[CW] N4YG de N1EA

[Joe]

Hi David,

It is probably not proper for me to get into this, but I just had to make a few comments. You can pass them on if you wish, but there are several errors in PA0WV's comments.

First, what it sounds like is all that is important.

Secondly, the amount of time for the shaping of the envelope depends on the frequency of the tone. Yes, if you want to reduce the transient effects in keying a tone, it takes some time. If you turn it on in zero time, there will be clicks. But the time it takes to do this is dependent on the tone. If a person is going to operate at higher speeds he has two choices, reduce the time for shaping the leading and trailing edges and accept the unpleasant sound or increase the sidetone frequency. Increasing the sidetone frequency for higher makes all the sense in the world. (And, by the way, the carrier is not 680 Hz, it is variable up to perhaps 1300 Hz.) If one is going to operate at 50 WPM, a nominal dot will have a duration of only 24 mS. If one uses a sidetone frequency of 500 Hz, there will be only 12 cycles of the sidetone for a single dot. I have found that it takes several cycles to shape the leading and trailing edges to produce a pleasing sound. This means that one is left with something like 8 cycles of relatively high amplitude sidetone - not so good. A better choice is to go to 1000 Hz and get twice as many cycles. I know, I like a lower tone, but I do not like clicks. There must be a tradeoff here.

You can be assured that beginning and ending the sidetone at zero phase is of value. All you have to do is consider a fourier series expansion of a string of dots which begins and end at 90 degrees phase. This will cause a very severe pop. I have not tried it, but I am going out on a limb and say that I do not think one could detect the difference between a phase discontinuity between adjacent dot of 90 degrees as compared to a situation where the dot length was chosen so that there was not dicontinuity.

Regarding, the D/A resistor ladder, 1 percent resistors are used and supplied with the kit. Again, I am going out on another limb because I have not tried this but I do not believe using 5 percent resistors would make a significant difference in the sound.

Thanks for your interest.

Joe, N4YG
David,I downloaded the track.

It sounds good, BUT

The slope takes 18 ms. That means that the minimum dit length is 36 ms, 
because for shorter dits the trailing slope starts earlier then the 
leading slope ends.

Dits of 36 ms means for that reason that the max speed is 33 wpm.

Furthermore some additional remarks:

The author wrote that his carrier of 680 Hz is not phase continue but 
starts with each dit and dah with zero phase. He claims that will 
additional degrade clicks.

That is not true. Morsecode is AM 100% modulated om the carrier of 680 
Hz. With a string of dits, at 33 wpm you got a sinewave envelope when 
the slope is 18 ms, hence only two sidebands at 680 plus and minus 27.8 
Hz. However when you start each dit with zero phase you get additional 
phase jumps in the signal which widens the spectrum.

There can not be advantage of zero phase starting, because you add 
unnecessary phase modulation on the carrier of 680 Hz.

Furthermore: He produces a sinewave with a set of resistors in the range 
R up to 128 R

When those resistors are 5 percent the 5% of the largest (128R) is 6 
times larger than the smallest resistor (R) in nominal value is.
When you pass digital level 01111111 to 10000000 which is only the step 
of decimal 127 to 128 you get a spike in the output that ruins your 
sinewave and also your slope, with 10% of the max amplitude in worst 
case with 5% resistors

So those resistors have to be carefully matched up to 0,04%, and I 
suspect that a R-2R network which contains only resistors R and 2R is a 
much better choice, but still requires matching.

PA0WV

----- Original Message ----- 
  From: D.J.J. Ring, Jr. 
  To: Joe 
  Sent: Monday, August 12, 2013 8:10 PM
  Subject: Re: [Spam?] N4YG de N1EA


  Joe,


  Move the link down with the other links or center it.  It is difficult to find.  I knew it was there and still I couldn't find it.


  Put a "new line" under it at least :-)


  Nice sounding I will tell the list.


  DR


  On Mon, Aug 12, 2013 at 6:51 PM, Joe <joe at n4yg.com> wrote:

    Hi David,

    Done. You should see it on the DDS Sidetone page. You may have to refresh the page.

    Joe
      ----- Original Message ----- 
      From: D.J.J. Ring, Jr. 
      To: joe at n4yg.com 
      Sent: Monday, August 12, 2013 4:15 PM
      Subject: [Spam?] N4YG de N1EA


      Hello Joe, 


      Some interest was shown concerning your module for sidetone and code practice oscillator on the qth.net CW reflector of which I am co-owner.


      i was wondering if you could post to your web site a very short recording of the code practice oscillator as some express doubts that a logarithmic slope produces a click free sound and that a cosine squared slope is needed.


      I'm glad you got some publicity on the CW reflector.


      David 

      David J. Ring, Jr., N1EA  Radio-Officers Group -- Join CW email list -- Historic Morse Recordings 
      Gopher Hole:  gopher://sdf.org/1/users/djringjr/ (native or with Firefox's Overbite extension) or via http to gopher gateway
      Chat Skype: djringjr MSN: djringjr (@) msn.com AIM: N1EA icq: 27380609


      =30=

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/cw/attachments/20130813/8e2cdccb/attachment.html>