Frequency Generation Tests

g3cwi

Various people have suggested that the jitter performance of the counters will be better when they are generating frequencies that are exact mutiples of the clock frequency. As I want to generate 10.140MHz I did some tests:

The plots below show 10.140MHz being generated from a 5MHz clock crystal (span 30MHz and 1MHz).

The next set show 10.140MHz being generated with a 10.140MHz clock crystal.

Perhaps you might notice the difference?

Next I will investigate jitter at small offsets.

Cheers

Richard

g3cwi

1 KHz offset - OMG

.

5Hz offset - so bad it even sounded dreadful on a receiver.

Conclusions : using a crystal close to or at your wanted frequency is a dreadful idea unless you want an unmodulated signal (CW).

Cheers

Richard

Phil Pilgrim (PhiPi)

Yup, small offsets are ugleee! I ran into the same problem with my PLL retrofit for the MFJ Cub 40m transceiver. Its VFO output includes 5 MHz (mixed with 12 MHz). I had to replace the Prop's 5 MHz crystal with one at 4.9152 MHz to get out of that awful jitter zone.

-Phil

g3cwi

Hi Phil

A better use of my 10.140MHz crystals will be to build a filter as you suggested previously. I have three so that's the next job! I will revert to a 5MHz clock for the prop.

Cheers

Richard

PS my Prop 8mW WSPR signal spanned the Atlantic last night on 20m. Code by KO7M.

g3cwi

The addition of a 2KHz wide 10.140MHz filter and a seven element low pass filter has resulted in a nice clean output. Interestingly the WSPR "drift" figure has also improved. Now running at 20dBm on 30m.

Cheers

Richard

Phil Pilgrim (PhiPi)

You could probably pour white noise through a filter like that and get clean output!

-Phil

Cluso99

g3cwi: Richard, perhaps you might like to show us your propeller circuit you are using? I am interested to see how much is being done by the prop.

g3cwi

Nothing fancy!

WSPR CCT.pdf