K3 CW Decoder, Not very good

[Joe Reaves]

Hello friends,
I have been trying to get my K3 decoder to work better by trying,
cutting back on the RF gain, using narrow filtering, to limit noise.
tuning back and forth across the signal to find a good copy spot.
Just can't seem to get it to read very well no matter what I try...?
Even strong signals don't seem to read very well... Is there a way to
improver the cw code reader / decoder in the K3 ?
Thanks, Best 73' Joe R. / W7Joe


     
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Don Cunningham]

Joe,
I do a LOT of RTTY work, and have since about 1972 and have tried nearly all
the supposed RTTY/CW decoders over those years and NONE of them have ever
decoded hand generated CW very well.  They do okay on machine (keyboard)
generated CW, but hardly any copy on a fist.  Humans just aren't consistent
enough, hi.
73,
Don, WB5HAK

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Don Wilhelm-4]

Joe.

If you are trying to decode hand sent CW. you will likely have to
provide a lot of "fills" unless the sending operator is quite skilled at
timing and spacing.  Machine sent (keyboard or computer) CW should be
easily decoded.

73,
Don W3FPR

Joe Reaves wrote:

> Hello friends,
> I have been trying to get my K3 decoder to work better by trying,
> cutting back on the RF gain, using narrow filtering, to limit noise.
> tuning back and forth across the signal to find a good copy spot.
> Just can't seem to get it to read very well no matter what I try...?
> Even strong signals don't seem to read very well... Is there a way to
> improver the cw code reader / decoder in the K3 ?
> Thanks, Best 73' Joe R. / W7Joe
>
>  

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Bill VanAlstyne W5WVO]

So that begs the question, I suppose -- is CW sent by a keyer with paddle
machine-sent or hand-sent? It's actually both, of course. Some of the
metrics are controlled by logic (hardware and/or software), but
inter-character and inter-word spacing are controlled by the human operator.
In my experience, this is usually enough to screw up the K3's decoding
algorithm. And of course once you start talking about bug-sent or
straight-key-sent code, that's another planet altogether.

Developing a really "brainy" hand-sent CW decoder would be very challenging,
to say the least -- and even if you could do it, it would almost certainly
require way more number-crunching horsepower than is available in the K3's
processor. It's one of those things that fall under the label,
"Theoretically possible, but why bother?" There are many much more useful
machine-generated digital encoding methods available today. I think Morse
code decoding should probably be allowed to remain in the wetware domain. If
you "need" a CW decoder, let me suggest that PRACTICE has been known to work
very well when applied consistently. :-)

Bill W5WVO

--------------------------------------------------
From: "Don Wilhelm" <[hidden email]>
Sent: Monday, June 14, 2010 9:06 PM
To: "Joe Reaves" <[hidden email]>
Cc: "elecraft mail" <[hidden email]>
Subject: Re: [Elecraft] K3 CW Decoder, Not very good

> Joe.
>
> If you are trying to decode hand sent CW. you will likely have to
> provide a lot of "fills" unless the sending operator is quite skilled at
> timing and spacing.  Machine sent (keyboard or computer) CW should be
> easily decoded.
>
> 73,
> Don W3FPR
>
> Joe Reaves wrote:
>> Hello friends,
>> I have been trying to get my K3 decoder to work better by trying,
>> cutting back on the RF gain, using narrow filtering, to limit noise.
>> tuning back and forth across the signal to find a good copy spot.
>> Just can't seem to get it to read very well no matter what I try...?
>> Even strong signals don't seem to read very well... Is there a way to
>> improver the cw code reader / decoder in the K3 ?
>> Thanks, Best 73' Joe R. / W7Joe
>>
>>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[wayne burdick]

Hi Joe,

If you've already narrowed the filter and tuned the signal in properly  
(with the aid of CWT), the next step is to adjust the threshold of the  
decoder. Hold the TEXT DEC switch, then use VFO A to adjust the  
threshold until the CWT icon blink is closely correlated with the  
incoming CW signal. You can set it to AUTO, too, but it takes several  
seconds to acquire. I find manually selecting a threshold works best.

Of course if the signal is weak, noisy, or fading in and out, the  
decoder will have a much harder time.

73,
Wayne
N6KR

On Jun 14, 2010, at 7:49 PM, Joe Reaves wrote:

> Hello friends,
> I have been trying to get my K3 decoder to work better by trying,
> cutting back on the RF gain, using narrow filtering, to limit noise.
> tuning back and forth across the signal to find a good copy spot.
> Just can't seem to get it to read very well no matter what I try...?
> Even strong signals don't seem to read very well... Is there a way to
> improver the cw code reader / decoder in the K3 ?
> Thanks, Best 73' Joe R. / W7Joe
>
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Julian, G4ILO]

I think the K3 CW decoder is surprisingly good, but it will lose the signal when it gets weak whereas I have seen the PC sound card decoder MRP40 continue to decode a signal that has gone into the noise and I can't hear. All computer decoders tend to be thrown by noise, interference and irregular sending.

[Guy, K2AV]

Being associated with people who have spent a career in
automating/computerizing "copy" of human directly generated
communication, they will tell you that the human mind has some
abilities to "fill", "connect," and "decode" in a manner that are
simply not understood by anyone, anywhere...only envied. In other
instances, the difficulty is that long running algorithms can decode,
but the processing and hardware cost makes it unmarketable for the
general public or useless in response-time-critical situations, such
as a ham contest.

They are working on it, because there is HUGE money in it for the
breakthroughs.  The amount of R&D money spent on these kinds of things
is staggering.

>From their perspective, you are lucky to have the degree of copy that you do.

Conspiracy theories abound that breakthroughs have been made but are
top secret, in use by military, CIA for monitoring voice channels for
"trigger" phrases. I tend to discount it because the end results seem
not to have the improvements such would bring.

Given the success that Skimmer is having, it would seem that export of
that function to a PC would garner resource levels to the problem not
available internally to the K3, if you are not satisfied with the most
recent K3 improvements.

73, Guy.

On Mon, Jun 14, 2010 at 11:38 PM, Bill VanAlstyne W5WVO
<[hidden email]> wrote:

> So that begs the question, I suppose -- is CW sent by a keyer with paddle
> machine-sent or hand-sent? It's actually both, of course. Some of the
> metrics are controlled by logic (hardware and/or software), but
> inter-character and inter-word spacing are controlled by the human operator.
> In my experience, this is usually enough to screw up the K3's decoding
> algorithm. And of course once you start talking about bug-sent or
> straight-key-sent code, that's another planet altogether.
>
> Developing a really "brainy" hand-sent CW decoder would be very challenging,
> to say the least -- and even if you could do it, it would almost certainly
> require way more number-crunching horsepower than is available in the K3's
> processor. It's one of those things that fall under the label,
> "Theoretically possible, but why bother?" There are many much more useful
> machine-generated digital encoding methods available today. I think Morse
> code decoding should probably be allowed to remain in the wetware domain. If
> you "need" a CW decoder, let me suggest that PRACTICE has been known to work
> very well when applied consistently. :-)
>
> Bill W5WVO
>
> --------------------------------------------------
> From: "Don Wilhelm" <[hidden email]>
> Sent: Monday, June 14, 2010 9:06 PM
> To: "Joe Reaves" <[hidden email]>
> Cc: "elecraft mail" <[hidden email]>
> Subject: Re: [Elecraft] K3 CW Decoder, Not very good
>
>> Joe.
>>
>> If you are trying to decode hand sent CW. you will likely have to
>> provide a lot of "fills" unless the sending operator is quite skilled at
>> timing and spacing.  Machine sent (keyboard or computer) CW should be
>> easily decoded.
>>
>> 73,
>> Don W3FPR
>>
>> Joe Reaves wrote:
>>> Hello friends,
>>> I have been trying to get my K3 decoder to work better by trying,
>>> cutting back on the RF gain, using narrow filtering, to limit noise.
>>> tuning back and forth across the signal to find a good copy spot.
>>> Just can't seem to get it to read very well no matter what I try...?
>>> Even strong signals don't seem to read very well... Is there a way to
>>> improver the cw code reader / decoder in the K3 ?
>>> Thanks, Best 73' Joe R. / W7Joe
>>>
>>>
>> ______________________________________________________________
>> Elecraft mailing list
>> Home: http://mailman.qth.net/mailman/listinfo/elecraft
>> Help: http://mailman.qth.net/mmfaq.htm
>> Post: mailto:[hidden email]
>>
>> This list hosted by: http://www.qsl.net
>> Please help support this email list: http://www.qsl.net/donate.html
>>
>>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html