[Elecraft] K3 AGC THR doesn't go low enough

[David Gilbert]

Don,

That's a good explanation.  Hopefully those folks will read it however 
many times it takes for them to understand that lowering the threshold 
is the opposite of what they want to do.  I suspect that MANY of the 
complaints about the K3 being a "noisy" rig are due to similar 
misunderstandings of how AGC actually works.  It appears that many folks 
think of AGC as being only gain reduction, and while that's probably 
true technically, it's probably much better to think of AGC as being 
gain equalization when you're trying to use it.

For anyone still unclear about how the AGC in the K3 works, I also 
recommend that  study Jack Smith's page on the subject.  It's a great 
explanation with several very informative graphs, and not surprisingly 
it is the very first link that is displayed when you Google "K3 AGC".   
The page doesn't appear to have been updated to include the newer "SOFT" 
AGC mode in the K3, but even so it is highly relevant.   
http://www.cliftonlaboratories.com/elecraft_k3_agc_and_s-meter.htm

I do a lot of contesting so I don't even want all signals to sound the 
same, and I certainly don't want the distortion issues that can 
sometimes be caused by AGC attack/decay slopes in the presence of 
multiple in-band signals.  Linearity is my friend, and AGC is by 
definition not linear over the range of interest.   I have my AGC THR 
set to 008 and I have my AGC SLP set to 000.  Jack's curves show that at 
even those settings, there is still some AGC action at low signals and 
protection at high signals.  For example, his data shows that even with 
AGC SLP = 000 it takes a 10 db increase in signal strength to give a 4 
db increase in audio level.  That works fine for me ...

73,
Dave   AB7E



On 10/8/2010 11:55 AM, Don Wilhelm wrote:
>    I seem to have trouble convincing some to raise the threshold to
> combat the noisy effect.  Please read all below carefully, and I believe
> I can offer some explanation.
>
> Let's study what the AGC SLP and AGC THR settings do and try to relate
> them to the perception that the K3 is noisy.
>
> IF --
> the threshold is so low that the noise triggers the AGC (and that will
> be happening if the noise is causing the S-meter to have a reading with
> no signal)
> AND --
> If the slope is set to a high value which makes all the signals have
> more or less the same audio volume,
> THEN --
> The noise will appear at about the SAME audio volume as the signals.
> RESULT --
> A K3 that sounds noisy :-( !  It is a combination of both the threshold
> and the slope that create this condition.
>
> The cure for that situation is to increase the threshold and reduce the
> slope.
> How much?
> That is an individual decision, but I have mine set to the extremes of
> THR at 008 and SLP at 002.
> I really believe the defaults are OK for a band where the ambient noise
> level is lower, but certainly gives problems on a band that is
> inherently noisy.
>
> Please make your evaluation while listening to signals and the breaks
> between signals - if you are hearing only noise, you will naturally turn
> the AF Gain up, and yes the noise will increase - not only that, with
> only noise, the perception is that one should use a lower threshold setting.
> That is because as you lower the threshold, the AGC response to the
> noise gets stronger, and the audio level is reduced by the AGC - what
> you are perceiving is  exactly the opposite of what you wish to achieve.
>
> In addition, turn the attenuator on and the preamp off - again, if you
> hear an increase in the band noise when the antenna is connected, you
> have all the sensitivity that you can use for that band and antenna.
> The K3 remembers those settings by band, so you don't have to make
> changes as you switch bands.
>
> 73,
> Don W3FPR