AnsweredAssumed Answered

Accuaracy of X-parameter measurements

Question asked by DSTEPHENS on Aug 22, 2014
Hi there,

I've experimented with X-parameters, by (i) doing an extraction in ADS from a transistor model and (ii) a real extraction by measuring a single transistor. 

Comparing the results in (i) i.e. transistor model used v.s. ADS extracted X-parameter model gave very very (very) impressive results. As the differences observed were almost entirely due to numerical noise. 

Comparing the results in (ii) i.e. transistor measured characteristics v.s. PNA-X extracted X-parameter model gave some good results, though as expected the delta's were somewhat larger. Agreement for the fundamental term and 2f0 were pretty decent, however, the 3f0 term was not as good. (3f0 power level ~ -85 dBm, low but within the dynamic range of the analyzer).

My question: Have the engineers at Keysight done any sort of benchmark tests to demonstrate the level of accuracy that one can expect from X-parameter models generated from actual measurements? E.g. in terms of power levels used, IF BW, number of harmonics included in the extraction. Anything that you can share would be helpful. 

Many thanks for your time.  

Outcomes