Arie,
Sorry about the added "n".
Yes, the DSP algorithms will not allow the passband to extend into the
opposite sideband.
What I was responding to was the initial statement about "centering" the
passband (on the CW sidetone pitch), and I was pointing out the fallacy
of that when the filter is wide.
The analog filters at the 8 MHz have only the task of keeping strong
nearby signals out of the ADC and not overloading it - they also reduce
the exposure of 'pumping' for the hardware AGC by those strong signals
near in frequency (such as may be encountered in a DX pileup or contest
conditions).
The DSP does the final filtering, and the DSP filters do not actually
need the roofing filters at all other than for the reasons stated above.
73,
Don W3FPR
On 2/6/2014 12:17 PM, Arie Kleingeld PA3A wrote:
Don,
I understand this.
Let me explain my thoughts.
Let's suppose we have the different IF's. One analog, the 8Mhz stuff,
and then a "DSP IF" which is fed by the second 15kHz IF.
I am not a DSP expert. But I would expect that the DSP filter would
filter out whatever bandwidth I would want it to, just like it were
analog and better. I expect this function of any IF filter, digital
and analog.
What we are tallking about is that the opposite sideband can pass
through the analog part of the K3 (because of the wide roofing filter)
and yes it does if the skirt of the filter allows it. If this is not
solved in the DSP, the /function/ of the DSP is not more than the old
type audio DSP filter we had in the old days (apart from partly being
inside the AGC-loop)
As Ralph DL5OAP already stated: Even the old FT1000MP with DSP
demodulation killed the opposite sideband. I had an MP for a long time
(before I got the K3) and I can confirm this. I had it switched on all
the time.
As you said Don, the K3 really needs the analog roofing filter to be
in place to reach its best performance. Good thing that I have them
for the bandwidths I use. And overall it works pretty well. There are
still some things to improve though.
73
Arie
(not Arnie :-))
Don Wilhelm schreef op 6-2-2014 17:21:
Arnie,
Think about it this way -- take a filter width of 1500 Hz and a
sidetone pitch of 600 Hz.
If you center the passband at 600 Hz, the upper passband limit will
be at 1350 Hz, but the lower edge will be 150 Hz into the opposite
sideband (single signal reception goes "bye-bye").
The proper position if the passband is to place the lower edge at
about 100 Hz, and the upper edge will extend to 1600 Hz.
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft@mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
