The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> How to simulate large signal NF in cadence?
https://designers-guide.org/forum/YaBB.pl?num=1146828924

Message started by ZYC on May 5^th, 2006, 4:35am

Title: How to simulate large signal NF in cadence?
Post by ZYC on May 5^th, 2006, 4:35am

Hi, all:

Now I am doing simulation for large signal NF of amplifier with cadence.
For example, there are two signals, Frf and Frf+∇F. The power of Frf is -60dBm and -20dBm for Frf+∇F. How could I simulate the NF for Frf?
I tried to use QPSS and QPnoise to do the simulation. Some setups made me confused.
1) Which signal should be set to the large signal, Frf or Frf+∇F?
2) How to set the reference side band? (0 0) or (1 0)?
Anybody know how to do this simulation? All suggestions are appreciated very much.

ZYC

Title: Re: How to simulate large signal NF in cadence?
Post by ACWWong on May 5^th, 2006, 9:00am

I also always forget which reference sideband i want. Often I revert to doing qpxf and fiddling with output frequency axis to absin.
Anyway, I do recall the spectreRF manual (in the cadence installation path) does have an example of exactly hwat you are trying to do.

cheers

aw

Title: Re: How to simulate large signal NF in cadence?
Post by Andrew Beckett on May 15^th, 2006, 1:38am

In this case the reference side band is most likely to be (0 0) because if it's an amplifier, the transfer function that you're interested in for the purposes of calculating the noise figure is that the output frequency is at the same frequency as the input. In general the reference sideband determines the input frequency given the output frequency sweep - i.e. for a specified output frequency, fin=fout+k1*f1+k2*f2 where f1 and f2 are the "large/moderate" signal frequencies, and k1 and k2 are the number of shifts to be applied.

Recent versions of DFII give you a means of selecting the reference side band from a list - so you can more easily see the frequency range. That said, you still need to ensure you've got the right transfer function, so simulating with qpxf is often useful anyway to ensure you've got a reasonable transfer function.

With an amplifier, the two large signals are effectively blockers which are not needed for correct operation of the amplifier - they're presumably only there because you want to see the effects of compression and/or intermodulation when doing the noise analysis.

Regards,

Andrew.