[time-nuts] Distribution amplifier phase noise

Tom Knox actast at hotmail.com
Thu Mar 14 16:00:45 EDT 2013


I think designing a good distribution amp is much more difficult then it would seem at first glance. But placing it in a system is the real challenge. If you have a ULN 5MHz oscillator that has phase noise of -125dB @ 1Hz and a noise floor of 167dB @10KHz it will take some real attention to detail from cable to amp design to get that signal to it's destination. I have done some past measurements and for BVA's and top of the line Wenzels, few distribution amps are up to the task. Symmetricom has data sheets on their amps and they vary widely close in. the 4036B seems like the was to go from them. The PTS amp also looks exceptional. I think Quartzlock also makes a great dist amp.  I am sure there are others. I am currently setting up a new multiple reference system and seeing all kinds of artifacts. Channel isolation and cable interaction can really affect signal quality even higher end Pomona act like soaker cables.   I think it is as much art as science where things like equipment grounding and isolated or unisolated patch panels can reduce or contribute to noise depending on very minor variables. And what effect does having the equipment net worked have? Are there any basic rules of thumb as far as dist amp set-up?

Thomas Knox



> Date: Fri, 15 Mar 2013 07:02:44 +1300
> From: bruce.griffiths at xtra.co.nz
> To: time-nuts at febo.com
> Subject: Re: [time-nuts] Distribution amplifier phase noise
> 
> Ulrich Bangert wrote:
> > Gentlemen,
> >
> > although it has been discussed here by Bruce et al how the noise floor of
> > integrated distribution amplifiers designs may compare to discrete designs I
> > have never seen a *real* phase noise measurement of a *real* amplifier here.
> >
> > My own measurements on a DIY AD8007 based distribution amplifier indicate a
> > noise floor of abt. -160 dBc @ 100 kHz which may not meet the specs of the
> > absolutely top line of oscillators by some 15 dB but is well in reach of
> > Bruce's predictions for integrated designs.
> >
> > Did any of you Timepod owners make measurements of your own on amplifier
> > noise? Please note that I am not only out for the noise floor specs. Instead
> > I am specially interested in specs for power supply induced spurs.
> >
> > I just received an article covering ultra low noise oscillators to be found
> > here:
> >
> > http://www.crystek.com/documents/appnotes/ImpactUltralow.pdf
> >    
> This paper is very sloppily written. e.g.:
> -174dBm/Hz is the available thermal noise power of any resistor at 290K 
> not just 1 ohm.
> There is no explicit consideration of the contribution of thermal noise 
> power to phase noise.
> 
> > Clearly the phase noise diagram shows a power induced spur @ -100 dBc. Is
> > this the measure for the amplifier to take care for? I.e. if the amplifiers
> > has spurs down -120 dBc, is that good enough?
> >
> > Or to put it another way: What is the best phase noise that we can expect
> > from an amplifier concerned power supply induced spurs? Can it be that the
> > spurs to be seen are not a problem of the amplifier itself but are inherent
> > to the measurement setup? If so, is there a recommended setup to minimize
> > such problems?
> >
> >    
> I routinely achieve mains related PN spurs below -150dBc and often below 
> -170dBc.
> Such spurs can be enhanced by low frequency ground loops and magnetic 
> fields.
> Using RF isolation transformers with capacitive grounding of the 
> isolated winding helps considerably in suppressing mains related ground 
> currents.
> > Thanks in advance for your suggestions. I appreciate the expertise in this
> > group a lot!
> >
> > Ulrich Bangert
> > www.ulrich-bangert.de
> > Ortholzer Weg 1
> > 27243 Gross Ippener
> >
> > _______________________________________________
> > time-nuts mailing list -- time-nuts at febo.com
> > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> > and follow the instructions there.
> >
> >    
> Bruce
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
 		 	   		  


More information about the time-nuts mailing list