[time-nuts] UK standard frequencies - where?

David Bobbett d.bobbett at tiscali.co.uk
Wed Oct 12 18:33:55 UTC 2011

The problem with all of these comparative techniques is that you are 
always dealing with relative error vs finite error. However the result 
is usually well in excess of that needed for ham ops.

Another trick is to load one memory with a 1kHz RWM offset in USB and 
another with a 1kHz offset in LSB so as to generate a tone for each 
sideband. You then feed the AF into SpectrumLab and repeatedly toggle 
between the two memories whilst adjusting the LO.

If you adjust the LO so that the difference between the LSB and USB 
audio tone is at a minimum, that will give you the most accurate 
setting. Any errors in the sound card will cancel out because you are 
comparing the /difference /between the two 1kHz signals, any remaining 
errors will be down to variables within the radio.

Hope that helps,

73 de David, G4IRQ

On 12/10/2011 13:54, David J Taylor wrote:
>> David,
>> I'm 70km north of London and have used the French 162kHz high 
>> stability signal as well as RWM (Moscow) on 4.996, 9.996 and 
>> 14.996MHz for frequency measurement and calibration before I got my 
>> Thunderbolt. RWM is particularly good because part of the schedule 
>> involves sending continuous carrier, which I used with SpectrumLab to 
>> calibrate transceivers - you simply use SSB, offset the transceiver 
>> by 1kHz to get an audio tone and measure the error using the 
>> waterfall on SPLab.
> []
>> Regards,
>> David, Milton Keynes, UK  (G4IRQ)
> David,
> This was an excellent suggestion!  Briefly, I have an audio oscillator 
> with a built-in counter where I could generate 1 KHz +/- 1Hz, and I 
> compared this against a computer generated 1 KHz tone using my 'scope.
>  http://www.satsignal.eu/software/audio.html#SweepGen
> I then installed Spectrum Lab (I had a very old version but wasn't 
> using it), and found that the 1 KHz displayed most close to 1 KHz when 
> the 12000 or 48000 sampling frequencies were selected.  Then found 
> that the RX was 60 Hz low at 14.996 MHz and about 400 Hz low on the 
> local BBC FM transmitters (88.9 - 94.3 MHz), which was a consistent 
> error.  LO on the R8500 tweaked against the 14.996 MHz, and the 9.996 
> MHz transmission was then spot-on.
> As a final check, the Edinburgh Tower ATC then showed just 6 Hz high, 
> so a most satisfactory result.  I can now tweak my FUNcube Dongle with 
> greater confidence!
> 73,
> David GM8ARV

More information about the time-nuts mailing list