Elecraft

ALC implementation

Classic

List

Threaded

32 messages Options

Bob McGraw - K4TAX

Re: ALC implementation

Don't confuse compression and clipping. Compression reduces dynamic range while clipping sets a hard ceiling and does generate distortion. Compression used correctly does not generate distortion.

73
Bob, K4TAX

Sent from my iPhone

> On Jul 12, 2019, at 4:00 PM, Wes <[hidden email]> wrote:
>
> What distortion?
>
>> On 7/12/2019 1:15 PM, Jim Rhodes wrote:
>> Nothing when it is done correctly on a mode that can stand the distortion. I personally won't use it on digital modes as my rigs tend to turn it off by default.
>>
>> Jim Rhodes
>> K0XU
>>
>> On Fri, Jul 12, 2019, 15:10 Wes <[hidden email] <mailto:[hidden email]>> wrote:
>>
>> And what exactly is the problem with compression? It's used all over the
>> place.
>>
>> Wes N7WS
>>
>> On 6/28/2019 11:38 AM, Brian Denley wrote:
>> > I have always thought that ALC is not appropriate for power control.
>> It’s a protection system. Over use of ALC acts like compression. No?
>> >
>> > Brian
>> > KB1VBF
>> > Sent from my iPad
>>
>> ______________________________________________________________
>> Elecraft mailing list
>> Home: http://mailman.qth.net/mailman/listinfo/elecraft
>> Help: http://mailman.qth.net/mmfaq.htm
>> Post: mailto:[hidden email] <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] <mailto:[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]

k6dgw

Re: ALC implementation

Well ... I think it depends on your definition of distortion. One can
argue that, if the RF envelope does not exactly equal the original AF
envelope, the signal has been distorted and that is certainly the case
when the dynamic range has been reduced with compression. It doesn't
increase the occupied bandwidth however [at least not by much].
Clipping mows down the peaks, leaving sharp "edges" or signal
transitions which do increase the occupied bandwidth.

73,
Fred ["Skip"] K6DGW
Sparks NV DM09dn
Washoe County

On 7/12/2019 2:41 PM, Bob McGraw K4TAX wrote:

Conrad PA5Y

Re: ALC implementation

It should be mentioned that FT8 and JT65 are constant envelope and therefore there should not be any AF envelope to follow. Actually when I recently measured the AF output from WSJT-X there was a tiny amount of AM, around 1.4% in fact which so small as to be of no consequence. ALC can introduce AM quite easily if not properly implemented or adjusted. The recommended method of adjusting the K3S AF drive for digital modes works very well but each radio is different. The recent Icoms (IC-7300, IC7610 and IC9700) 50% indication on the ALC meter does not cause any degradation or spreading even for MSK144 which has 16.5% AM. Presumably there is a small delay introduced to prevent overshoots, a so called 'look ahead' algorithm.

Regards

Conrad PA5Y

-----Original Message-----
From: [hidden email] <[hidden email]> On Behalf Of Fred Jensen
Sent: 13 July 2019 00:33
To: [hidden email]
Subject: Re: [Elecraft] ALC implementation

Well ... I think it depends on your definition of distortion. One can argue that, if the RF envelope does not exactly equal the original AF envelope, the signal has been distorted and that is certainly the case when the dynamic range has been reduced with compression. It doesn't increase the occupied bandwidth however [at least not by much]. Clipping mows down the peaks, leaving sharp "edges" or signal transitions which do increase the occupied bandwidth.

73,
Fred ["Skip"] K6DGW
Sparks NV DM09dn
Washoe County

On 7/12/2019 2:41 PM, Bob McGraw K4TAX 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]

Wes Stewart-2

Re: ALC implementation

In reply to this post by Bob McGraw - K4TAX

Actually, gain compression reduces "gain" not dynamic range. All signals are
reduced in amplitude by the same number of dB. This is the only way it doesn't
introduce distortion.

This reminds me of the endless discussions about RX AGC where everyone has their
favorite combination of rf-gain, threshold, slope, etc, that somehow magically
reduces strong signals while allowing weaker ones to "pop out of the noise." If
you have one of these receivers, you best only use it for FM reception.

There is nothing inherently wrong with using compression, i.e. ALC on digital
modes if the time constants are appropriate. I have just run some tests on my
K3S on FT8 while looking at the spectrum on a spectrum analyzer and listening
with a second K3 for the dreaded in-band harmonics while "overdriving" the audio
into the ALC region. With the present setup, I got the "recommended" 4-5 bars
with the Line In at about 7 to 8. Cranking it up to 30 and maxing out the ALC
at 8 bars made no difference in detectable in-band harmonics, even using a 300
Hz offset. (Split was off)

It's funny that this guy, https://www.youtube.com/watch?v=smWp0TOEtlQ decides
that speech processing has no undesired effects yet is adamant that ALC should
not be used.

Wes N7WS

On 7/12/2019 2:41 PM, Bob McGraw K4TAX wrote:

> Don't confuse compression and clipping. Compression reduces dynamic range while clipping sets a hard ceiling and does generate distortion. Compression used correctly does not generate distortion.
>
> 73
> Bob, K4TAX
>
>
> Sent from my iPhone
>
>> On Jul 12, 2019, at 4:00 PM, Wes <[hidden email]> wrote:
>>
>> What distortion?
>>
>>> On 7/12/2019 1:15 PM, Jim Rhodes wrote:
>>> Nothing when it is done correctly on a mode that can stand the distortion. I personally won't use it on digital modes as my rigs tend to turn it off by default.
>>>
>>> Jim Rhodes
>>> K0XU
>>>
>>> On Fri, Jul 12, 2019, 15:10 Wes <[hidden email] <mailto:[hidden email]>> wrote:
>>>
>>> And what exactly is the problem with compression? It's used all over the
>>> place.
>>>
>>> Wes N7WS
>>>
>>> On 6/28/2019 11:38 AM, Brian Denley wrote:
>>> > I have always thought that ALC is not appropriate for power control.
>>> It’s a protection system. Over use of ALC acts like compression. No?
>>> >
>>> > Brian
>>> > KB1VBF
>>> > Sent from my iPad
>>>
>>> ______________________________________________________________
>>> Elecraft mailing list
>>> Home: http://mailman.qth.net/mailman/listinfo/elecraft
>>> Help: http://mailman.qth.net/mmfaq.htm
>>> Post: mailto:[hidden email] <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] <mailto:[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]
>
>

Jim Brown-10

Re: ALC implementation

In reply to this post by Conrad PA5Y

A common cause of incidental AM is ripple in the crystal filter(s) in
use. One reason I went to the 2.8 kHz filter was that it had less
ripple, and the difference was obvious on RTTY.

73, Jim K9YC

On 7/12/2019 3:45 PM, Conrad PA5Y wrote:
> It should be mentioned that FT8 and JT65 are constant envelope and therefore there should not be any AF envelope to follow. Actually when I recently measured the AF output from WSJT-X there was a tiny amount of AM, around 1.4% in fact which so small as to be of no consequence.

______________________________________________________________
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]

bdenley

Re: ALC implementation

In reply to this post by Wes Stewart-2

Nothing when that’s what you want. But using ALC to control xcvr power is uncontrolled compression that may or may not be desired.

Brian
KB1VBF
Sent from my iPad

> On Jul 12, 2019, at 4:09 PM, Wes <[hidden email]> wrote:
>
> And what exactly is the problem with compression? It's used all over the place.
>
> Wes N7WS
>
>> On 6/28/2019 11:38 AM, Brian Denley wrote:
>> I have always thought that ALC is not appropriate for power control. It’s a protection system. Over use of ALC acts like compression. No?
>>
>> Brian
>> KB1VBF
>> Sent from my iPad
>
> ______________________________________________________________
> 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]

Wes Stewart-2

Re: ALC implementation

You do know that the K3(S) uses a slow ALC system for power control don't you?

Wes N7WS

On 7/12/2019 8:53 PM, Brian Denley wrote:

> Nothing when that’s what you want. But using ALC to control xcvr power is uncontrolled compression that may or may not be desired.
>
> Brian
> KB1VBF
> Sent from my iPad
>
>> On Jul 12, 2019, at 4:09 PM, Wes <[hidden email]> wrote:
>>
>> And what exactly is the problem with compression? It's used all over the place.
>>
>> Wes N7WS
>>
>>> On 6/28/2019 11:38 AM, Brian Denley wrote:
>>> I have always thought that ALC is not appropriate for power control. It’s a protection system. Over use of ALC acts like compression. No?
>>>
>>> Brian
>>> KB1VBF
>>> Sent from my iPad
>> ______________________________________________________________
>> 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]
>

bdenley

Re: ALC implementation

Wes:
Yes but I assume you have have used or use other xcvrs as well. As a general comment, that doesn’t change the idea that using ALC to control a xcvr’s power to one’s linear isn’t the purpose of ALC and would seem to not be a very healthy practice. If one were to choose that path, then one should be mindful that the result acts much like compression. In some cases, lots of compression!

Brian Denley
KB1VBF
Sent from my iPad

> On Jul 13, 2019, at 12:32 AM, Wes <[hidden email]> wrote:
>
> You do know that the K3(S) uses a slow ALC system for power control don't you?
>
> Wes N7WS
>
>> On 7/12/2019 8:53 PM, Brian Denley wrote:
>> Nothing when that’s what you want. But using ALC to control xcvr power is uncontrolled compression that may or may not be desired.
>>
>> Brian
>> KB1VBF
>> Sent from my iPad
>>
>>> On Jul 12, 2019, at 4:09 PM, Wes <[hidden email]> wrote:
>>>
>>> And what exactly is the problem with compression? It's used all over the place.
>>>
>>> Wes N7WS
>>>
>>>> On 6/28/2019 11:38 AM, Brian Denley wrote:
>>>> I have always thought that ALC is not appropriate for power control. It’s a protection system. Over use of ALC acts like compression. No?
>>>>
>>>> Brian
>>>> KB1VBF
>>>> Sent from my iPad
>>> ______________________________________________________________
>>> 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]
>>
>

Richard Corfield

Re: ALC implementation

In reply to this post by Wes Stewart-2

The compression modulates the signal. That will have a fourier transform of
its own. If we think of it as a simple amplitude modulation then, as for AM
signals, the spectrum of the original signal will be combined (convolved)
with upper and lower sidebands representing the spectrum of the compressor.
Our ears don't notice (unless we're sound guys listening to a pop song
thinking "They've squashed those dynamics to death" ;-) ) but maybe a
digital mode will.

The transmitted signal is constrained to the IF passband so you stay within
band. The compression frequencies and hopefully amplitude are both small so
the artefacts it introduces into the spectrum should also be both narrow
and small.

- Richard

On Fri, 12 Jul 2019 at 22:01, Wes <[hidden email]> wrote:

> What distortion?
>
> On 7/12/2019 1:15 PM, Jim Rhodes wrote:
> > Nothing when it is done correctly on a mode that can stand the
> distortion. I
> > personally won't use it on digital modes as my rigs tend to turn it off
> by
> > default.
> >
> > Jim Rhodes
> > K0XU
> >
> > On Fri, Jul 12, 2019, 15:10 Wes <[hidden email]
> > <mailto:[hidden email]>> wrote:
> >
> > And what exactly is the problem with compression? It's used all
> over the
> > place.
> >
> > Wes N7WS
> >
> > On 6/28/2019 11:38 AM, Brian Denley wrote:
> > > I have always thought that ALC is not appropriate for power
> control.
> > It’s a protection system. Over use of ALC acts like compression.
> No?
> > >
> > > Brian
> > > KB1VBF
> > > Sent from my iPad
> >
> > ______________________________________________________________
> > Elecraft mailing list
> > Home: http://mailman.qth.net/mailman/listinfo/elecraft
> > Help: http://mailman.qth.net/mmfaq.htm
> > Post: mailto:[hidden email] <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] <mailto:[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]

gm3sek

Re: ALC implementation

gm3sek

Re: ALC implementation

In reply to this post by Richard Corfield

The original concept of ALC was devised by Collins engineers as part of
their development program for speech processing. Speech clipping
flattens the peak amplitude, but also introduces some splatter which
then has to be filtered out - but the filtering re-introduces a small
amount of level variation (called 're-peaking'). Collins then added a
very small amount of ALC, only 2-3dB, to keep the peak envelope power
more constant. This mild ALC introduced very little additional splatter
of its own, so the result was a large increase in intelligibility with
very little increase in bandwidth.

Where ALC got a bad name was when other manufacturers started mis-using
it for other purposes - in particular for manual power control from 100W
down to 1W or less. ALC was also being used to compensate for variations
in TX gain between different bands. All of this required much more gain
in the control loop (20-30dB) which made teh ALC much more aggressive
and created major problems with transient behavior - in other words,
quite serious splatter due to the ALC system itself.

Then someone decided it would be a good idea to apply ALC from the power
amplifier, thus wrapping an external control loop around the existing
ALC loop within the transceiver, so the transient behavior became
totally unpredictable. Add to this the tendency of many hams to
perpetually overdrive their transmitters while *also* applying speech
processing, and the result was that both ALC and speech processing got a
very bad name.

Enter the K3. With a little persuasion from early adopters, Elecraft
returned to the original Collins concept of applying only a small amount
of ALC to minimise re-peaking *without* significantly re-introducing
splatter. The SDR concept allowed other means to be used to implement
the non-dynamic power control functions such as manual power setting and
TX gain compensation. When correctly configured according to the User
Manual, the K3 family of transceivers are capable of excellent speech
processing without worrying about also creating splatter. (Also note
that speech processing is automatically disabled for data modes.)

As far as I know, there is still no other manufacturer who is doing this.

References:

[1] W E Sabin and E O Schoenike (Collins), 'Single-Sideband Systems and
Circuits'. McGraw-Hill Book Company 1987. ISBN 0-07-054407-7

[2] Leif Åsbrink, SM5BSZ (http://sm5bsz.com), ‘Real Life Dynamic Range
of Modern Amateur Transceivers’ and ‘Speech Processing for SSB
Transmitters’.

73 from Ian GM3SEK

On 13/07/2019 08:35, Richard Corfield wrote:

> The compression modulates the signal. That will have a fourier transform of
> its own. If we think of it as a simple amplitude modulation then, as for AM
> signals, the spectrum of the original signal will be combined (convolved)
> with upper and lower sidebands representing the spectrum of the compressor.
> Our ears don't notice (unless we're sound guys listening to a pop song
> thinking "They've squashed those dynamics to death" ;-) ) but maybe a
> digital mode will.
>
> The transmitted signal is constrained to the IF passband so you stay within
> band. The compression frequencies and hopefully amplitude are both small so
> the artefacts it introduces into the spectrum should also be both narrow
> and small.
>
> - Richard
>
>

K9MA

Re: ALC implementation

In reply to this post by Wes Stewart-2

I like to think of ALC as receiver AGC in reverse. Done properly, it
results in minimal distortion while keeping the signal level nearly
constant. As far as I can tell, commercial transceivers have all been
doing this pretty well for a long time.

What I've never been able to figure out is why it's so hard to close the
ALC loop around an external amplifier. If it works with the 100 W
amplifier built into the transceiver, why won't it work with a bigger
amplifier? (And it seems about universally agreed that it won't.)

73,
Scott K9MA

On 7/12/2019 23:32, Wes wrote:

> You do know that the K3(S) uses a slow ALC system for power control
> don't you?
>
> Wes N7WS
>
> On 7/12/2019 8:53 PM, Brian Denley wrote:
>> Nothing when that’s what you want. But using ALC to control xcvr
>> power is uncontrolled compression that may or may not be desired.
>>
>> Brian
>> KB1VBF
>> Sent from my iPad
>>
>>> On Jul 12, 2019, at 4:09 PM, Wes <[hidden email]> wrote:
>>>
>>> And what exactly is the problem with compression? It's used all
>>> over the place.
>>>
>>> Wes N7WS
>>>
>>>> On 6/28/2019 11:38 AM, Brian Denley wrote:
>>>> I have always thought that ALC is not appropriate for power
>>>> control. It’s a protection system. Over use of ALC acts like
>>>> compression. No?
>>>>
>>>> Brian
>>>> KB1VBF
>>>> Sent from my iPad
>>> ______________________________________________________________

--
Scott K9MA

[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]