I want to know whether NQS model is included in TSMC.18 model when NF is simulated in Cadence SpectreRF.
NF I postlayout- simulated is about 1.3 dB in consideration of
LNA's matching network and pad parasitics , package and corner fluctuation at room temperature. but, after testing, i found NF is about 1.9 dB at same current consumption, good matching and same gain. i can't find the reason noise figure
degraded. so i'm doubted its noise model.
0.2 to 0.5 dB from sim to meas is a typ degradation, commonlyfound also by good LNA designers.
To give you an answer a lot of details should be known. The NQS model could be one. Other can be: layout parasitic extraction, noise from bias/supply, model of passives (in particular inductors, if you use it), good PCB or fixture gnd and so on.
I hope it can help.
Mazz