ACWWong wrote on Aug 11th, 2008, 12:30am:I'm assuming your simulation gives -100dBc/Hz
In such a case your measurement shows about -95dBc/Hz at 1MHz offset.
That is -60dBc in a 3kHz RBW, so you need to adjust the marker by 10*logRBW=34.8 dB.
Often the analyser has the marker on noise function which will take into account the RBW automatically (and more accurately).
So in fact you are 5 to 6 dB of where you expect to be. The reasons for this can be many e.g. discrepenacy in models, simulation setup issues, measurement issues, PCB design, noise pickup etc.....
Great.
from the test picture, we can find some details:
1. RBW is 3kHz, so we should normilized the PN to dBm/Hz.
2. the Spectrum analiyzer power dector is peak dector , you should change the power dector to rms dector.
2.