[Elecraft] Let's Try This Again -- was "[K3] ... A 750 Hz, 8-Pole ... Filter?"

[Ian White GM3SEK]

David Woolley wrote:

>Ian White GM3SEK wrote:
>
>>
>> However, contesters are always looking for an extra edge, and a true
>> 250-270Hz 8-pole filter with a steeper transition from the passband into
>> the stopband would be the next step to try.
>>
>At that sort of bandwidth you are going to be clipping a non-trivial
>amount of sideband power.  Are you sure that a brick wall filter would
>be a good idea; I would wave thought it would cause significant dispersion.
>

That theory only applies to the copy of weak RTTY signals against a 
background of noise. But when the main problem is QRM, the best 
*available* copy is obtained by reducing the bandwidth and accepting a 
small reduction in accuracy.

The recommendation for a 250-270Hz filter is based on many years of 
practical experience in heavy contest QRM, starting with different 
combinations of cascaded filters in the FT-1000MP, and then moving on to 
different combinations of roofing filters and DSP in the K3. The 
measurable performance parameters have been certificates and a modest 
amount of silverware.

With off-the-shelf roofing filters in the K3, the quality of RTTY copy 
in extreme QRM was inferior to the FT-1000. The 400Hz filter let in too 
much QRM which was routinely triggering the hardware AGC, while the 
200Hz 5-pole required manual fine tuning for each new caller which made 
it unusable for serious contesting. With the modified 270Hz 5-pole, the 
performance of the two radios is now about the same... so the next 
logical step forward would be a 250-270Hz 8-pole, a "gaussian to 6dB" 
design with improved roll-off outside the passband.



-- 

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