Question on two stage AGC.

I understand the general idea about AGC is to measure the input signal
power. If it is too big, lower it down. If it is too small, scale it
up.

I am now given a two stage AGC and try to understand it. Basically,
it's two consecutive sequences say AGC1 and AGC2 and it is claimed
that AGC1 and AGC2 together help receiver accomplish the overall AGC
task.

I think this is some kind of two stage or two level AGC. But why do we
need two level to finish the AGC, which in my naive mind is a pretty
simple task? Please share some thoughts on why and how to make the two
level AGC work? THanks a lot.

Phil

phil wrote:
> I understand the general idea about AGC is to measure the input signal
> power. If it is too big, lower it down. If it is too small, scale it
> up.
> 
> I am now given a two stage AGC and try to understand it. Basically,
> it's two consecutive sequences say AGC1 and AGC2 and it is claimed
> that AGC1 and AGC2 together help receiver accomplish the overall AGC
> task.
> 
> I think this is some kind of two stage or two level AGC. But why do we
> need two level to finish the AGC, which in my naive mind is a pretty
> simple task? Please share some thoughts on why and how to make the two
> level AGC work? THanks a lot.
> 

Generally because if you need too much gain adjustment you would lose 
too much dynamic range if you did it at only one stage.

-- 
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

On Apr 10, 7:39 pm, phil <phi...@gmail.com> wrote:
> I understand the general idea about AGC is to measure the input signal
> power. If it is too big, lower it down. If it is too small, scale it
> up.
>
> I am now given a two stage AGC and try to understand it. Basically,
> it's two consecutive sequences say AGC1 and AGC2 and it is claimed
> that AGC1 and AGC2 together help receiver accomplish the overall AGC
> task.
>
> I think this is some kind of two stage or two level AGC. But why do we
> need two level to finish the AGC, which in my naive mind is a pretty
> simple task? Please share some thoughts on why and how to make the two
> level AGC work? THanks a lot.
>
> Phil

In a radio receiver, one frequently finds attenuation control points
before and after the IF strip. When turning the gain down, the
question of which one to adjust first depends on whether intercept
point or noise figure is more important.

John

phil wrote:
(snip)

> I am now given a two stage AGC and try to understand it. Basically,
> it's two consecutive sequences say AGC1 and AGC2 and it is claimed
> that AGC1 and AGC2 together help receiver accomplish the overall AGC
> task.

> I think this is some kind of two stage or two level AGC. But why do we
> need two level to finish the AGC, which in my naive mind is a pretty
> simple task? Please share some thoughts on why and how to make the two
> level AGC work? THanks a lot.

Having never actually designed one, I am sometimes surprised at how
well car radios work far away from the station or right next to
the transmitter.

Last summer out camping away from any large city one night we tried
to see how far away we could find an AM station.  It had to be clear
enough to hear the city it was from.

-- glen

glen herrmannsfeldt wrote:

   ...

> Last summer out camping away from any large city one night we tried
> to see how far away we could find an AM station.  It had to be clear
> enough to hear the city it was from.

And ...?

Jerry

P.S. AM or FM?
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Jerry Avins wrote:

> glen herrmannsfeldt wrote:

>> Last summer out camping away from any large city one night we tried
>> to see how far away we could find an AM station.  It had to be clear
>> enough to hear the city it was from.

> And ...?

The farthest I remember was about 1000 miles. (San Francisco from
somewhere north of Seattle.)   That one was very clear, so some
could have been farther away but didn't indicate the city.
Sometimes they didn't say the city in the time we listened to each one.

> P.S. AM or FM?

AM.  As some here know, at night AM stations will bounce off the
ionosphere and can be heard at long distances.  FM stations
(with their higher frequency) don't do that limiting the possible
distance.

-- glen

Jerry Avins wrote:
(previous snip on AGC)

> And ...?

Another place where AGC is important and you might not expect
it is the input to a cable modem.  It seems that ours has
a +16dBm input signal, which is very high.  (Close to enough
for a small light bulb, though I didn't try it.)
It will work without it, but I now have an 8dB attenuator.
I have run it with a 28dB attenuator and it works fine.
Maybe smaller dynamic range than expected for a car radio,
but higher than you might expect for a constant source
like cable.

-- glen

On Fri, 11 Apr 2008 13:34:58 -0800, glen herrmannsfeldt
<gah@ugcs.caltech.edu> wrote:

>Jerry Avins wrote:
>(previous snip on AGC)
>
>> And ...?
>
>Another place where AGC is important and you might not expect
>it is the input to a cable modem.  It seems that ours has
>a +16dBm input signal, which is very high.  (Close to enough
>for a small light bulb, though I didn't try it.)
>It will work without it, but I now have an 8dB attenuator.
>I have run it with a 28dB attenuator and it works fine.
>Maybe smaller dynamic range than expected for a car radio,
>but higher than you might expect for a constant source
>like cable.
>
>-- glen

The cable modem architect can't assume how many splitters or hybrids
there are between the source and the modem, or even how much cable,
for that matter.    So the dynamic range requirement is still pretty
significant.
Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.ericjacobsen.org

Eric Jacobsen wrote:
(snip)

> The cable modem architect can't assume how many splitters or hybrids
> there are between the source and the modem, or even how much cable,
> for that matter.    So the dynamic range requirement is still pretty
> significant.

In our case, the neighborhood amplifier is in our front yard and
no splitter for cable TV.   Some cable modems won't work well
with +16dBm input.

-- glen

On Mon, 14 Apr 2008 11:02:49 -0800, glen herrmannsfeldt
<gah@ugcs.caltech.edu> wrote:

>Eric Jacobsen wrote:
>(snip)
>
>> The cable modem architect can't assume how many splitters or hybrids
>> there are between the source and the modem, or even how much cable,
>> for that matter.    So the dynamic range requirement is still pretty
>> significant.
>
>In our case, the neighborhood amplifier is in our front yard and
>no splitter for cable TV.   Some cable modems won't work well
>with +16dBm input.
>
>-- glen


:)

Yeah, I can see that being a problem.   ;)

Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.ericjacobsen.org