[time-nuts] DDS module

Graham planophore at aei.ca
Sun Jul 21 13:56:56 UTC 2013


Good morning Joe (and all),

I have been playing around with a few of these DDS modules as well as a 
couple of the lower end DDS chips (i.e. AD9834, AD9850, AD9851) and they 
do work well but they are not perfect.

I have a Fox Delta Antenna SWR analyzer which has an AD9850 and a 125Mhz 
surface mount oscillator and although I have not done any extensive 
testing of the devices stability it is relatively stable to a N3ZI 
design DDS VFO which uses a AD9834 and a 80Mhz dip can (ie. not surface 
mount) oscillator.

The AD9834/80Mhz dip can oscillator DDS wanders quite a bit and when I 
listen to it on a receiver it sounds "raspy". I have simple heater 
attached to the dip can - it is just as bad with as without. There has 
been some discussion on the Knights QRSS reflector on the subject. If 
you are not yet a member, you can join here: 
http://mail.cnts.be/mailman/listinfo/knightsqrss_cnts.be

I am guessing "the amateur in town" is Dave WA5DJJ. He and I have 
exchanged a few emails on this subject (and others). There is a fellow 
in Australia as well which I have exchanged a few ideas.

One common method of taming some of wandering habits of the inexpensive 
DDS modules is to put a small coin on the dip can oscillator. I noted 
this behavior as well some time ago. I can't quite put my finger on why 
it helps. Some postulate that it is adding just enough thermal mass to 
keep drift down when there is moving air around the can oscillator.

What I have noticed is that the DDS with the surface mount oscillator is 
more stable (remember, no extensive and exhaustive testing at this 
point, just simple observation). I have also noticed that if I have one 
of the DDS modules on my bench that when I monitor the output of said 
device that tapping on the desk also seems to cause some wandering of 
the synthesized frequency indicating that the the dip can oscillator is 
possibly microphonic in some way.

It seems at this point as though the SMD oscillators are a better choice 
between these two low cost options. Of course, TCXO or OCXO would be a 
better choice.

I stumbled across an Analog Devices application note AN-419 title "A 
Discrete Low Phase Noise 125Mhz Crystal Oscillator for the AD9850 
Complete Direct Digital Synthesizer" which has some interesting tid bits 
of information. I am using this as a starting point for some 
experimenting with reference clock sources for DDS's. One bit of 
information that no doubt is in the DDS chip's data sheet is that the 
reference clock needs to be at least 3 Vp-p and centered on 1/2 VDD. 
Seems so obvious but makes me wonder if adding a small coin to top of 
the DIP can oscillator is not only helping thermal wise but is somehow 
helping improve phase noise.

I think we have tended to think of a DDS as being "it". That is, we can 
generate something stable and precise digitally and it will be good no 
matter what. What we (I) forgot when first exploring the devices was 
that what we get out the synthesized end is (relatively speaking) no 
better than what we put in (i.e. the reference clock), another example 
of GIGO.

cheers, Graham ve3gtc



On 13-07-21 04:03 AM, Joseph Gray wrote:
> I have a few of those $5 AD9850 DDS modules from China. I'm going to use
> one to replace a crystal in a transmitter. I attached it to a TI MSP430
> Launchpad and programmed a fixed frequency.
>
> Just out of curiosity, I wanted to see how stable this thing is. Using
> cellophane tape, I put a few pieces of foam around the oscillator can. Then
> I hooked it up to my HP counter, which is locked to my GPSDO.
>
> I'm not logging, but I have kept an eye on the setup for well over 24 hours
> now. So far, the frequency has held to within 0.01-0.08 of the programmed
> value.
>
> The house thermostat is set at 75 F, but this room gets somewhat warmer due
> to the equipment in it. Also, this room gets the morning sun.
>
> Considering the cost, I am impressed with how well this little module works
> and how stable it is. There is an Amateur in town who has been using
> several of these modules to drive WSPR and QRSS transmitters. He has done
> up a poor man's oven to keep his transmitters on frequency.
>
> Joe Gray
> W5JG




More information about the Time-nuts_lists.febo.com mailing list