[Elecraft] K3 CW rise time mod - Definition

[Joe Subich, W4TV]

Looking at the scope picture of a single Morse dit at G4AON's 
site http://www.astromag.co.uk/k3/k3_envelope2.jpg indicates 
the trailing edge might be a bigger contributor to potential 
clicks than the leading edge of the waveform.  Note the upper 
right corner has a fairly "sharp" transition from "on" to 
"falling" and compare that to the nicely shaped transition at 
the upper left corner. 

>From all reports the K3 is more than acceptable but if any 
additional effort is put into CW shaping, the "upper right" 
transition may be a place to look. 

73, 

   ... Joe, W4TV 
     

> -----Original Message-----
> From: elecraft-bounces at mailman.qth.net 
> [mailto:elecraft-bounces at mailman.qth.net] On Behalf Of Brian Alsop
> Sent: Sunday, March 30, 2008 11:01 AM
> Cc: elecraft at mailman.qth.net
> Subject: Re: [Elecraft] K3 CW rise time mod - Definition
> 
> 
> For what it is worth,  I've just completed a Fourier analysis of the 
> original K3 and revised K3 waveforms presented by G4AON on 
> his website.
> 
> I was a bit concerned that the sharpening of the leading edge would 
> adversely affect bandwidth and possibly generate clicks.
> 
> Bottom line for the region of the spectrum between 0 and +/- 200 Hz 
> there is a slight increase in average power (maybe 3 db) for the new 
> "sharper" waveform.  In the range of 200 to 300 Hz the average power 
> increases maybe 10 db.  However at this point the spectral energy is 
> down 50 db anyhow. Beyond this 350 Hz,  the spectrum energy of the 
> "sharper" waveform actually falls off relative to the 8ms case.
> 
> These conclusions are "eyeball averages" of the sawtooth 
> looking output 
> produced.  The teeth go above and below the zero db level.
> 
> I was pleasantly surprised.  Apparently it isn't all in "rise 
> time" it 
> is also how that waveform is shaped.
> 
> Here are the assumptions/limitations of the analysis.
> 1) Code used "Fast Fourier Transform"  by M.F. Hajen
> 2) A single dit was modeled.
> 3) Both dits had the same length in the model (about 70ms 
> representing 
> about 20 wpm).
> 4) Waveforms were digitized in 1 ms steps.
> 5) The assumption that the waveforms given represented 
> voltage and hence 
> db was calculated by 20 log (amplitude/amplitude at 0 Hz)
> 
> This doesn't reflect repeating dits or CW made up of dots and 
> dashes.  
> However, it is probably a worse case analysis or close to it.
> I don't have a web site to post on.  There is a .jpg file 
> available of 
> results.
> 
> Disclaimer
> I'm not an expert on this or on using the calculational tools.   The 
> output does at least agree with the position of the first zero point 
> predicted by a sinx/x  formula for a single pulse of near 
> infinite rise 
> time and same width.  Clearly it doesn't represent other 
> effects in the 
> rig which affect spectrum. 
> 
> Like a lawyer once said: "I'm giving you some free advice, it 
> is worth 
> every penny paid for it"
> 
> 73 de Brian/K3KO
> .
> 
> Paul Christensen wrote:
> 
> >>    I would like to see someone please take this discussion 
> one step 
> >> further and define the terms "hard" keying and "soft" keying, and 
> >> describe how either an operator or someone listening would 
> be able to 
> >> discriminate between the two.
> >
> >
> > Gary, this topic has been covered in great detail in nearly 
> every ARRL 
> > Handbook for at least the last fifty years -- as well as other 
> > publications. Some of the most relevant material is covered in 
> > articles during the '40s and '50s by By Goodman and George Garmmer. 
> > Not all of the material pertaining to "optimum key-shaping" is 
> > accurate, but otherwise, the content is still accurate.
> >
> > Many of us do not like to be held hostage to one rise/fall 
> setting as 
> > a "one size fits all" answer to our operating practices. With 
> > DSP-generated raised-cosine key shaping, it's now possible 
> to tighten 
> > up key shaping at high CW speeds without appreciably consuming 
> > additional bandwidth.
> >
> > Paul, W9AC
> > _______________________________________________
> > Elecraft mailing list
> > Post to: Elecraft at mailman.qth.net
> > You must be a subscriber to post to the list.
> > Subscriber Info (Addr. Change, sub, unsub etc.):
> > http://mailman.qth.net/mailman/listinfo/elecraft   
> > Help: http://mailman.qth.net/subscribers.htm
> > Elecraft web page: http://www.elecraft.com
> >
> 
> _______________________________________________
> Elecraft mailing list
> Post to: Elecraft at mailman.qth.net
> You must be a subscriber to post to the list.
> Subscriber Info (Addr. Change, sub, unsub etc.):
>  http://mailman.qth.net/mailman/listinfo/elecraft    
> 
> Help: http://mailman.qth.net/subscribers.htm
> Elecraft web page: http://www.elecraft.com