[Fwd: LF: Frequency calibrations]

Andre' Kesteloot akestelo@bellatlantic.net
Thu, 12 Nov 1998 07:48:41 -0500

drassew2@interalpha.co.uk wrote:

> >From G4JNT
> I have been following the recent discussions on frequency calibration and
> measurment from VK2ZTO et. al. with interest.  This is something we have
> been very concerned with on the microwave bands recently and most operators
> now regularly achieve better than 1 ppm from home built equipment,
> calibrated against off air sources.  That is 10 kHz frequency error on the
> 10 GHz band and signals still usually have to be searched for - especially
> very weak CW ones!  The new 47GHz SSB / CW operators are looking for better
> than 10^-7 stability from portable equipment - and they are beginning to
> manage it.
> The hardware used is typically :
> Series resonant butler oscillator in the 100 MHz region, using a simple clip
> on heater on the crystal to keep it at around 35 deg C; These devices cost
> around 1 to 2 pounds each.  1 ppm is typical for this combination, and the
> same stability is probably just as achievable from a lower frequency
> crystal.  If such an oscillator is left switched on as a calibration or
> reference source  in a constant temperature environment (such as a house),
> 0.1ppm is possible.
> Off air frequency standards, such as ones based around MSF as published in
> RadCom a few years ago are commonly employed.  Another option is to use TV
> line sync pulses.  The BBC maintain their's using a Caesium standard most of
> the time;  ITV and Channel 4 both use Rubidium standards - this fact was
> confirmed to me, separately, by three employees of the various companies.
> However, I have observed a few times when there is an abrupt phase shift on
> the ITV network which could be due to a change in programming where local /
> regional network switching occurs.  Channel 4, being national, was
> recommended as being one of the better ones to use as the ITV networks
> employ many separate Rubidium standards at each of their centres.
> Microwave beacons coming on air are encouraged to use as good a drive source
> as possible - two of the recent ones use oscillators locked to oven
> controlled oscillators (available from rallies / surplus for 10 pounds each)
> giving a few parts in 10^-8 stability.  They therefore form a good
> calibration source for portable operation, or for operators with no other
> means of generating an accurate standard
> In Germany, TV sync is used as a means of propagating a time / frequency
> standard over the country and this extends to their satellite TV broadcasts;
> so anyone with satellite TV has access to GUARANTEED certifibale frequency
> accuracy.  This practice was originally started in the old DDR as a means of
> low cost  time / frequency distribution,  and extended over the whole
> country after unification.  (DO any DL operators on this group have further
> information ?)
> The Droitwich transmission on 198kHz is Caesium controlled but has an
> annoying 45 degrees PSK transmission on the carrier which makes a simple
> standard difficult to make; I believe (although unconfirmed) that this is an
> National Physical Laboratory controlled standard.  The same applies to the
> Allouis LW station on 162kHz.   Other BBC transmissions on MW appear to be
> at least Rubidium locked - at least based on my observations - but the same
> is definitely not true of the commercial MW stations;  one local station has
> been observed 2ppm off frequency!
> TCXOs as used in the mobile phone industry are two-a-penny at the right
> rallies and radio bootsales now - I was given ten of these for free, but the
> more normal price is from 50p to a couple of pounds or so each.  These tiny
> postage stamp sived will give 2ppm guaranteed, and frequently 1 ppm accuarcy
> from turn on. Frequency is usually 9.6MHz or 13 MHz although others do exist.
> So there is no excuse for anyone not being able to measure frequency quite
> accurately - all that is needed is the willingness to look around at rallies
> and build a bit of hardware of whatever complexity is felt necessary for the
> accuracies wanted.
> Incidently - I use a Caesium standard, found in a skip a few years ago and
> fully operational.  There are at least two other amateurs in the UK with Cs
> standards, all thrown away because they were 'officially' at end of life but
> with plenty of operational time left.  We only tend to use the Caesium tube
> part for a few hours or so every few months (prolonging its life - they do
> wear out) to calibrate the internal crystal oscillator.  This part alone
> free runs at  7 * 10^-10 accuracy in its own right  - within the
> specification for a Rubidium source - and the seconds ticks are always
> within 2ms when compared with MSF each time I recalibrate at 2 - 3 monthly
> intervals.
> The 5 MHz output from this is piped around the shack and used for the RA1792
> receiver, a homebrew SSB driver for HF, frequency counter, clock and (soon)
> the 10 GHz transverter local oscillator
> Andy  G4JNT / G8IMR