Re: VLF: DK7FC's "big" VLF PA - Outphasing

[Andy Talbot <andy.g4jnt@gmail.com>]

Two stereo outputs, could be digitised and the resulting values used to  access a lookup table that contains the pulse width.   It may just about be possible that PIC could be fast enough;  a DSPPIC certainly would be.

 

Andy

www.g4jnt.com

 

 

On 4 January 2011 18:38, wolf_dl4yhf <dl4yhf@freenet.de> wrote:

Hello Markus and Stefan,

Phew, that sounds interesting.. the question is how to convert the output of the inverse FFT (with the 'analog' chirped Hell signal) into your described digital (square wave) output signal. Could it be easier to synthesize this signal directly, i.e. without an FFT at all ?

Cheers,
  Wolf  DL4YHF .

Am 04.01.2011 00:00, schrieb Markus Vester:

Dear Stefan,

 

thanks very much for sharing this very useful design!

 

Just in case you would like to extend this to linear modulation: The H-bridge topology naturally lends itself to the Outphasing technique (aka "Chireix" modulation", "Ampliphase", "LINC"), which was invented in the 1930ies to increase AM transmitter efficiency. The idea is to drive the two half-bridges separately with a variable phase difference: 180° gives full power, 90° gives -3 dB, and 0° gives no differential output across the load.

 

Instead of generating the two carriers by special envelope-driven phaseshifter hardware, you could also take a software approach through the two outputs of a stereo soundcard. It may then be best to feed the PA drivers directly, avoiding the dividers which can introduce an unwanted 180° ambiguity.

 

If you feel like trying this concept, I could be persuaded to adapt the Chirped Hell software for this purpose. What day is your birthday? ;-))

 

Best regards,

Markus (DF6NM)

 

 

PS: In case you decide to use small audio transformers at the input, make sure their insulation can safely withstand the line voltage!