[time-nuts] DMTD Question1
bruce.griffiths at xtra.co.nz
Tue Jan 26 23:48:56 UTC 2010
The system noise floor should have been ~<1E-13/Tau for tau < 1000 sec
if one wishes to replicate the performance of the Early (1976?) NIST
DMTD design where wite phase noise dominates for tau up to 1000 sec or so.
Bruce Griffiths wrote:
> Knowing the magnitude of various effects allows one to make rational
> choices based on one's actual or perceived requirements.
> Just ignoring effects which may or may not be significant and hoping
> that they wont be significant isn't particularly rational.
> The aim was a system noise floor of around 1E-13/SQRT(Tau) (for Tau up
> to a few hundred seconds at least) or better to ensure that this is
> well below the expected stability of sources being compared.
> There are also other uses for DMTD systems such as measuring phase
> instability of cables etc, that some may wish to pursue.
> NIST found a mixer phase shift TC of up to 75ps/K for uncompensated
> mixers with 5MHz inputs.
> If you carefully analyse your favourite frequency comparison system
> you may just discover why this particular technique was quietly
> dropped as not being particularly useful for Tau much greater than the
> PLL inverse loop bandwidth.
> The real problem is that there is no simple way to actually determine
> the actual Tau range for which such a system is useful other than
> comparing the results obtained with those from a more conventional
> WarrenS wrote:
>> OK, Now what am I missing?
>> I understand there are 'Freq nuts' and just 'plane NUTS', this sound
>> like the later.
>> With a little effort, not hard to hold the temp drift rate change to
>> 1deg / hr (and a couple deg change altogether).
>> Lets say the Tempco balance of the mixers was as bad a 1ns / deg
>> That means temp change would add an addition freq error offset term
>> of half of 3e-13 at tau of 3600 sec (36ns/hr = 1e-11)
>> Even this gross amount would not effect the results of most of the
>> things that any of these freq nuts would be testing including
>> probable the best CS or RB Osc.
>> AND if they are doing 1e-15 things, they probable want to be hold the
>> temp to better than 0.01deg for many other reasons as well.
>> So why even talk about TC phase shifts at the ps / deg level? Who
>> cares and who needs it?
>> Does not sound like you are doing the Nuts any service to address
>> things that are orders of Magnitude below usefulness.
>> ----- Original Message ----- From: "Bruce Griffiths"
>> <bruce.griffiths at xtra.co.nz>
>> To: "Discussion of precise time and frequency measurement"
>> <time-nuts at febo.com>
>> Sent: Tuesday, January 26, 2010 1:51 AM
>> Subject: Re: [time-nuts] DMTD Question
>>> If one carefully the value of a resistor in series with both the
>>> mixer RF and LO ports and adds a 6dB attenuator between the series
>>> resistor and the output of the associated isolation amplifier its
>>> possible to reduce the mixer phase shift tempco by a factor of 10 or
>>> so. The resistor value is chosen to minimise the mixer port VSWR for
>>> the given IF port termination.
>>> At 5Mhz this can reduce the mixer phase shift tempco to less than
>>> The mixer phase shift tempco also reduces as the input frequency
>>> increases (at least up to 100Mhz or so) by a factor equal to the
>>> ratio of the frequencies for which the phase shift tempco is being
>>> Thus carefully matching the mixer inputs can reduce the required
>>> mixer temperature stability.
>>> It is believed that the matching reduces the effect of reflections
>>> in the input cables, so a similar effect may be achieved by using an
>>> isolation amplifier located close to each mixer input port.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> and follow the instructions there.
More information about the time-nuts