LF: HiTech: DDS controlled via soundcard (crazy idea)

[DL4YHF@aol.com]

Hi all,

During a nice QSO in Jason mode today, and thinking about this and that, a crazy idea went through my brain. Maybe it's only good for the waste bin but how about this:

Let the PC produce a digital control word for a DDS  VIA THE SOUNDCARD ?

A DDS like the AD9832 needs a special serial control word, which is generated by a PIC processor in most cases. Three digital lines are required. How about "producing" the control word for the DDS via the SOUNDCARD ? We could either use a stereo output for clock+data (producing digital signals) , a few clamping diodes, two or three transistors and possibly a monoflop for the FSYNC signal (a third control line from PC to PIC). An alternative would be using different levels for the FSYNC, but proper level adjustment and clamping would be required because most audio outputs are only "AC coupled".

Johan (SM6LKM) has written a routine in "C" for PIC controllers which I have adapted to run under DOS, using the parallel port. But it's complicated to do this under windoze, especially if the DDS shall be controlled in a highly "synchronous" way through the parallel port. 


What would be the benefit of using the soundcard to drive a DDS ? 
- you wouldn't need a programmed PIC processor acting as an interface between PC and DDS.
- you wouldn't need a LPT interface which is tricky to access under Windoze NT + XP.
- you wouldn't need a serial interface which does not exist on some cheap notebooks.
- no need to fool around with USB programming
- The serial control word for the DDS chip can be timed very precisely under windoze. A 
continuous output (with update the DDS every 100 ms or so) is possible. 


Disadvantages:
- It's not a 'stand-alone' solution. A PC (or at least one of these tiny PDA's) is always needed. But which of the "modern modes" runs without a PC ?
- To implement this, basic knowledge of Windoze API programming is required. But I know at least one other reader of this reflector who has.. ;-)


If this sounds interesting, let me know and I'll do some more "research". The result could be a routine which produces raw 16-bit analogue samples to drive an AD9832 (which is the only DDS I have at the moment). 

73, 
  Wolf DL4YHF.