[time-nuts] time-nuts Question concerning ADEV measurements

Mark Spencer mspencer12345 at yahoo.ca
Sat Oct 12 13:29:07 EDT 2013


Hi Bob,  All of my actively used HP5370B's exhibit behavior similar to yours.   I routinely use a BVA oscillator as the clock for most of my HP5370B's.   I've also done some testing comparing the reported adev of other signals using various references and swapping between the  internal oscillator and the BVA does seem to make a small difference.   

 That being said looking at linear residual plots I have occasionally seen some odd results (probably stair case type plots with a recurring pattern iirc sorry I am on the road right now and can't check my files re this) on occasion when using the clock source as either the start or stop signal.   One day when I have some time I will spend a day or so going thru the various use cases systematically on all of my active 5370B's and will hopefully get to the bottom of this.  

As a side note at least with my HP5370B's I need to be very meticulous about cable types, signal levels, trigger settings etc. to be able to notice the small difference in performance I get by using the BVA as a clock source.  

Best regards Mark S

Sent from my iPad

On 2013-10-12, at 11:41 AM, time-nuts-request at febo.com wrote:

> ------------------------------
> 
> Message: 2
> Date: Fri, 11 Oct 2013 22:27:26 -0400
> From: Robert Darby <bobdarby at triad.rr.com>
> To: time-nuts at febo.com
> Subject: Re: [time-nuts] Question concerning ADEV measurements
> Message-ID: <5258B38E.8030303 at triad.rr.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> Thanks John and Magnus.
> 
> The reason I asked was that when I check the noise floor of my one and 
> only 5370, the result is noticably superior using the external clock.  
> Since the signal is coming from a single source fed through a MC ZFSC-2 
> splitter and unequal lengths of co-ax, variability in the signal source 
> should have no effect.  When I measure the 5370 with an external clock I 
> see 2.24E-11 at t0=1s with a straight -1 slope; with the internal clock 
> I see 4.43e-11 with periodic noise at about 11, 22, .... seconds.  This 
> suggests to me that the internal oscillator is not so hot.
> 
> Having only one 5065 I want to use it to both clock the 5370 and serve 
> as the reference; hence the question.  It seems to me by applying the 
> DUT to the start channel one avoids the issue of hitting the same range 
> of start values, assuming some phase change relative to the DUT.
> 
> Bob Darby
> 
> 
> On 10/11/2013 5:36 PM, Magnus Danielson wrote:
>> It can be a benefit or a cause of problem. In one end, it could lower
>> your noise. The DG535 actually has that as a recommendation, that the
>> signal source and the DG535 operate on the same clock.
>> 
>> Another aspect is that you keep banging the same little range of values,
>> and that may for that time-base to stop delay be systematically off in
>> an unfavourable fashion. The noise of a counter is not always linear
>> with that delay, so an asynchronous clock would sweep over the range.
>> Then again, if the delay is just right you can utilize the minimum.
>> 
>> Cheers,
>> Magnus
>> 
>> 
> 


More information about the time-nuts mailing list