[time-nuts] Is it Hz or MHz ?

Tom Van Baak tvb at LeapSecond.com
Mon Jun 2 05:28:28 EDT 2014


Hi Kit,

For convenience, accuracy is often given as a relative, unit-less value. Think of values like 1%, or 20 ppm, or 5e-11. These are all are unit-less numbers. To translate to absolute units just multiply:

50 Hz times 1% is 0.5 Hz.
1 meter times 20 ppm is 20 microns.
10 MHz times 5e-11 is 500 uHz.
5 MHz times 5e-11 is 250 uHz.
1 MHz times 5e-11 is 50 uHz.
1PPS times 5e-11 is 50 ps.

Two issues arise with GPSDO specs:

1) When you talk about frequency, you have to specify the averaging interval(s). Performance can vary considerably depending on interval. It sounds your data sheet does this correctly.

2a) When you talk about time, there are often two meanings. One is relative time, as in pulse to pulse interval accuracy or stability. This is almost always the spec used in the commodity GPS receivers we play with. It's the only spec that matters if the receiver is used for a GPSDO.

2b) The other meaning is absolute time, as in conformance with UTC(k), where k is your local National Measurement Institute (NMI). This is a much more difficult problem, since it requires independent antenna + cable + receiver + TIC calibration against some portable "gold" standard, and likely re-calibrations every year.

/tvb

----- Original Message ----- 
From: "Kit Scally" <kitscally at iprimus.com.au>
To: <time-nuts at febo.com>
Sent: Monday, June 02, 2014 12:28 AM
Subject: [time-nuts] Is it Hz or MHz ?


> Team Nutters,
> 
> 
> GPSDO's are often specified for stability as so many parts per x.  Using a
> Symmetricom 8040C as an example, this has a quoted accuracy of typically
> 5E^-11 at shipment.  Ageing (monthly and yearly) and ADEV stability rates
> for 1, 10 & 100 seconds are also given. This particular device as 1, 5 and
> 10 MHz outputs.
> 
> The question - is this "accuracy" with respect to 1Hz - or 10MHz ?  If the
> latter, then the accuracy figures quoted needs to be modified by 10^6 (Hz to
> MHz) to get a feel for the actual frequency changes at 10MHz.
> 
> Glossies of many GPSDO manufacturers seem deathly silent on this point, so
> clearly, "something is assumed".  I have seen an oscillator comparison chart
> by Meinberg (a German company) that in their comparison chart, the accuracy
> references need scaling to relate to 1Hz.
> 
> Appreciate some advice and/or direction on this before madness overtakes me
> !
> 
> 
> Kit 
> Canberra
> Australia




More information about the time-nuts mailing list