EVM of 256-QAM with envlp analysis

Hello,

I am a student, and new to Cadence and SpectreRF. I need to the determine the EVM introduced by outphasing PA's in case of 256-QAM, preferably by using a fast envlp analysis (instead of a much slower transient analysis).

First, I did some envlp and EVM simulations by using the standard 802.11.bin files (from rfVsource ) as baseband sources combined with the EVM function, the results were as expected, but the maximal constellation size in this case is 64-QAM.

So I tried to extend this to 256-QAM, but since I could not open these .bin-files, I used pwl sources as I/Q sources, and (longer) pwl files which I made in Matlab, by baseband filtering the symbols with a square root raised cosine pulse. Now, I try to determine the EVM in this case, but directly calling the EVM function gives the "continuous time", filtered constellation diagram, since I did not find a way to specify the baseband receiver filter.

I did not find a way to do this, so I added a downconverter and demodulator to my schematic and I did the baseband demodulation by using the square root cosine block in rfLib. But now, when I do an envlp analysis and save the baseband output waveform, it is constant and zero. I think this is because SpectreRF does not realise that the output signal is already a baseband signal, and it determines the envelope of something which is already the envelope; so it obtains 0. (When I do a normal transient analysis on the same schematic, the output waveform is correct. So I could save this output waveform and to the sampling and EVM determination in Matlab, but this is less convenient.).

I suspect that my last step is not necessary,  and that it is possible to define a '"communication standard" similar to the 802.11 standards which are already available in the EVM function, by providing the used transmit/receive pulse. Then SpectreRF directly knows how to demodulate the envelope and should output an EVM and constellation diagram as if an 802.11 standard was used.

How can I do this?

Greetings,

Thank you for your time,

Joris Lambrecht