[time-nuts] Thunderbolt spurs on 10MHz output at 100Hz and 200Hz from signal.

[Dan Rae]

On 9/21/2016 8:01 AM, David C. Partridge wrote:
> Looks like all I've managed to do is measure the baseline noise of the analyser.  Ho-hum looks like I need something a *lot* quieter to do these measurements.
>
Indeed Dave, yes.  I made the same mistake a few years ago when I tried 
to compare the results of three different DDS clocks on the output of my 
homebrew transceiver DDS LO using a 3585A spectrum analyser and John's 
program.  It took me a while to figure out why they were identical and 
then the penny dropped that I was just measuring the 3585A...

The plot of the three thunderbolts I sent before using my -hp- phase 
noise system is interesting in that the white plot was from the Tbolt I 
keep on the bench to run all the stuff there, and it uses the switch 
mode power supply that came with the TAPR ones, built into an Extron 
distribution amp case.  All the spurs that are shown with dotted lines 
aren't all spurious, most are probably real.  The other two are from 
units with linear supplies and are significantly improved.

Most older phase noise systems use some kind of low noise oscillator for 
reference, mixing the signal under test down to AF and going from there 
with low noise amps and then the actual measurement. Wenzel has a very 
good application note describing a system that can be copied in the home 
lab.

I attach a plot of my -hp- 3048A system noise floor showing what levels 
can be achieved with twenty year old technology.  The spurs the system 
shows dotted may not be real.  The modern system John Miles designed can 
no doubt do better and in about one hundredth of the volume, but is a 
little more expensive :^)

Dan





-------------- next part --------------
A non-text attachment was scrubbed...
Name: noisefloor.png
Type: image/png
Size: 54927 bytes
Desc: not available
URL: <http://www.febo.com/pipermail/time-nuts/attachments/20160921/fb4a2474/attachment-0001.png>