Guys,
While you wait for the USRP files here's some background info using the SDR-IQ:
The screen shot above is the IF output of the 70 cm Receiver with an RF input of -60 dBm. FFT averaging is set to 100 and the span is 200 Hz.
I then matched the level of the Receiver's IF output using a step attenuator and fed a signal from a battery-powered HP reference oscillator I salvaged from an old signal generator into the SDR-IQ. The SDR-IQ settings remain exactly the same. You can see the difference.
Since the reference oscillator runs on batteries I was able to take it to work last month and test the phase noise using a hugely expensive Agilent E5500 phase noise measurement subsystem. The shot above is the results of the phase noise test on the oscillator so you can correlate the SDR-IQ against the Agilent. So for example, the battery-powered oscillator phase noise is down -130 dBc/Hz at an offset of 100 Hz. The SDR-IQ says it's down about -115 dBc. That's not bad since we're comparing a $400 box against something that probably cost over $100,000. If we assume that adding another -15 dB to the SDR-IQ brings it into line with the Agilent then the 70 cm Receiver has a phase noise of about -85 dBc/Hz at 200 Hz offset. I'll be curious to see what the USRP has to say about this.
There are some external sources of phase noise that need to be considered - The SDR-IQ has its own issues of course, and the Receiver was using my GPS disciplined clock as an external reference. The clock is very clean but nothing is perfect. I can't use the internal reference since it jumps all over the place and will be changed out in the next rev anyway.
IMPORTANT - Last night I think I noticed that the receiver phase noise was greatly affected by the quality of the external reference. If this is true then we need to consider how that external reference will be generated and distributed on Eagle and what sort of quality we can expect - or the Receiver needs to be designed so it can function with a dirty external reference. I'll look into this as time permits. We hope to get the USRP over here this weekend and rerun these tests with it.
73,
Juan
