6300 CW filter function at narrow settings

  • 1
  • Question
  • Updated 5 years ago
At 50Hz., does anyone know what the function of the 6300's CW filter is, Bessel, Butterworth, number of poles, Etc? I find the Flex audio to ring so much at this narrow setting compared to my TS-590 that copy is very difficult with the Flex while very smooth on the Kenwood.
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes

Posted 5 years ago

  • 1
Photo of Tom Warren

Tom Warren

  • 59 Posts
  • 24 Reply Likes
Bob,
    It will be interesting to hear what kind of reply you get since I've mentioned this in the past and sure hope the ringing is improved in a major way with 1.4. I try to use the very narrow filter settings while doing weak signal work with my 6300 with 2M and 70cm transverters and in comparison to the F5K, it seems (by ear) the 6300 filters are not as quiet.
Tom   

Bob wrote:
  At 50Hz., does anyone know what the function of the 6300's CW filter is, Bessel, Butterworth, number of poles, Etc? I find the Flex audio to ring so much at this narrow setting compared to my TS-590 that copy is very difficult with the Flex while very smooth on the Kenwood.
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Gentlemen,  especially Tim & Ken,

Thanks for your help...I was indeed mistaking too much gain ahead of the 50Hz CW filter for filter ringing.  The filter, of course, picked out the "on-frequency" components in the noise and presented the results to the loud speaker which sounded to me, at the time, like filter ringing. 

The question that I have now have to answer is whether adjusting the AGC down to mid-range should be my default setting and what the impact is going to be on hearing the signals I am trying to discriminate.

Also Tim, if you could pursue the answer to my original question as to the filter characteristics, I would appreciate knowing this information. Had I had this information I would have come to the proper conclusion with a lot less confusion.  I could run a sweep of the filter but doing that through all the other DSP functions may mask the true nature of the filtering itself.

Thank you and warm regards,

Bob, W7KWS
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4185 Posts
  • 1333 Reply Likes
If you adjust your AGC-T down to the "Sweet-spot" or "Knee" which is where the background noise just begins to drop, you will not miss any stations.  This control adjusts the threshold at which the AGC begins to drop the gain.   Anything higher than the "knee" doesn't add any effective gain because it is all noise anyway.  If you are on a noisy band with nothing but strong stations, drop it down some more.

This adjustment will vary depending upon the band, antenna, band conditions, , filter width, RX EQ settings, APF settings, and static level at that particular time of day.   This is why many of us have one of the buttons on our FlexControl Knob set up to control the level of AGC-T.  It becomes second nature to quickly adjust AGC-t to maximize the Signal to Noise Ratio.  
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4189 Posts
  • 1334 Reply Likes
The key to eliminate "ringing" at narrow bandwidths is judicious adjustment of the AGC-T.  Find a clear frequency and begin adjusting the control to the left until the noise just begins to drop.  A little further for stronger signals.  It will make the signal begin to pop out of the noise and it will keep the background noise from ringing in your ears.  The rig was not designed to run with AGC-T at full open.  it is an unnecessary noise source.
BTW.  I always listen to CW for the AGC in "Fast" mode.

Also, if you want better performance, adapt to a CW offset of 500 Hz and set your RX EQ with the 500 Hz up all the way and all the others down all the way.  It will add some audio channel filtering and reduce noise in that arena also.

When I get really weak stations, I turn on APF, but always need to adjust the AGC-T to maximize the signal to noise ratio. 

Once you learn how to adjust these controls, it is a very quiet rig, even at 50 Hz filtering.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9186 Posts
  • 3541 Reply Likes
The DSP filters cannot ring. Do you have APF enabled with a narrow CW filter?  If so disable it.
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Hi Tim. Thanks for your interst. I'll check that setting but I think that you will find after doing some research that DSP filters are just mirrors of their hardware equivelant and if the design says mathimatically that it will ring... Ring it will.
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Tim...you are probably closest to being in a position to obtain the answer to my original question. That answer would be very helpfull to me in sorting this out.
Namely, things like filter type or passband ripple and phase response, number of poles, their Q & positions when the 6300 is set to 50Hz.

Thanks
(Edited)
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Hi Ken,


Thanks for your input but it seems to me that all that adjusting for every signal is a lot of unnecessary work when I'm busy with other aspects of the communication I'm trying to complete.


Wouldn't a better solution be to consider if a more optimum filter could be designed for 50Hz? After all, it's only software. Maybe one with a more linear phase response and some addational poles to compensate for the reduced shape that would result. Might also consider if there are more optimum pole positions for a better result at 50 Hz. bandwidth.


Wish I were a software guy but all my filter work was hardware where it was too expensive for this kind of thinking, and that was so many years ago.


73
(Edited)
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9183 Posts
  • 3529 Reply Likes
In CW, you have to balance creating a super sharp filter vs latency due to the QSK requirements.
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Hi Tim,

Maybe at narrow bandwidths, and to accommodate communications in an environment which dictate using such a narrow bandwidth, QSK isn't so important as hearing the signal in the first place.  Thus, a different design for the filter at this bandwidth and not worrying about QSK would be a more optimum approach to completing the communication at hand.

However, I doubt that the minimal group delay, through this type filter, will hinder QSK but I'm not experienced at that type of CW so I might not have a good handle on QSK requirements.

Thanks,

Bob
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1842 Posts
  • 669 Reply Likes
Hi Tim,
Are you suggesting that the 50HZ filter could be improved when not running QSK?    

Would it be possible to enhance the logic to detect whether QSK is in effect and if not, switch to the enhanced filter mode?

Would the non-QSK enhanced mode also benefit the other filters 100HZ, etc?

To be clear, I'm not suggesting the filters aren't very good now.  They work great  compared to those in other radios I've used.  And I also get good results with the 50HZ filter but I'm always interested in ideas to make things even better.    

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com 
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9183 Posts
  • 3529 Reply Likes
No, I am saying that by not using APF with narrow CW filters it cuts down on the noise inside the filter.  This reduces the audio effect that is being reported as "ringing".  The number of taps for the filters is fixed no matter if QSK mode is enabled or not.  CW is a very strict timing mode, regardless of QSK and the filters have been optimized to balance latency verses sharpness while taking into account the computational overhead to pass the signal through the DSP filter.
(Edited)
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4189 Posts
  • 1334 Reply Likes
If I Remember Correctly, Steve once said that the narrow CW filter is done with a fourier transform calculation that does not ring.  but what appears to be "ringing" is the left over noise in the narrow passband, thus the need to adjust AGC-Threshold correctly.
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1848 Posts
  • 670 Reply Likes
Ken,
I read the same comment from Steve.  And agree that the AGC-T setting is a key to reducing the noise and reducing the artifacts that can sound like ringing.

Bob,
The optimum AGC-T setting changes based on the band, antenna direction (if you have a rotatable antenna), propagation at the moment, etc.   It can require a lot of attention.  There is an idea posted to automate the AGC-T setting.  Check it out and if you like it, you can add a vote.

https://community.flexradio.com/flexradio/topics/option-to-automate-the-agc-t-setting

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com


  
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Ken,

What you suggest may be a solution but one that I supect may be sub-optimum. I've designed very narrow signalling filters for mobile telephony, some on the order of 10Hz. These filters had no ringing but they did pull out frequency noise components out of FM squelch noise and the result sounded a bit like ringing.


Since a narrow filter will pick out its natural frequency from noise and can sound like a weak signal, amplitude control can moderate how much noise is picked out like this.


If this were my problem I would expect the Kenwood to act simillarly but it does not. I will make some AGC adjustments, however just to see if the default Kenwood AGC setting is set perfectly just by a rare coincidence.


Thanks, Bob
(Edited)
Photo of Steve - N5AC

Steve - N5AC, VP Engineering / CTO

  • 1053 Posts
  • 1076 Reply Likes
The CW filter for 50Hz has 2048 taps and it is a Butterworth filter.  These are EXTREMELY sharp filters, but they are FIR and not IIR so there is no ringing.  Because you restrict all energy to the 50Hz region of interest, the only sound you hear will be tones in that range of frequencies.  For example, if your CW pitch is 600Hz, you will only hear sounds from 575-625Hz.  If you increase your noise level where you hear noise, all of the noise will be colored so that you only hear those frequencies.  This is not "ringing."  Ringing is where the filter has feedback inside the filter and an impulse to the filter can continue to echo in the filter long after the input stimulus has been removed.  Also, there is no limit to how narrow we could make the filter.  I think we arbitrarily limit to to 50Hz in the presets and 10Hz in the manual adjustment.  They would work just fine at 1Hz or 0.1Hz, but I don't think you would want to listen to them ;-)

There are several approaches: 1) reduce the level of the noise where you are not listening to noise bounce in and our from the frequencies of interest [adjust AGC] or 2) widen your filter so that your ear [brain] does the DSP and separates the tone from the surrounding noise which has many frequencies in it.  For me, turning up AGC-T and listening in a narrow filter is difficult especially in a low SNR environment.  3) widen the filter and use APF to boost the amplitude of a particular set of frequencies.  This provides the soothing background noise and pulls the the CW tone higher out of the noise.  Which of these you use will depend on what you are used to and what you prefer.  I have friends that prefer to decode CW in a full USB bandwidth (2.6+kHz).  This is too much noise for me, generally, but I do understand why they like it.
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
Steve, thanks for this detail. It is very helpful to my understanding.

Best regards,

Bob, W7KWS
Photo of Bob - W7KWS -

Bob - W7KWS -

  • 321 Posts
  • 52 Reply Likes
By the way Steve, sometime ago, maybe the mid 1960s, Mal Mobily, a broadcast engineer from L.A. came to Seattle to measure co-channel interference beteen KMPC in L.A. & KIRO here in Seattle, both on 710kHz. He had one station move up 5Hz & the other down 5Hz and he flew around in an airplane with a 1Hz filter in his receiver and ploted signal strength contours for each station. 1Hz is a most valuable tool sometimes. ;-)

Bob, W7KWS