[time-nuts] DDS - higher frequecies

Jim Lux jimlux at earthlink.net
Mon Nov 26 01:39:13 UTC 2012


On 11/25/12 5:19 PM, Said Jackson wrote:
> Hal,
>
> Check out the Analog Devices website. Good info on DDS Dacs there.
>
> You want to stay a bit away from the 1/2fs Nyquist limit in your DA. The reason is the image coming down from your 1MHz clock.
>
> If you output say 0.45MHz, you have an image at 0.55 MHz already (1MHz - 0.45MHz) so your filter has to be extremely steep to make that work and remove the spur at 0.55 MHz.. Check out the Mini Circuits LFCN low pass filters, they work at higher frequencies, and are very steep.. Your filter quality is going to determine how close you can get to Nyquist.
>

There's also a variety of intermod type products that show up, 
particularly when you talk about spurs from phase truncation and the 
like.  So you get not only the phase truncation spurs, but also all the 
aliases of those spurs.


There's a fair amount of literature around about this, especially from 
about 10-20 years ago, when 1GHz ADCs and the logic to drive them 
weren't easy to come by.  People wanted to generate signals in the 
hundreds of MHz range, but with logic and DACs that were slower.

There's a reason that people do dithering in these sorts of 
applications: it degrades the peak performance, but at least it keeps 
you from having a big spur in the wrong place.

There's a nice PhD dissertation out there (which name escapes me right 
now) that has a whole matlab code to simulate/analyze it.





More information about the time-nuts mailing list