[K3] autonotch now distorts SSB audio

[inventor61 .]

Hello


I noticed some unusual distortion of SSB phone (voice) signals this evening
when using autonotch on my K3

Running 4.60 uC, 2.80 d1/d2, FL1.15, dr 1.03

I have noticed minor coloration of the signals in the past when using
autonotch, but not this manner of change, until after I recently upgraded
the firmware.

Could it be something I'm doing?


Steve KZ1X/4
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Bill Clarke]

You are not alone regarding problems with ANF (NTCH) distortion of SSB
received signals.  Too bad the problem continues unaddressed - as it is
the single remaining receive issue I have with my K3.

There is a thread about it here:
http://www.mail-archive.com/elecraft@.../msg136366.html

A couple of responses were actually explanatory and helpful - others
were laughable detractors. The thread is worth reading. Some responders
say they have no problem with the distortion and find it acceptable. I
personally do not accept it, as it detracts from an otherwise nearly
perfect receiver.

As an interim fix, I suggest you reduce the AGC SLP to a low number,
such as 2 or 3 (the settings I use).

Bill W2BLC


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[David McClain]

Hi,

The AutoNotch on the K3 is too effective! It’s learning rate is too rapid, and that causes it to lock on and off of rapidly varying voice peaks during SSB phone reception. The effect sounds much like someone overdriving their mic input at the transmitter.

Last night I was listening in headphone diversity mode, with my K3 on the left and an IC-7700 on the right, to some SSB comms. The left ear (K3) sounded crackly and distorting, like the op was overdriving his transmitter, but the right ear (IC-7700) sounded just fine. I removed diversity by listening only to one receiver or the other. The IC-7700 continued to sound okay, but the K3 was gritty and abrasive, and I could not tolerate listening for any long duration.

So I started tracking things down…. I normally listen to the SSB subbands with the auto notch enabled to avoid all the tuner uppers, and with the NB enabled at a very low setting just to get rid of incidental impulses. After disabling the NB no change occurred.

I then examined both the K3 and the IC-7700 audio output on an audio spectrum analyzer. There were no obvious indications of IMD in either of them. So perhaps I need to feed 2 tones? I did so, and still no IMD. I tried lowering the Line Output setting, fearing that I might have it too high and I was inducing harmonic distortion, but the spectrum analyzer showed nothing down to -60 dBc.

Just by chance, I removed the AutoNotch while listening to some more SSB phone, and immediately noticed that the speaker’s voice quality dramatically improved. Aha! The AutoNotch is too quickly adapting to varying voice peaks, and the sudden disappearance / appearance of nearby treble peaks in the voice was causing a crackling sound in my headphone.

I notice that on heterodyne tones the K3 almost instantly removes them, while the IC-7700 takes a fraction of a second longer, perhaps 100-200 ms. But that sluggish behavior in the IC-7700 is not objectionable, and apparently is long enough to avoid it training quickly on voice peaks.

73 de Dave, N7AIG

David McClain
[hidden email]



______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]