Frequncy generator

[GPG]

Is this frequency plan viable.
http://www.geocities.com/gpg212/freqgen2.GIF

 

[GPG]

Is this frequency plan viable.

http://www.geocities.com/gpg212/freqgen2.GIF

oops

 

[Andrew Holme]

http://www.geocities.com/gpg212/freqgen2.GIF

oops

I would say not.

The 40Hz comparison frequency is too low. That's going to force you to use
a very low loop bandwidth, which will make it slow to lock.

You'd probably need to convert the mixer inputs to sine waves first; and
it'll produce sum and difference products, so you'd need a good low pass
filter to remove the 10.4-13 MHz sum product whilst passing the .4-4 MHz
difference. Finally, you'd need to square-up the sine waves again before
feeding them to the dividers.

If the aim is to produce low frequencies with high resolution, how about a
DDS?

If you want a PLL with small steps, consider Fractional-N.

 

[GPG]

I would say not.

The 40Hz comparison frequency is too low. That's going to force you to use
a very low loop bandwidth, which will make it slow to lock.

That is a concern. Once locked the loop filter can be slow, and I had
thought to use a lock detect circuit to change the current of the
detector for speeding up lock time.

You'd probably need to convert the mixer inputs to sine waves first; and
it'll produce sum and difference products, so you'd need a good low pass
filter to remove the 10.4-13 MHz sum product whilst passing the .4-4 MHz
difference. Finally, you'd need to square-up the sine waves again before
feeding them to the dividers.

Had a D FF in mind for that part, only the difference out.

If the aim is to produce low frequencies with high resolution, how about a
DDS?

Had a brief look, still reading.

If you want a PLL with small steps, consider Fractional-N.

Will research. Thanks.

 

[John Larkin]

Is this frequency plan viable.
http://www.geocities.com/gpg212/freqgen2.GIF

Only if you can tolerate immense amounts of phase noise.

This sort of thing looks ideal for a cheap DDS chip. I'd maybe double
the clock first, to 20 MHz, to not run too close to Nyquist.

John