[time-nuts] Equipment question: OCXO versus GPSDO + XTAL

Bruce Griffiths bruce.griffiths at xtra.co.nz
Sat Mar 19 20:23:47 UTC 2011


Since the 53320A TCXO option has a 1s ADEV spec of 1E-9 whereas the OCXO 
option has a 1s ADEV spec of 1E-11 and the PLL bandwidth is unlikely (no 
info on the PLL bandwidth available from datasheet) to be very large the 
OCXO contribution to measurement uncertainty will likely be 
significantly larger than that of the OCXO even when phase locked to an 
external standard particularly for short gate times (less than the 
reciprocal of the PLL bandwidth i.e. probably below a few tens of 
millisec). For sufficiently long gate times (> 1s??) their will likely 
be no discernable difference between the the OCXO and the TX) options 
when locked to a sufficiently quiet external frequency standard (ie one 
with an ADEV << 1E-11 or so for Tau > 1s).

Locking the counter timebase to an external standard isnt as flexible as 
a direct input where one could use something exotic like a low noise 
OCXO simultaneously locked to a hydrogen maser and a Caesium frequency 
standard (or GPSDO) to achieve low ADEV over a wide range of Tau values. 
Implementing such a system doesnt necessarily require that the hydrogen 
maser be phase locked to the Caesium standard (or GPSDO).

In some cases trigger jitter may dominate as it isn't easy (but not 
impossible for low noise sources: 
http://www.holzworth.com/Spec_sheets/HX2410_Web_Datasheet.pdf ) to 
produce subpicosec trigger jitter when converting from sine wave inputs 
to logic level signals.

Bruce


Bob Camp wrote:
> Hi
>
> Alan variance / short term stability / medium term stability is what will impact your measurements "noise". That's not to surprising - noise on the standard gives you noise on the measurement.
>
> If for some reason an internal standard is more quiet than an external standard, the internal may give you "better" measurements. This is not a so much function of the oscillator, but of the PLL used in the instrument.  If the PLL changes between your choice A and choice B, that may impact the result, if there is much noise on your external standard.
>
> As you measurement interval gets shorter, the noise level of all sources will go up. Fortunately, the noise tolerance of the measurement does the same thing. Depending on exactly which specific oscillators are being compared (not just types, and possibly not just models) there may be a cross over between devices. That could also give you a cross over between choice A and choice B.
>
> All that said, normally you are better off with the highest priced standard. It's going to have the lowest intrinsic noise. It also should  be locked through the narrowest PLL. That's a most of the measurements most of the time judgement.
>
> Bob
>
>
> On Mar 18, 2011, at 10:18 PM, Wolfgang Wieser wrote:
>
>    
>> Hello all,
>>
>> first, thanks for all the answers!
>>
>> Maybe let me ask the more fundamental question: Completely independent
>> of use scenarios and so on: Assume device A uses standard reference tied
>> to ultra-stable zero-drift magic 10 MHz refclock. Device B uses TXCO option.
>>
>> Is there any measurement type with any strange setup that you could imagine
>> where B outperforms A just because the combo "standard XO + refclock"
>> is inferior in any respect compared to "TXCO". (Lab/shop equipment,
>> no portability needed.)
>>
>> I mean - not necessarily for this particular Agilent 53230A but for anything
>> of this type and class.
>>
>> @Bill:
>> thanks - exactly the type of advice/opinion I was hoping to get.
>>
>> @Paul Swed:
>>      
>>> Unfortunately in the first round of the thread you did not say what you are
>>> actually doing with the system and whats important to you.
>>>        
>> Measuring time and frequency? Maybe there are some unknown issues that I am
>> not aware of but what I would like to know is whether for any usual
>> measurement made with a timer/freq counter the choice of XO+refclock could
>> be inferior to TXCO. If there are significantly different scenarios in
>> your mind - it would be nice to let me know for what type of measurement
>> you think that the one is better and for what not.
>> Anything except portability issues?
>>
>>      
>>> The last answers actually the best considering the base system initial
>>> investment.
>>>        
>> Well, once a year I buy something more expensive...
>> However, while 2500 EUR seems reasonable for the counter, additional
>> 1000 EUR seems a bit much for the OCXO option.
>>
>> @Joe
>>      
>>> You will have to look at the stability measurements of the OCXO versus the
>>> GPSDO to choose that one.
>>>        
>> OK, do you have an educated guess about the outcome of that?
>> Because acutally doing this comparison exceeds my scope ATM.
>>
>> @Greg
>>      
>>> My opinion is that first you need to tell more about how you
>>> are going to use this counter. For work in the shop where
>>> a distributed reference is always available (...)
>>>
>>>        
>> Hm.. OK.. I re-formulated into the question on the top.
>>
>>      
>>> I am suspicious of the note that this OCXO upgrade can only
>>> be done at the factory.  See if you can get the control sequence
>>> to cause reference autocal.
>>>        
>> I've got a test unit ATM and there actually is a security code to
>> unlock the calibration. I don't know the code right now.
>>
>> - Wolfgang, DL1SKY
>>
>> _______________________________________________
>> 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.
>>      
>
> _______________________________________________
> 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