[Elecraft] K3--Too much ALC creates energy on opposite sideband?

[Wes]

This thread instigated some experiments here.  In analyzing TX IMD in both my K3 
and K3S, I've made probably hundreds of two-tone measurements using the built-in 
two-tone generator.  But in all of these measurements I never questioned what 
the audio level was, since it's not adjustable, except for balancing the levels 
of the tones. So I set up a measurement and went to transmit and then using the 
K3 utility used the get "BG" command to read the ALC bar graph.  The answer was 
zero bars.  I found this interesting.

To explore further, as a proxy audio signal generator, I opened WSJT-X, FT-8 
mode and used the "Test" button to generate an audio tone.  I used both the K3S 
internal sound card and the built-in card in a Lenovo Think Pad as a double 
check against generator distortion.  I saw similar results with both.

And what are those results you ask.

Using a spectrum analyzer and looking at the K3S output on 14 MHz in this case, 
using a low tone (500+ Hz) I noted that at any audio level that drove the ALC 
indicator to four bars (even flickering), there was low level garbage that 
appeared in the TX passband.  It took reducing the level to three bars to clean 
it up and in my radio there are either three bars or none, no one or two, when 
reducing the gain.  The "BG" command confirmed this.  Above four bars, things 
only got worse. Understand this was low-level stuff down in the weeds and not 
the dreaded audio harmonics that WSJT-X "split" is supposed to prevent.

I've never been one of those handwringers worrying about using ALC, as I believe 
just like AGC in receivers that we all use, properly designed it should work 
well for its intended purpose. Now, I'm not so sure, perhaps I've been wrong or 
I'm saved by the "properly designed" clause.

Wes  N7WS



On 8/28/2019 2:22 PM, Fred Jensen wrote:
> Yes, I know what the manual says.  However the IMD on my K3 goes up when I set 
> the mic gain to flicker bar 7 on voice peaks so I run it with almost no ALC.  
> I also set the TX EQ using the "K9YC Method," and regularly get unsolicited 
> "great audio" comments so I'm a fairly happy ham.
>
> I'm not on SSB often, but when I do get on such as NAQP SSB, I'm always a bit 
> surprised at how many signals have all the sideband energy crammed right up 
> against the suppressed carrier.  Not too many fill up their BW.  For those of 
> us with compromised hearing, it makes copy difficult ... one of the several 
> reasons I don't get on phone very often.  All that low frequency energy 
> contributes almost nothing to intelligibility.
>
> 73,
> Fred ["Skip"] K6DGW
> Sparks NV DM09dn
> Washoe County
>