Sunday, July 31, 2016

Boys with Toys 2

My last posting was about using my RIGOL Spectrum Analyzer with frequency range up to 1.5GHz to show characteristic of a 10GHz filter. That little test made me wonder if I can use the same Spectrum Analyzer to display a signal in the 10GHz range on that analyzer. Obviously, I have to mix the signal down below 1.5GHz using a transposing signal from another synthesizer.

I have used my 24GHz transverter with 11.88GHz local oscillator, because that signal is conveniently fed to a connector on the transverter's front panel. The same mixer from my previous test (IF range only 0-150MHz) was used for the first test today:

Yes! It really works! My ELCOM on the left side is programmed to 10.805GHz (+13dBm) and the gray transverter on the right side delivers 11.88GHz (+10 dBm). Difference of these two signals is exactly 1.075GHz which is what we see on the analyzer. What we also see is the signal level below -40dBm. Hmm ...

I have tried to "zoom in" and changed the RBW ratio on my RIGOL from 1MHz to 300kHz. Signal still looks nice and clean, just not too strong:

Can we do something about the attenuation? Of course we can. I tried 3 things:
  • Find better flexible cables
  • Replace flexible cables by Semirigids
  • Use different mixer
Luckily, I have bought good sized bundle of these flexible black cables with SMA connectors cheap on eBay. Maybe I could find some better pieces in that bundle? Tried and ... the signal strength improved +30dB:

Wow! This is almost perfect! Maybe Semirigids will make it even better:

Hmmm, not too much. Only about 2dB. Surprising, really. These black cables are RG-178 which is OK but not as good as semirigid, at least in theory. Whatever ... let's try another mixer:

Nice! Improvement about 8dB! This mixer has IF range 0-2GHz, but it has apparently been exposed to salty water somewhere. I have bought it cheap on eBay. Lets try better one:

No surprise here, less than 2dB below zero. This mixer behaves as expected. Note that I am feeding it LO signal 11.88GHz about +10dBm and the test signal is 10.8GHz about +13dBm. My RIGOL has internal attenuator of 10dB, so the conversion loss in this case is 13 - (10-2) = about 5dB. That is certainly very acceptable!

My next test will be with a Harmonic Mixer (same thing, but the mixer will generate harmonics from a strong base signal to be able to display higher frequencies, such as 24GHz). But that will be the topic for some other day.

73 Herbert


  1. Thanks ... this seems to be "hot topic", judging by the number of responses and questions I am getting. I will try to continue with more difficult measurements, such as Noise Figures etc. I have never done that with Spectrum Analyzer myself, so it's learning curve for me too.