generate the two tone and a spliter at the ouput of the DUT to send the test signal to port 2 of each Libre VNA would work ?

Yes, from a hardware perspective this should work. The software is almost there: you can measure IMD3 in the spectrum analyzer mode already but you need to provide all sources externally. You have a tracking generator in the spectrum analyzer mode as well but it always tracks the measured frequency (with an optional offset). For this measurement to work, you would need a "non-tracking" generator, that is a fixed frequency.
Neither the source nor the receiver levels in the LibreVNA are particular accurate, so I am somewhat doubting how useful this measurement would turn out (unless you calibrate them).

Finally, for optimal return on investment, the second librevna could also be used to create a virtual 4 port VNA provided the PC app is modified adequately too.

That is already possible, you just lose phase information for transmission measurements between two physical VNAs (e.g. phase information is there for S21 but not for S31 if port 1/2 are on one VNA and port 3/4 on the other). This is another hardware limitation as the LO can't be shared between the VNAs.

[satfan52]

Thanks. I understand very well that a single LibreVNA could be used in spectrum analyzer mode with two external power sources to measure the IMD3 in dBc. This is what I do it with the spectrum analyzer that I own already....but this is a slow and extremely painful procedure when it has to be repeated at different frequencies and different power levels. This very painfull issue is very well known by all people who characterize amplifiers and can't afford professional VNA-X equipment costing thousands of Euros. Doing such measurements with rhe LibreVNA in spectrum analyzer mode would of course require a proper linearity of the LibreVNA spectrum analyzer itself which I assume is the case because otherwise, the IMD3 measured will be the one of the LibreVNA instead of the DUT one ! If the linearity is not good enough, there is indeed no point using the LibreVna in spectrum analyzer mode for such IMD measurements. My initial point was about automating the IMD3 measurement over a given frequency sweep (at a given power) or over a power sweep (at a given frequency). This is where I find tremendous value in investing in two LibreVNAs provided the software allows to automate such IMD3 measurents '(in dBc) over frequency or powersweeps. Thank you for you patience and your great work on the LibreVNA Regards Peter Le mer. 15 nov. 2023, 13:34, Jan Käberich ***@***.***> a écrit :

…

maybe an approach with two LibreVNA used in T|Rmode using a combiner on port 1 to generate the two tone and a spliter at the ouput of the DUT to send the test signal to port 2 of each Libre VNA would work ? Yes, from a hardware perspective this should work. The software is almost there: you can measure IMD3 in the spectrum analyzer mode already but you need to provide all sources externally. You have a tracking generator in the spectrum analyzer mode as well but it always tracks the measured frequency (with an optional offset). For this measurement to work, you would need a "non-tracking" generator, that is a fixed frequency. Neither the source nor the receiver levels in the LibreVNA are particular accurate, so I am somewhat doubting how useful this measurement would turn out (unless you calibrate them). Finally, for optimal return on investment, the second librevna could also be used to create a virtual 4 port VNA provided the PC app is modified adequately too. That is already possible, you just lose phase information for transmission measurements between two physical VNAs (e.g. phase information is there for S21 but not for S31 if port 1/2 are on one VNA and port 3/4 on the other). This is another hardware limitation as the LO can't be shared between the VNAs. — Reply to this email directly, view it on GitHub <#227 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADR72WP4FF7QH2BSGUHHGVLYESZGHAVCNFSM6AAAAAA7DWEGESVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TKNZWGUZDE> . You are receiving this because you authored the thread.Message ID: ***@***.***>