LCW and UCW, not just CW. Band Pass picture is incorrect

  • 2
  • Problem
  • Updated 5 years ago
The bandpass sketch has a small line at the tuned frequency, inferring that it is symmetrical around that frequency. which is not true. While this does not affect operation on ssb (there are + and - signs to guide you, to where you want to be, either all above the frequency, or all below the frequency. Upper or lower sideband.

However, on CW, the bandpass really has been placed symmetrical around the the
tuned frequency, instead of around the desired frequency, which is above or below that frequency by the preset "tone freq." In other words, at 400 Hz selectivity, it should not be +/- 200 Hz, but properly be +400 / +800 Hz or -400 / -800 Hz, depending upon whether you want to listen to the upper or lower sideband. This assumes you have selected 600Hz tone freq..

This is why increasing selectivity does not work. The filter is probably working fine, but at the wrong frequency. And this is why the receiver needs UCW and LCW.

Tim: lets get together . I'm less than an hour away.
Photo of mike_w1bfn

mike_w1bfn

  • 9 Posts
  • 1 Reply Like

Posted 5 years ago

  • 2
Photo of Stu Phillips - K6TU

Stu Phillips - K6TU, Elmer

  • 642 Posts
  • 256 Reply Likes
Mike,

This isn't the cause of filter selectivity in the 0.13.10 - the number of taps in the DSP filter are the known reason for this and are being increased.

The filter is correctly symmetric about the indicated frequency.

Stu K6TU
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4025 Posts
  • 1248 Reply Likes
Stu, Do you know if they are considering an adjustable parameter to adjust the sharpness of the filter skirts, especially on CW? There are times I want a sharp filter and other times I would settle for a little less with less ringing.
Photo of Stu Phillips - K6TU

Stu Phillips - K6TU, Elmer

  • 642 Posts
  • 256 Reply Likes
Sorry I don't...

Tim - is there a US on this?
Stu
Photo of Steve - N5AC

Steve - N5AC, VP Engineering / CTO

  • 1031 Posts
  • 1002 Reply Likes
The filters in the FLEX-6000 are finite impulse response filters which do not ring. We have dramatically increased the sharpness of the filters recently (0.14.x as I recall). We have discussed making the skirt steepness selectable, but again ringing would not be a side effect with FIR filters. The key side effect of a sharper filter is increased latency. Filter latency in 0.14.x is about 43ms.

CW works a little different in SmartSDR than in PowerSDR. In PowerSDR, we placed the carrier indicator where the CW signal would be IF the tone frequency was 0Hz. Since it never is, you see the signal of interest at ~600Hz or whatever tone frequency you selected. This makes the actual carrier indicator, which is a red line, misleading. So in SmartSDR, the carrier line is where you will transmit and where you will hear the other station. The sides of the filter show how far away from the desired signal your skirts are. But, you hear everything offset by your selected CW pitch so if you select 600Hz, everything on the red line is at 600Hz audio frequency.

Steve
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4025 Posts
  • 1248 Reply Likes
Thanks, Steve. That is good news. I can't wait to hear the new filters.

Re: the offset tones, it really makes no difference whether you are USB or LSB, except if you are using a wide filter and trying to do cross-mode with a station on SSB, or using a Digi CW program that expects the shift in tone to follow one standard direction or the other. Then it would make a big difference.

I think that may be why some have asked for a CWU & CWL option, so the Digi programs don't get screwed up.