[RTTY] K3

[Ian White GM3SEK]

Kok Chen wrote:
>
>On Jan 6, 2009, at 3:20 PM, Joe Subich, W4TV wrote:
>
>>>
>>> The AD9834 DDS that it uses is definitely capable of FSK and
>>> PSK, but both the FSELECT and PSELECT lines of the chip are
>>> grounded in the K3, so I have a sneaky suspicion that the rig
>>> might be a Keyed AFSK rig.
>>
>> I suspect Lyle is generating both PSK31 and FSK in the DSP and
>> not creating external audio.
>
>However, I suspect that it is the PA that is contributing to the poor
>IMD (because the IMD becomes quite respectable when I turn off the 100
>watt PA).  If the PSK31 envelope goes through any nonlinearity, you
>get IMD, no matter how you had generated the initial modulation in the
>first place.
>
>As I mentioned earlier, you can slew FSK tones in a phase continuous
>way (instead of applying envelope waveshaping to each mark and space
>tones) and create an FSK signal whose envelope is constant and yet has
>bounded keying sidebands.  That might solve the RTTY problem, but it
>still wouldn't solve the PSK31 problem.
>

Thanks to everyone for the thoughtful comments (which arrived here 
overnight).

Like Joe, I feel sure that the K3 generates its "direct" FSK entirely in 
DSP. However, we don't have any details about the implementation of 
frequency slewing, or about the levels of keying sidebands compared with 
the AFSK alternative.

We do know a little more about the FT-1000. Starting from the 1000MP, 
Yaesu has been implementing direct FSK by DDS. However, I suspect they 
are simply banging in new frequency instructions and leaving it to the 
DDS chip to handle the transition.


Jumping back to the discussion about narrow roofing filters for RTTY, 
W4TV pointed out that Elecraft don't presently offer a filter with a 
true bandwidth of about 250Hz. The so-called "250Hz" 8-pole filter has a 
-6dB bandwidth of almost 350Hz, which is almost the same as the 400Hz 
filter, while the 200Hz 5-pole filter is too narrow for practical use.

Joe mentioned a simple modification to increase the bandwidth of the 
200Hz filter to about 260Hz. Here is a summary.

The schematic of the 5-pole K3 filters is on page 9 of:
http://www.elecraft.com/manual/K3_Schematics_Oct2008.pdf

The modification was suggested by Wayne, N6KR, and involves changing 
both C4 and C5 from 1000pF to about 800pF. The modification involves 
removing and replacing two 0604 SMD chips.

Lower capacitance will give a wider bandwidth. I used 830pF (680 + 
150pF, 5% NP0) which gave a bandwidth of about 260Hz. The center 
frequency remains almost exactly the same. The measured results are 
shown at:
http://www.ifwtech.co.uk/g3sek/misc/kfl3-200_mod260.gif

The increase in bandwidth to 260Hz makes the filter much more practical 
for RTTY under heavy QRM conditions, without losing its effectiveness as 
a roofing filter for CW.


-- 

73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek