[time-nuts] Update -- Comparing 10 MHz Oscillators at 10 GHz

Brucekareen at aol.com Brucekareen at aol.com
Thu Jan 27 00:20:42 UTC 2011

Thanks to some help from Joe Ruggieri, I finally managed to get a pair of  
10.000 GHz microwave "brick" oscillators phase-locked to 100 MHz inputs.   
These require an input of about 0 dBm at 100 MHz and have an output of about 
+  13 dBm at 10 GHz.  To gauge whether or not loop noise would interfere 
with  using these to compare 10 MHz oscillators, I assembled the following test 
 setup as I have not yet constructed the required 10 MHz to 100  MHz 
Both 10 GHz microwave bricks were driven through a splitter from an HP  
signal generator locked to 100 MHz.  The 10 GHz output of one brick was  routed 
through an HP X885A phase-shifter equipped with WG/coax adapters and on  to 
one input of an Anaren hybrid combiner.  The output of the other brick  was 
routed through a 0-10 dB SMA attenuator then on to the combiner.  The  
reject side of the combiner was terminated.  The main combiner output was  
routed through a 10 dB attenuator to a Systron-Donner coaxial detector.   The DC 
output of the detector was monitored with a digital voltmeter.
First the phase shifter and attenuator were adjusted to cause the two  
sources to arrive at the combiner with equal amplitudes and 180 degrees out of  
phase  and produce a null at the detector output.  I measured as  little as 
0.01 millivolt on the DVM at the null point.  Then the  phase-shifter was 
adjusted to rotate the phase 180 degrees to place the  combiner inputs in 
phase.  The detector output was about 128  millivolts.  
This experiment seems to confirm the earlier results by Ulrich Bangert and  
his friend that noise does not appear to be a problem with this method for  
rapid frequency comparison.  However it would be convenient to have an  
attenuator and phase shifter to null the sources.    
Bruce Hunter, KG6OJI.  

More information about the time-nuts mailing list