[Elecraft] Re: Displaying the keying waveshape on Spectrogram.

[Masleid, Michael A.]

Hello Mike,

>Dale Boresz wrote...
>>After reading Guy's (K2AV) excellent explanation of measuring the K2's =
keying
>>waveform with Spectrogram, I decided to measure my own K2, using the
>>ICOM IC-746PRO as the monitoring receiver.
>>
>>The Spectrogram plot of the results can be viewed at:
>>http://64.105.107.250/kqrv/k2keying.aspx

Mike wrote...
>Without the keying speed being specified, the results aren't very =
meaningful. The amount of energy in the
>sidebands will vary with keying speed.

It is necessary to be precise:  We can talk about power (watts) or =
energy (joules) here.  It makes a
difference.  You have a point, but it is easy to confuse.  The amount of =
energy in a click doesn't change
with keying speed.  The number of clicks per second increases with =
keying speed, so the amount of
power in the clicks will increase.  The amount of power in the dots =
should not increase with keying speed
unless something odd happens to the dot space ratio.  Over a useful =
range of keying speed, say 15 to 30 WPM,
the amount of power in the click will double - a change of 3 dB.  Dale's =
measurements are at 30 dB per
division, so that change could hardly be noticed, even if there really =
was a change.

Since Dale was measuring peak power levels, things get complicated very =
quickly.  Roughly speaking, the
energy in a click, divided by the duration of the click, gives the power =
- during the click.  This doesn't
change with keying speed either.  So, depending on how spectrogram =
determines peak power, it might still
report no change as keying speed changes.

The devil in the details has to do with how impulse type noise, like a =
click, gets processed by a filter.
The amount of power in a pure tone should not change with bandwidth, =
although doubling bandwidth might
double the number of tones that you hear - so you might get 3 dB more =
power by doubling bandwidth.  With
something like a click, doubling bandwidth does something strange.  =
Since you are getting the click twice
at the same time, so to speak, you are getting the voltage twice at the =
same time.  Since doubling the
bandwidth doubles the voltage for impulse noise, the power goes up by 4 =
times (volts squared), or 6 dB.

The moral of the story:  Yes, dots per second changes the amount of =
power in the clicks, but it doesn't
change the peak power in the clicks.  The apparent peak power in the =
clicks depends on the bandwidth
of the filter being used to measure the power.

Narrow filters =3D long sample time =3D average power =3D clicks go =
away.

Wide filters =3D short sample time =3D more of the peak power =3D clicks =
get bigger.

Really wide filter =3D shorter sampling time than the click =3D all of =
the peak power =3D big as it gets.

I guess this means that there is a perfect receive bandwidth to make a =
given rig's clicks look their
worst compared to another's.

On our ears?  Double the width of the filter doesn't increase the amount =
of signal, but it does double
the amount of thermal noise, and it should re-double the amount of =
click, up to the point.

73, de Michael AB9GV