EMRFD Message Archive 2231

Message Date From Subject
2231 2008-10-12 00:25:00 Ashhar Farhan Audio DDS as a reference for PLL
If we consider a 7MHz VCO fed to divide-by-1024 counter, we would get
a 6.83KHz output, that is in the audio range.
We can then program a PC card to generate a sine wave of 6.83KHz
(using it as a DDS) and use it as a reference to lock the VCO.

The current day PC sound cards can be used to generate frequencies of
22KHz and more with resolution of 16 bits and more. My specific issues
are:
1. How will the phase noise of the audio card effect the VCO? can we
use an active band-pass filter to clean up the sound card output
before feeding it to the phase detector?
2. Though tuning can be very fine, is there a way to theoretical
predict the in-band spurs of such a system?

using a non-dedicated dds for frequency control is interesting for me
as a pc is available in every home lab.

here is a link (though i have no idea of how well it performs, onno -
are on this list?)
http://www.geocities.com/pa2ohh/04dds1.htm

- farhan
2234 2008-10-12 03:35:39 Lasse Re: Audio DDS as a reference for PLL
Ashhar Farhan skrev:
> 1. How will the phase noise of the audio card effect the VCO? can we
> use an active band-pass filter to clean up the sound card output
> before feeding it to the phase detector?
>
I am sure you need a lot of filtering to get a resonably clean
reference. With a divisor of 1000 you get a 60 dB multiplication of
phase noise and spurs from the reference to show up witin your loop
bandwith.
Outside of your loopfilter the VCO will set the noise.


> 2. Though tuning can be very fine, is there a way to theoretical
> predict the in-band spurs of such a system?
>
See above, with-in the loop filter the spurs from your reference will be
mulitplied 60 dB, but outside of the filter you will mainly have the
reference leak through every 6 kHz. The amount of this is a function of
what phase detector, filter and sheilding you have.

*TANSTAAFL*


or
*There Ain't No Such Thing As A Free Lunch*
2235 2008-10-12 16:49:38 Murray Re: Audio DDS as a reference for PLL
Hi Farhan

I've built a version of Onno's PLL using his discreet phase detector but
a different PIC for the DDS.
My reference runs 120khz - 230khz or so. I use this to give me a stable
VFO for an old FT-101 so
I can experiment with the various digital modes. I did (do!) have
issues with noise from the DDS
and/or the phase detector but because I don't need to change frequencies
quickly I can damp them
out of the control loop. While it wouldn't work for break-in, the VCO
settles in less than a second.
I'm reasonably confident that any noise phase or otherwise I suffer is
just normal oscillator noise
in the VCO.

Cheers
Murray
VK3ACF
Melbourne



Ashhar Farhan wrote:
> If we consider a 7MHz VCO fed to divide-by-1024 counter, we would get
> a 6.83KHz output, that is in the audio range.
> We can then program a PC card to generate a sine wave of 6.83KHz
> (using it as a DDS) and use it as a reference to lock the VCO.
>
> The current day PC sound cards can be used to generate frequencies of
> 22KHz and more with resolution of 16 bits and more. My specific issues
> are:
> 1. How will the phase noise of the audio card effect the VCO? can we
> use an active band-pass filter to clean up the sound card output
> before feeding it to the phase detector?
> 2. Though tuning can be very fine, is there a way to theoretical
> predict the in-band spurs of such a system?
>
> using a non-dedicated dds for frequency control is interesting for me
> as a pc is available in every home lab.
>
> here is a link (though i have no idea of how well it performs, onno -
> are on this list?)
> http://www.geocities.com/pa2ohh/04dds1.htm
>
> - farhan
>
>
2236 2008-10-12 16:49:41 Gheorghian Romeo Re: Audio DDS as a reference for PLL
There will be no significant spurs, at the output VCO if the final
divided frequency will be like 6kHz or better more in a PLL HF system.
The PLL loop filter will suppress to a insignificant level any spur if
the frequency at the fq and phase comparator will be more than 1 KHz,
better 6 or much better 12kHz.
There are very good PLL synthesizers made in the past years using
VCO's which frequency was divided using programmable CMOS or ECL&CMOS
dividers (like MC145170...) to 1 kHz or around and have very high
performance, see the K2 project made by Elecraft. I see no reas
2237 2008-10-12 16:49:41 Gheorghian Romeo Re: Audio DDS as a reference for PLL
Correcti
2238 2008-10-12 22:01:33 Ashhar Farhan Re: Audio DDS as a reference for PLL
the software required to generate this reference shouldn't be too
difficult. a few years ago, I had written a dsp shell to experiment
with the dsp algorithms of emrfd, it can be easily used to generate
the reference.
this is intriguing enough to 'throw code at the problem'. i'll finish
the pc end this week and post results soon.
- farhan

On 10/13/08, Gheorghian Romeo <yo8can@yahoo.com> wrote:
> Correcti