Elecraft

[K3] AGC White Paper

Classic

List

Threaded

24 messages Options

alorona

[K3] AGC White Paper

I spent a little bit of time this weekend and put together yet another K3 AGC (YAKA) "white paper" to put some measurements and discussion of the results down in the hope that others may benefit. Maybe it'll help folks understand AGC better by demystifying some of the K3 idiosyncrasies. I hope it makes you think about a few things in a way that maybe never occurred to you. I put the document in my Dropbox at the following link:

https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%20K3%20AGC%20System.pdf?dl=0

I'll leave the file there for several days in case any of y'all are interested.

73,

Al W6LX
______________________________________________________________
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]

Bill Frantz

Re: [K3] AGC White Paper

Thanks Al. A very useful paper.

Now to use SLP=15 for digital, where I want the AtoD to not have
to deal with a big range, while using something with more
dynamic range for voice/CW.

73 Bill AE6JV

On 3/6/17 at 9:50 PM, [hidden email] (Al Lorona) wrote:

>I spent a little bit of time this weekend and put together yet
>another K3 AGC (YAKA) "white paper" to put some measurements
>and discussion of the results down in the hope that others may
>benefit. Maybe it'll help folks understand AGC better by
>demystifying some of the K3 idiosyncrasies. I hope it makes you
>think about a few things in a way that maybe never occurred to
>you. I put the document in my Dropbox at the following link:
>
>
>https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%20K3%20AGC%20System.pdf?dl=
>0
>
>
>I'll leave the file there for several days in case any of y'all are interested.

-----------------------------------------------------------------------
Bill Frantz | Security is like Government | Periwinkle
(408)356-8506 | services. The market doesn't | 16345
Englewood Ave
www.pwpconsult.com | want to pay for them. | Los Gatos,
CA 95032

______________________________________________________________
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]

OE5CSP-Chris

Re: [K3] AGC White Paper

In reply to this post by alorona

Thank you for your great explanation!!
I recently came across an article on the internet( can´t find it right now, maybe it was in German) about the biological attenuation of the inner ear. The author says that at a certain point the biological attenuation kicks in to safe your ears.It also kicks in before you start to talk but then you loose lots of dbs. So I´m trying to find a setting where I get a maximum of dynamic range before my personal attenuator is engaged.When I was younger, this attenuator worked quite well in the gym, now I wear hearing protection all the time(I´m a gym teacher). A setting of SLP 05 and THR 14 is too aggressive for my ears.
73,Chris-OE5CSP

Cady, Fred-2

Re: [K3] AGC White Paper

In reply to this post by alorona

Very nice Al Thanks.

73,

Fred KE7X

________________________________
From: Elecraft <[hidden email]> on behalf of Al Lorona <[hidden email]>
Sent: Sunday, March 5, 2017 10:50 PM
To: Elecraft Reflector
Subject: [Elecraft] [K3] AGC White Paper

I spent a little bit of time this weekend and put together yet another K3 AGC (YAKA) "white paper" to put some measurements and discussion of the results down in the hope that others may benefit. Maybe it'll help folks understand AGC better by demystifying some of the K3 idiosyncrasies. I hope it makes you think about a few things in a way that maybe never occurred to you. I put the document in my Dropbox at the following link:

https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%20K3%20AGC%20System.pdf?dl=0

I'll leave the file there for several days in case any of y'all are interested.

73,

Al W6LX
______________________________________________________________
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]

______________________________________________________________
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]

Sverre Holm (LA3ZA)

Re: [K3] AGC White Paper

In reply to this post by alorona

Very nice report - thanks!

Sverre, LA3ZA

K2 #2198, K3 #3391,
LA3ZA Blog: http://la3za.blogspot.com,
LA3ZA Unofficial Guide to K2 modifications: http://la3za.blogspot.com/p/la3za-unofficial-guide-to-elecraft-k2.html

Wes Stewart-2

Re: [K3] AGC White Paper

In reply to this post by alorona

In Al's very nicely done paper he has this paragraph:

"The purpose of Automatic Gain Control (AGC) is to reduce the range of the
signals seen by the sensitive stages in the receiver. The AGC stage is
designed to vary its gain depending on the input signal; stronger signals
get less gain, and this has the effect of compressing the amplitude range.
This is the desired response."

I believe this paragraph and the accompanying graphic can be misleading to the
unwary. AGC does not compress the range of signals, it simply lowers the gain
through the receiver. The range (difference between) signals might well be 130
dB at the input but it better be 130 dB everywhere else in the receiver too. If
lower level signals are driven into the internal noise level because of gain
reduction, so be it; that should be the only reduction in range.

Al continues:

"But signals above this threshold will be acted on by the AGC. Even though
in real life an S9 signal is 5 S-units stronger than an S4 signal, because
of the AGC it will sound only 11.1 dB louder – less than 2 S-units louder.
This is because, reading from the above graph, an S4 (-103 dBm) signal
produces -15.3 dBV of audio output and an S9 (-73 dBm) signal produces -4.2
dBV, a difference of 11.1 dB. A five S-unit difference has been reduced to a
less-than-two S-unit difference."

Some are going to read this and mistakenly believe that while receiving both an
S9 signal and an S4 signal, AGC is going to reduce the ratio between them from 5
S-units to two S-units.

Hence, I suspect that "mush" the proponents believe that AGC somehow magically
reduces the amplitude of the stronger signal that activates the AGC but allows
the weaker ones to have full gain, so that they "catch up" and become
indistinguishable from the stronger one. That's a limiting receiver, nice for
FM, not so nice for CW/SSB.

All IMHO, of course,

Wes N7WS

On 3/5/2017 10:50 PM, Al Lorona wrote:

> I spent a little bit of time this weekend and put together yet another K3 AGC (YAKA) "white paper" to put some measurements and discussion of the results down in the hope that others may benefit. Maybe it'll help folks understand AGC better by demystifying some of the K3 idiosyncrasies. I hope it makes you think about a few things in a way that maybe never occurred to you. I put the document in my Dropbox at the following link:
>
>
> https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%20K3%20AGC%20System.pdf?dl=0
>
>
> I'll leave the file there for several days in case any of y'all are interested.
>
> 73,
>
> Al W6LX
> ______________________________________________________________
> 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 [hidden email]
>

alorona

Re: [K3] AGC White Paper

Yup, I knew that would cause confusion. That's NOT what I meant. I meant that if you listen to ONE signal at a time, their apparent amplitudes do not reflect reality, because of the AGC. If BOTH signals are present in the passband at the same time, then my comments DO NOT apply. I was ONLY talking about hearing ONE signal AT A TIME.

NONE of my discussion EVER presupposed more than ONE signal being received at a time.

THANK YOU. :^)

I am beginning to believe that it's impossible to shed any more light on this topic. It continues to be one of the most misunderstood areas of all, and I'm not sure if any effort can clear up the confusion without adding more confusion of its own. At least, I know I've failed! The next time someone asks a question about AGC, I'm going to chomp down on a towel, duct tape my arms to the chair, and start my breathing exercises: deep breath.... hold..... exhale..... again.....

Al W6LX

________________________________
From: Wes Stewart <[hidden email]>
To: [hidden email]
Sent: Monday, March 6, 2017 10:19 AM
Subject: Re: [Elecraft] [K3] AGC White Paper

In Al's very nicely done paper he has this paragraph:

"The purpose of Automatic Gain Control (AGC) is to reduce the range of the
signals seen by the sensitive stages in the receiver. The AGC stage is
designed to vary its gain depending on the input signal; stronger signals
get less gain, and this has the effect of compressing the amplitude range.
This is the desired response."

I believe this paragraph and the accompanying graphic can be misleading to the
unwary. AGC does not compress the range of signals, it simply lowers the gain
through the receiver. The range (difference between) signals might well be 130
dB at the input but it better be 130 dB everywhere else in the receiver too. If
lower level signals are driven into the internal noise level because of gain
reduction, so be it; that should be the only reduction in range.

Al continues:

"But signals above this threshold will be acted on by the AGC. Even though
in real life an S9 signal is 5 S-units stronger than an S4 signal, because
of the AGC it will sound only 11.1 dB louder – less than 2 S-units louder.
This is because, reading from the above graph, an S4 (-103 dBm) signal
produces -15.3 dBV of audio output and an S9 (-73 dBm) signal produces -4.2
dBV, a difference of 11.1 dB. A five S-unit difference has been reduced to a
less-than-two S-unit difference."

Some are going to read this and mistakenly believe that while receiving both an
S9 signal and an S4 signal, AGC is going to reduce the ratio between them from 5
S-units to two S-units.

Hence, I suspect that "mush" the proponents believe that AGC somehow magically
reduces the amplitude of the stronger signal that activates the AGC but allows
the weaker ones to have full gain, so that they "catch up" and become
indistinguishable from the stronger one. That's a limiting receiver, nice for
FM, not so nice for CW/SSB.

All IMHO, of course,

Wes N7WS

On 3/5/2017 10:50 PM, Al Lorona wrote:

______________________________________________________________
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]
______________________________________________________________
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]

Don Wilhelm

Re: [K3] AGC White Paper

In reply to this post by Wes Stewart-2

Wes is correct. Let me say it in a slightly different way.
The AGC simply reduces the gain of the receiver - and it will respond
mainly to the strongest signal in the passband. Actually it responds to
the integral of all the signals in the passband, but if there is a
single stronger one it will predominate.

When the receiver gain is reduced, ALL the signals will be
proportionally reduced in amplitude - including the weaker ones. This
is often referred to as "AGC Pumping".

73,
Don W3FPR

On 3/6/2017 1:19 PM, Wes Stewart wrote:

> In Al's very nicely done paper he has this paragraph:
>
> "The purpose of Automatic Gain Control (AGC) is to reduce the range
> of the
> signals seen by the sensitive stages in the receiver. The AGC stage is
> designed to vary its gain depending on the input signal; stronger
> signals
> get less gain, and this has the effect of compressing the amplitude
> range.
> This is the desired response."
>
> I believe this paragraph and the accompanying graphic can be misleading
> to the unwary. AGC does not compress the range of signals, it simply
> lowers the gain through the receiver. The range (difference between)
> signals might well be 130 dB at the input but it better be 130 dB
> everywhere else in the receiver too. If lower level signals are driven
> into the internal noise level because of gain reduction, so be it; that
> should be the only reduction in range.
>
> Al continues:
>
> "But signals above this threshold will be acted on by the AGC. Even
> though
> in real life an S9 signal is 5 S-units stronger than an S4 signal,
> because
> of the AGC it will sound only 11.1 dB louder – less than 2 S-units
> louder.
> This is because, reading from the above graph, an S4 (-103 dBm) signal
> produces -15.3 dBV of audio output and an S9 (-73 dBm) signal
> produces -4.2
> dBV, a difference of 11.1 dB. A five S-unit difference has been
> reduced to a
> less-than-two S-unit difference."
>
> Some are going to read this and mistakenly believe that while receiving
> both an S9 signal and an S4 signal, AGC is going to reduce the ratio
> between them from 5 S-units to two S-units.
>

Jim Brown-10

Re: [K3] AGC White Paper

On Mon,3/6/2017 11:29 AM, Don Wilhelm wrote:
> This is often referred to as "AGC Pumping".

Yes, BUT -- there is distortion associated with AGC action, and that
distortion is what is perceived as "mush."

[Remember that the fundamental definition of distortion is "anything in
the output that's not the input multiplied by a constant. Distortion can
include time-variable gain, non-constant amplitude and phase response as
a function of time or frequency, and amplitude non-linearity, variable
time response. ALL of these distortions have the potential for confusing
the ear-brain when trying to get intelligence from a signal, or
providing aural satisfaction when listening to music. The K2 provides a
great example. It sounds great and audio is easy to copy when listening
to a SSB signal through the TX filter, but sounds awful and makes copy
very difficult when the crystal filters are stagger tuned to provide
narrower bandwidth. That's because the frequency response is pretty flat
for the TX filter, but looks like a side view of the Rocky Mountains
with the stagger tuned filters, and because every change in amplitude
response is accompanied by non-constant phase response.]

More important -- this "mush" problem is NOT a theoretical concept --
the problem has been experienced by hundreds of great CW operators in
contests and DX pileups. Those operators have, by careful experimenting
based on their understanding of how AGC works, and by listening to the
results of tweaking AGC parameters, developed settings that SOLVE THE
PROBLEM. Caps added for emphasis. In general, their solutions are quite
similar, differing only in degree.

This is NOT a problem for casual contesters or DXers, who either don't
call CQ in contests or don't have a big enough station or a callsign
that's rare enough in a given contest to generate a pileup. Most of the
time I'm operating I fall into the first category, but occasionally the
second. I've tried the recommendations of several top CW contesters, and
I do agree that the result is making multiple callers easier for my
brain to separate.

73, Jim K9YC

______________________________________________________________
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]

briancom

Re: [K3] AGC White Paper

In reply to this post by Cady, Fred-2

For curiosity we looked at the lower end of the curve with SpectrumLab
having a 0.27 Hz effective noise bandwidth at audio.

The K3 (not S here) shows AF linearity to more than 40 dB below S1!

So you guys with a low noise floor and ultra narrow digital modes could
potentially do wonders.

73 de Brian/K3KO

On 3/6/2017 15:11 PM, Cady, Fred wrote:

> Very nice Al Thanks.
>
> 73,
>
> Fred KE7X
>
>
>
> ________________________________
> From: Elecraft <[hidden email]> on behalf of Al Lorona <[hidden email]>
> Sent: Sunday, March 5, 2017 10:50 PM
> To: Elecraft Reflector
> Subject: [Elecraft] [K3] AGC White Paper
>
> I spent a little bit of time this weekend and put together yet another K3 AGC (YAKA) "white paper" to put some measurements and discussion of the results down in the hope that others may benefit. Maybe it'll help folks understand AGC better by demystifying some of the K3 idiosyncrasies. I hope it makes you think about a few things in a way that maybe never occurred to you. I put the document in my Dropbox at the following link:
>
>
> https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%20K3%20AGC%20System.pdf?dl=0
>
>
> I'll leave the file there for several days in case any of y'all are interested.
>
> 73,
>
> Al W6LX
> ______________________________________________________________
> 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]
>
> ______________________________________________________________
> 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]
>

Joe Subich, W4TV-4

Re: [K3] AGC White Paper

In reply to this post by Don Wilhelm

On 3/6/2017 2:29 PM, Don Wilhelm wrote:
> Actually it responds to the integral of all the signals in the
> passband, but if there is a single stronger one it will predominate.

In truth, depending on the AGC *attack time constant* the AGC system
responds to the *instantaneous peak voltage* of all signals in the
passband. Even for two signals where one signal is *only* 10 dB
stronger than the other, the *instantaneous peak voltage* can be as
much as 20 dB greater than the stronger signal alone.

With multiple signals, e.g. a large pile-up, and [relatively] slow
attack times, the instantaneous peak voltage can drive the ADC into
clipping before the AGC can react while with very fast attack and
decay times AGC can "pump" the receiver gain at a syllabic rate and
further "muddy" the recovered audio.

Steady state (single signal, single tone) measurements can show the
general AGC behavior but it is still an "art" to find the best
combination of attack time, decay time, threshold, slope (gain
reduction) above the threshold, "hold" times and *overall gain*
to keep peak signal voltages below the ADC clipping point and
simultaneously keeping the AGC response from "pumping" receiver
gain in such a way as to generate IMD (mush).

The key here is the *DYNAMIC* response of the AGC system.

73,

... Joe, W4TV

On 3/6/2017 2:29 PM, Don Wilhelm wrote:

> Wes is correct. Let me say it in a slightly different way.
> The AGC simply reduces the gain of the receiver - and it will respond
> mainly to the strongest signal in the passband. Actually it responds to
> the integral of all the signals in the passband, but if there is a
> single stronger one it will predominate.
>
> When the receiver gain is reduced, ALL the signals will be
> proportionally reduced in amplitude - including the weaker ones. This
> is often referred to as "AGC Pumping".
>
> 73,
> Don W3FPR
>
>
>
> On 3/6/2017 1:19 PM, Wes Stewart wrote:
>> In Al's very nicely done paper he has this paragraph:
>>
>> "The purpose of Automatic Gain Control (AGC) is to reduce the range
>> of the
>> signals seen by the sensitive stages in the receiver. The AGC stage is
>> designed to vary its gain depending on the input signal; stronger
>> signals
>> get less gain, and this has the effect of compressing the amplitude
>> range.
>> This is the desired response."
>>
>> I believe this paragraph and the accompanying graphic can be misleading
>> to the unwary. AGC does not compress the range of signals, it simply
>> lowers the gain through the receiver. The range (difference between)
>> signals might well be 130 dB at the input but it better be 130 dB
>> everywhere else in the receiver too. If lower level signals are driven
>> into the internal noise level because of gain reduction, so be it; that
>> should be the only reduction in range.
>>
>> Al continues:
>>
>> "But signals above this threshold will be acted on by the AGC. Even
>> though
>> in real life an S9 signal is 5 S-units stronger than an S4 signal,
>> because
>> of the AGC it will sound only 11.1 dB louder – less than 2 S-units
>> louder.
>> This is because, reading from the above graph, an S4 (-103 dBm) signal
>> produces -15.3 dBV of audio output and an S9 (-73 dBm) signal
>> produces -4.2
>> dBV, a difference of 11.1 dB. A five S-unit difference has been
>> reduced to a
>> less-than-two S-unit difference."
>>
>> Some are going to read this and mistakenly believe that while receiving
>> both an S9 signal and an S4 signal, AGC is going to reduce the ratio
>> between them from 5 S-units to two S-units.
>>
> ______________________________________________________________
> 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]

Brian Hunt

Re: [K3] AGC White Paper

In reply to this post by alorona

I noticed that your AGC curves were made with the RF gain full max. I rarely operate that way. I usually turn the RF gain down to where the noise peaks are barely kicking the S-meter. By doing so, is it fair to say I'm moving the curve to the right? I.e, a higher signal amplitude before the AGC kicks in for a given THR number?

Nice presentation Al. Thanks.

73,
Brian, K0DTJ

> On Mar 5, 2017, at 21:50, Al Lorona <[hidden email]> wrote:
>
> I spent a little bit of time this weekend and put together yet another K3 AGC (YAKA) "white paper" to put some measurements and discussion of the results down in the hope that others may benefit.

______________________________________________________________
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]

Matt Murphy

Re: [K3] AGC White Paper

Is there a way to simulate different agc settings by generating sine wave
CW signals and varying the relevant parameters?

I'd like to write some code to allow for straightforward observation of the
phenomena we are discussing and various assumptions about the input signals
to be tested.

It should be possible to create "mush" and also the desired response, I
would think. It would be useful to those wishing to understand the theory
in a more concrete way.

Matt NQ6N

On Mon, Mar 6, 2017 at 3:14 PM Brian Hunt <[hidden email]> wrote:

> I noticed that your AGC curves were made with the RF gain full max. I
> rarely operate that way. I usually turn the RF gain down to where the noise
> peaks are barely kicking the S-meter. By doing so, is it fair to say I'm
> moving the curve to the right? I.e, a higher signal amplitude before the
> AGC kicks in for a given THR number?
>
> Nice presentation Al. Thanks.
>
> 73,
> Brian, K0DTJ
>
> > On Mar 5, 2017, at 21:50, Al Lorona <[hidden email]> wrote:
> >
> > I spent a little bit of time this weekend and put together yet another
> K3 AGC (YAKA) "white paper" to put some measurements and discussion of the
> results down in the hope that others may benefit.
>
> ______________________________________________________________
> 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]
>

Don Wilhelm

Re: [K3] AGC White Paper

In reply to this post by Brian Hunt

Brian,

The effect of reducing the RF Gain is *about* the same as increasing the
AGC threshold. Not quite, but that is a simplistic way of looking at it.

73,
Don W3FPR

On 3/6/2017 6:13 PM, Brian Hunt wrote:
> I noticed that your AGC curves were made with the RF gain full max. I rarely operate that way. I usually turn the RF gain down to where the noise peaks are barely kicking the S-meter. By doing so, is it fair to say I'm moving the curve to the right? I.e, a higher signal amplitude before the AGC kicks in for a given THR number?
______________________________________________________________
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]

Ed Muns

Re: [K3] AGC White Paper

In reply to this post by Bill Frantz

Just the opposite. Most digital signal decoders work best with the least
AGC action possible, even AGC Off. AGC constrains decoder performance.

Ed W0YK
_____________________________________________________________________

-----Original Message-----
From: Elecraft [mailto:[hidden email]] On Behalf Of Bill
Frantz
Sent: 05 March, 2017 23:17
To: [hidden email]
Subject: Re: [Elecraft] [K3] AGC White Paper

Thanks Al. A very useful paper.

Now to use SLP=15 for digital, where I want the AtoD to not have
to deal with a big range, while using something with more
dynamic range for voice/CW.

73 Bill AE6JV

On 3/6/17 at 9:50 PM, [hidden email] (Al Lorona) wrote:

>I spent a little bit of time this weekend and put together yet
>another K3 AGC (YAKA) "white paper" to put some measurements
>and discussion of the results down in the hope that others may
>benefit. Maybe it'll help folks understand AGC better by
>demystifying some of the K3 idiosyncrasies. I hope it makes you
>think about a few things in a way that maybe never occurred to
>you. I put the document in my Dropbox at the following link:
>
>
><a href="https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%2">https://www.dropbox.com/s/drfujqupr4pcet3/Getting%20the%20Most%20Out%20of%2

0K3%20AGC%20System.pdf?dl=
>0
>
>
>I'll leave the file there for several days in case any of y'all are
interested.

-----------------------------------------------------------------------
Bill Frantz | Security is like Government | Periwinkle
(408)356-8506 | services. The market doesn't | 16345
Englewood Ave
www.pwpconsult.com | want to pay for them. | Los Gatos,
CA 95032

______________________________________________________________
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]

______________________________________________________________
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]

Bill Frantz

Re: [K3] AGC White Paper

What I get using a relatively flat AGC is automation of my
riding the RF gain control. I don't see how this kind of
automation affects decoder performance. It should help
performance as the decoder has a nearly constant level signal to
work with. Please explain.

Thanks - Bill AE6JV

On 3/7/17 at 11:37 AM, [hidden email] (Ed Muns) wrote:

>Just the opposite. Most digital signal decoders work best with the least
>AGC action possible, even AGC Off. AGC constrains decoder performance.
>
>Ed W0YK
>_____________________________________________________________________
>
>-----Original Message-----
>From: Elecraft [mailto:[hidden email]] On Behalf Of Bill
>Frantz
>...
>Now to use SLP=15 for digital, where I want the AtoD to not
>have to deal with a big range, while using something with more
>dynamic range for voice/CW.

-----------------------------------------------------------------------
Bill Frantz |Security, like correctness, is| Periwinkle
(408)356-8506 |not an add-on feature. - Attr-| 16345
Englewood Ave
www.pwpconsult.com |ibuted to Andrew Tanenbaum | Los Gatos,
CA 95032

______________________________________________________________
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]

Ed Muns

Re: [K3] AGC White Paper

Decoders operate on the audio stream presented to them from the receiver.
The decoding algorithms use the difference in audio level at each instant in
time to make bit decisions. AGC reduces those differences and thus
increases the bit error rate.

Of course, it is important to adjust the absolute audio levels so that the
dynamic range of the signal(s) is placed within the dynamic range of the
decoder. But, compressing the dynamic range, i.e., AGC, increases the error
rate of the decoder.

Ed W0YK
_________________________________________________________________

-----Original Message-----
From: Bill Frantz [mailto:[hidden email]]
Sent: 07 March, 2017 13:05
To: [hidden email]
Cc: [hidden email]
Subject: RE: [Elecraft] [K3] AGC White Paper

What I get using a relatively flat AGC is automation of my
riding the RF gain control. I don't see how this kind of
automation affects decoder performance. It should help
performance as the decoder has a nearly constant level signal to
work with. Please explain.

Thanks - Bill AE6JV

On 3/7/17 at 11:37 AM, [hidden email] (Ed Muns) wrote:

Alan Bloom

Re: [K3] AGC White Paper

If the AGC is working right, it is equivalent to manually riding the RF
gain control. It shouldn't affect the decoding.

If the AGC time constant is too fast then, yes, it can cause distortion
and degrade the decoding. But that should never happen in a
properly-designed AGC system.

AGC makes the decoder's job easier, not harder, because it reduces the
dynamic range the decoder has to deal with.

Alan N1AL

On 03/07/2017 01:19 PM, Ed Muns wrote:

> Decoders operate on the audio stream presented to them from the receiver.
> The decoding algorithms use the difference in audio level at each instant in
> time to make bit decisions. AGC reduces those differences and thus
> increases the bit error rate.
>
> Of course, it is important to adjust the absolute audio levels so that the
> dynamic range of the signal(s) is placed within the dynamic range of the
> decoder. But, compressing the dynamic range, i.e., AGC, increases the error
> rate of the decoder.
>
> Ed W0YK
> _________________________________________________________________
>
> -----Original Message-----
> From: Bill Frantz [mailto:[hidden email]]
> Sent: 07 March, 2017 13:05
> To: [hidden email]
> Cc: [hidden email]
> Subject: RE: [Elecraft] [K3] AGC White Paper
>
> What I get using a relatively flat AGC is automation of my
> riding the RF gain control. I don't see how this kind of
> automation affects decoder performance. It should help
> performance as the decoder has a nearly constant level signal to
> work with. Please explain.
>
> Thanks - Bill AE6JV
>
> On 3/7/17 at 11:37 AM, [hidden email] (Ed Muns) wrote:
>
>> Just the opposite. Most digital signal decoders work best with the least
>> AGC action possible, even AGC Off. AGC constrains decoder performance.
>>
>> Ed W0YK
>> _____________________________________________________________________
>>
>> -----Original Message-----
>> From: Elecraft [mailto:[hidden email]] On Behalf Of Bill
>> Frantz
>> ...
>> Now to use SLP=15 for digital, where I want the AtoD to not
>> have to deal with a big range, while using something with more
>> dynamic range for voice/CW.
>
> -----------------------------------------------------------------------
> Bill Frantz |Security, like correctness, is| Periwinkle
> (408)356-8506 |not an add-on feature. - Attr-| 16345
> Englewood Ave
> www.pwpconsult.com |ibuted to Andrew Tanenbaum | Los Gatos,
> CA 95032
>

Wes Stewart-2

Re: [K3] AGC White Paper

In reply to this post by Ed Muns

AGC adjusts the receiver gain, compared to the audio bits, rather slowly and
DOES NOT change the "dynamic range."

If you're listening to a signal and it's too loud for your ears and you turn the
audio gain down 10 dB, does that change the amplitude ratio of any signal you're
listening to? No, it lowers everything by 10 dB. AGC is no different, it just
operates at another location in the radio.

IMHO, this thinking is the result of looking at too many ACG slope, threshold,
and phase of the moon charts and assuming that bigger signals have their gain
reduced more that smaller ones at the same instant in time.

Wes N7WS

On 3/7/2017 2:19 PM, Ed Muns wrote:

Jim Brown-10

Re: [K3] AGC White Paper

Wes and Ed,

It's important to realize that in the K3 and later Elecraft rigs, all of
this is happening in the digital domain. One thing we learned about DSP
in pro audio is that dynamics processing burns a LOT processor cycles.
DSP SIMULATES the behavior of analog circuitry, but depending on the
code and the processor, the quality of the emulation can be varying
degrees of ideal.

ALSO -- those of us who have worked in broadcast audio know that
dynamics processing can be a VERY complex business, especially when
you're trying to do a lot of it, and even with a budget far larger than
in a ham rig. Even in analog, it's tricky to do a LOT of gain change and
still sound good. A handful of designers were good at it, and made a lot
of money for themselves, or their employer, or both.

And I'll repeat my previous observation that a signal path with gain
that varies with time or with the signal amplitude is NOT a linear
device, and non-linearity => distortion.

73, Jim K9YC

On Tue,3/7/2017 1:53 PM, Wes Stewart wrote:

> AGC adjusts the receiver gain, compared to the audio bits, rather
> slowly and DOES NOT change the "dynamic range."
>
> If you're listening to a signal and it's too loud for your ears and
> you turn the audio gain down 10 dB, does that change the amplitude
> ratio of any signal you're listening to? No, it lowers everything by
> 10 dB. AGC is no different, it just operates at another location in
> the radio.
>
> IMHO, this thinking is the result of looking at too many ACG slope,
> threshold, and phase of the moon charts and assuming that bigger
> signals have their gain reduced more that smaller ones at the same
> instant in time.
>
> Wes N7WS
>
> On 3/7/2017 2:19 PM, Ed Muns wrote:
>> Decoders operate on the audio stream presented to them from the
>> receiver.
>> The decoding algorithms use the difference in audio level at each
>> instant in
>> time to make bit decisions. AGC reduces those differences and thus
>> increases the bit error rate.
>>
>> Of course, it is important to adjust the absolute audio levels so
>> that the
>> dynamic range of the signal(s) is placed within the dynamic range of the
>> decoder. But, compressing the dynamic range, i.e., AGC, increases
>> the error
>> rate of the decoder.