From: Andy Talbot <[email protected]>
Subject: Re: LF: WSPR without an SSB TRX ? FRM Joe Taylor NPwinner
To: [email protected]
Date: Wednesday, 7 April, 2010, 21:53

Not quite sure what you're trying to suggest...

 

There is no 'magic' to a phase continuous oscillator.   Any single oscillator whose frequency is controlled by adjusting the frequency determining element will inherently have to be phase continuous.  Simple examples are a varicap tuned LC or crystal source, or a DDS where changing the increment to the NCO is the tuning element. 

 

In the former you may have a problem getting sufficiently close enough to a 1.46Hz shift (actually the value is 12000/8192 = 1.46484375Hz).   But for a DDS no problem, even with many stages of frequency multiplication.   I have only ever used the DDS implementation, but there are many people out there who have built highly sucessful low power WSPR beacons - many based around varicap controlled crystal oscillators controlled from a microcontroller.  

 

As mentioned before,  JT4 is much the same as WSPR, and I've tested many successful beacon sources for that, - in fact the 10GHz beacon GB3SCX transmits JT4G using a DDS in a PLL, and controlled to set one of four frequencies from another PIC generating the JT4G symbols.  http://www.scrbg.org/JT4_GB3SCX.pdf

 

The only really critical thing about the mode is gettign the symbol timing right.    Be just a fraction of a percent out and after the 110 second transmission period, if you have drifted by more than about 20% of the 680ms long symbol period, decoding will suffer.    Actual start time of the transmission can vary by a few seconds either way, but symbol timing MUST be better than 0.1%.   No big deal for a crystal oscillator, but the PIC code must gets its counters and division ratios right!   (Certain cheap sound cards cannot manage it though)

 

But, while on the same subject.  Why reject a PC out of hand.   They are now so ubiquitous, that old laptops appear at rallies/boot sales/junk fairs and cost next to nothing.   Almost anything running Win 98 or later will do for WSPR and WSJT - although you may have to use an external USB headphone dongle if the internal soundcard is too ancient.  (a couple of quid from Ebay)

At the recent Bournemouth Sale, I saw Toshiba laptops (135MHz clock running Win 98)  identical to the one I use currently 24/7 for the 5MHz beacon monitoring project, being  sold by a trader for a mere £15 each.    You could probably find a suitable one on a rubbish tip even, fully working.   I also picked up a very nice Toshiba lightweight notebook (700 MHz clock) Win XP  for £40 a couple of years ago, and have to stop buying old laptops as they're just too plentiful now.

 

 

Andy
www.g4jnt.com

On 7 April 2010 20:55, M0FMT <[email protected]> wrote:

Please see below reply from Joe Taylor on this very subject September last year, This is the guy who actually wrote this application (you might say from the Horse Mouth) and he said :-   On Mon, 7/9/09, Joe Taylor <[email protected]> wrote: From: Joe Taylor <[email protected]> Subject: Re: WSPR To: "M0FMT" <[email protected]> Date: Monday, 7 September, 2009, 1:47 PM Hi Peter, > Does your Software have a data output that gives a hi lo i.e 0,1 on the serial port other than the PTT Hi Lo? No. > If the software doesn't have this facility why not? Because there is no standard, nor even anything remotely approaching one, that would permit hard-keying an oscillator to produce continuous-phase 4-FSK with offset 1.465 Hz between tones. The WSPR program, as distributed, is essentially a sound card mode. With that said, it should be noted that a number of people have built special-purpose WSPR transmitters that implement the necessary 4-tone FSK capability in hardware.  Their keying is typically controlled from a  microprocessor, with the encoded message (sequence of tones) stored in ROM.     -- 73, Joe, K1JT 73 es GL petefmt --- On Wed, 7/4/10, Rik Strobbe <[email protected]> wrote: From: Rik Strobbe <[email protected]> Subject: RE: LF: WSPR without an SSB TRX ? To: "[email protected]" <[email protected]> Date: Wednesday, 7 April, 2010, 14:23 hello Chris, Jim, group, an alternative to a transverted would be a divider, let's say from 5MHz (from a HF tranceiver) to 500kHz. That would limit the hardware to some resistors an a single CMOS or TTL IC. I have written a small apllication that generates the appropriate WSPR audio signal based on the input frequency, output frequency and divider ratio. It worked fine for some local tests, but due to lack of I did no real "DX tests" so. But if anyone wants to try this route I will be happy to send the software. 73, Rik  ON7YD - OR7T ________________________________________ Van: [email protected] [[email protected]] namens James Moritz [[email protected]] Verzonden: woensdag 7 april 2010 15:09 Aan: [email protected] Onderwerp: Re: LF: WSPR without an SSB TRX ? Dear Chris, Andy, LF Group, What Andy suggests is fine for generating a WSPR signal, but receiving WSPR via the PIC might be a bit more difficult... But also at http://www.g4jnt.com/LFUpconv.pdf is Andy's simple phasing up-converter, which could be used to convert a PC audio output to 500kHz and be used in place of the VCO in your TX. Trawling through the Forum topics on the WSPRnet.org website (you might have to sign up to see all the material) will reveal many designs for homebrew HF WSPR beacons that could mostly be adapted to 500k. Cheers, Jim Moritz 73 de M0BMU