[RTTY] Crystal filter width preferences for RTTY contesting

[Ian White GM3SEK]

Richard Ferch wrote:
>Jeff wrote:
>
>> The only reason I am continuing to beat the dead horse is that the filter,
>> when pressed in double-duty CW service as well as RTTY service, a wider
>> number than necessary for RTTY may cause an a drop in user happiness - if
>> the bandwidth were allowed to grow too wide.
>
>I'm not so sure that's a major concern. Maybe I am atypical, but on
>average I tend to use narrower IF filtering in RTTY than I do in CW.
>
>I have a K3 with the "250 Hz" (actually 370 Hz) roofing filter,
>configured to cut in at the 350 Hz setting. On RTTY I move my DSP filter
>settings back and forth between 400 Hz (with a 500 Hz IF filter) and 350
>Hz (370 Hz IF filter). I can't say I feel that one is all that much
>better on average than the other, but it's psychologically comforting
>(if nothing else) to have the choice. On reasonably strong signals with
>nearby very strong signals I sometimes narrow the DSP filter down even
>more, but as Chen says, on weak signals this is counter-productive.
>
>On the other hand, in CW I don't use the narrower filter as much as I do
>in RTTY, except in very specific situations. I find that narrower
>filtering doesn't help me all that much most of the time in CW. Indeed,
>it can actually make the wetware filtering less effective by making the
>noise sound more like the signal. Anyway, so much of the interference
>problem in CW is due to keying clicks and phase noise that I am not
>convinced that using narrower IF filters is all that much help on
>average. To put it another way, although narrower AF or DSP filtering
>can be useful in a number of situations in CW, I am not sure that doing
>the narrower filtering at IF is all that much better, other than when
>the neighbouring station is pumping your AGC, and you often have other
>weapons at your disposal to deal with that (turning the RF gain down
>and/or turning off AGC, combined with AF or DSP filtering).
>
>Be that as it may, I also have the impression, rightly or wrongly, that
>RTTY contesters actually tend to crowd in closer to one another than CW
>contesters typically do, although that might be a function of the types
>of contests I am most likely to be active in. It seems to me as if RTTY
>contesters are used to sidling up right beside one another, whereas CW
>contesters tend to stay far enough apart so that the key clicks from
>neighbouring stations aren't too distracting, regardless of what filters
>they are using.
>
>Just a thought...
>
>73,
>Rich VE3KI
>

My experience is very similar to Rich's. No argument about the optimum 
bandwidth under QRM-free conditions being 350-400Hz; but in intense QRM 
the word "optimum" takes on a whole different meaning.

During 40m RTTY contests in Europe, the big guns tend to settle into a 
reasonably comfortable spacing between their 500Hz filter passbands. 
Everyone else then has to squeeze through the gaps in between. That 
means we routinely have to copy signals at S3-4 with an S9+30dB 
interfering tone only about 200Hz away.

In this situation the "optimum" crystal filter bandwidth is one that 
will be narrow enough to prevent the extremely strong QRM signal from 
capturing the IF AGC and/or excessively challenging the IF DSP. 
Certainly the keying sidebands of the QRM will overlap the wanted 
signal, and decoding of the wanted signal will be degraded by the 
narrower bandwidth... but even degraded decoding is better than none at 
all.

When I still had the FT-1000, it was often necessary to operate 
substantial parts of a 40m RTTY contest with the two cascaded "250Hz" 
filters [1]. This worked very well, and allowed decoding in intense QRM 
where the cascaded "500Hz" filters were simply too wide. So with the new 
K3, it came as no surprise that the 400Hz 8-pole roofing filter was too 
wide as well. (The strong-signal  effects in the K3 are different from 
those in the FT-1000, but that isn't the main issue here.)

With help and advice from Wayne at Elecraft, I changed a couple of 
capacitor values in the 200Hz 5-pole filter to increase the -6dB 
bandwidth to 260Hz [2]. This is also my narrowest roofing filter for CW, 
which is a slight but acceptable compromise compared with the original 
195Hz bandwidth. For RTTY it is normally used with a DSP bandwidth of 
350Hz, to avoid any further narrowing of the nose bandwidth. This brings 
the K3 back into a similar level of performance that I enjoyed with the 
fully loaded FT-1000MP.

I didn't cut myself any extra slack for tuning errors - it's a luxury we 
cannot afford. It isn't hard to learn how to tune accurately and 
quickly.

When running, the second RX is usually on the same frequency with a 
400Hz roofing filter and its own MMTTY window. In this QRM-limited 
environment, the narrower bandwidth fairly consistently gives the better 
decoding,  while the wider bandwidth contributes a better awareness of 
what's happening nearby.

Then it's all down to the usual crossword skills.





[1] As Joe has pointed out,  the bandwidths of the individual 8.215MHz 
and 455kHz filters in the FT-1000 series are actually closer to 350Hz. 
Yaesu labeled them "250Hz" because that is the approximate bandwidth 
when the two filters are cascaded. Inrad then did the same for their 
aftermarket filters.

The 8.215MHz 8-pole "250Hz" roofing filter for the K3 is supplied by 
Inrad and is identical to the Yaesu version (except for the adapter 
board), so despite what the label says, the "250Hz" roofing filter 
bandwidth is actually about 350Hz. Inrad and Elecraft do not supply a 
true 250Hz roofing filter; the choices from stock are the 8-pole 350Hz 
or the 5-pole 200Hz.

[2] Details are at:
<http://www.ifwtech.co.uk/g3sek/misc/kfl3-200_mod260.gif>
The filter schematic is in Elecraft's K3 schematics file.
Although Elecraft provided helpful advice, this is not an official 
Elecraft modification.


-- 

73 from Ian GM3SEK