# Measurements using spectrum analyzer

I need to measure the following characteristics of a frequency synthesizer using a spectrum analyzer:

• Frequency range: 2.286 GHz to 2.346 GHz

• Output power: -2.5 dBm to 2.5 dBm

• SSB phase noise: typical: -122 dBc/Hz, maximum: -116 dBc/Hz, at 100 kHz offset

• Integrated SSB phase noise: -45 dBc, at 100 Hz to 1MHz

• Reference spurious suppression: typical: -106 dBc, maximum: -80 dBc, with reference frequency = 15 MHz

• Non-harmonic spurious suppression: -90 dBc

• Harmonic suppression: typical: -30 dBc, maximum: -20 dBc

Am I able to measure all those specifications using a Rohde and Schwarz FSC6 model .06: datasheet ?

I can choose any spectrum snalyzer in the world, but I was told that Rohde and Schwarz is the one of the best brands, and that's the one I found more useful for the specifications I need to measure, because of its frequency range and phase noise (I don't know if I'm reading it correctly). What do you think?

• Will the R&S provide the integrated phase noise measurement? Or must you export the phase-noise, and perform the integration? Commented Dec 12, 2018 at 4:44
• With -116dBc/rtHz at 100KHz, if we assume flat DC to 1MHz, then at 3dB increase per doubling of the bandwidth (-113dBc at 2Hz BW, -110 at 4Hz, -107 at 8Hz, -107 + 30 = -77dBc for 8,000 Hz BW, -77 + 21 = -56dBc at 1MHz integrated bandwidth, you've got some margin for high levels of 1/F closein (and 1/F^2). Commented Dec 12, 2018 at 4:48
• I don't know how to do the integrated phase noise measurement, but I still didn't get there. I supposed that there would be an option in the S.A. that says "Integrated Phase Noise" or something like that :S. What do you think about the Signal Analyzer E5052B? I think that instrument can measure the phase noise and, hopefully, the integrated phase noise Commented Dec 12, 2018 at 5:58
• Would you please stop saying that you have "measure" things? Your assignment is to write up the procedures for measuring these things. If you would just say thay, then you could stop having to explain that you can "buy any piece of equipment no matter how expensive." Just for pete's sake say that you writing up the procedures, and can base them on any piece of equipment available.
– JRE
Commented Dec 12, 2018 at 6:22
• You also haven't voted for or accepted the answers to all of your other questions on this subject.
– JRE
Commented Dec 12, 2018 at 6:24

## 1 Answer

The phase noise of the spectrum analyser is -100 dBc worst case, -110 dBc typical. This is (hopefully) worse than the synthesiser you are trying to measure. So no, you won't have much luck seeing your synth phase noise above that of the S/A local oscillator. This is not unusual. Spectrum analysers often have worse phase noise than synthesisers. I haven't needed to do this test for a while so I'm not up to date with the test equipment available. One technique I have used is to phase-lock a low phase noise signal generator to the synth output. A ring mixer is used as a phase comparator, which results in the signal generator and synth being in quadrature. The phase noise sidebands then appear at the mixer output at baseband. That is, phase noise at 100 kHz offset appears at 100 kHz at the mixer output, both sidebands superimposed. Your signal generator needs to have FM modulation capability down to DC in order to lock. You may also be able to infer phase noise performance in your final system by measuring reciprocal mixing performance. You may have trouble seeing -106 dBc ref spurs in the noise.

• Thanks! But I didn’t understand all the procedure that you do to measure the phase noise :(. Do you have any place where I can read it with more detail or images (I know you described it very well but I don’t have too much expertise with these instruments). What do you think about the Signal Analyzer E5052B by Keysight? Commented Dec 12, 2018 at 1:12
• I'm glad the answer was of use. I'd appreciate you marking it as such. Here is a link to a presentation describing the phase detector method. google.com/… Commented Dec 12, 2018 at 1:32
• The signal analyser you mention looks like a nice piece of kit at first glance, if you can afford it. I don't have the time to look at it in detail so you should satisfy yourself as to its suitability. Commented Dec 12, 2018 at 1:34
• I can spend all the money in the world because I have to make the measurements in paper, not experimentally. Is the method you mentioned In the keysight link you posted? Commented Dec 12, 2018 at 2:46
• Yes, it starts page 31. Commented Dec 12, 2018 at 5:42