# Official Definition of Noise Gain

M

#### Michael A. Terrell

Jan 1, 1970
0
Eeyore said:
You are clearly mentally defective.

I've been saying for ages that the Alzheimers was destroying your
credibility but you insist on posting nonsense to prove the point.

--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

F

#### Fred Bartoli

Jan 1, 1970
0
Eeyore a écrit :
Since when was the gain for noise any different to the gain for signal ?
Or have you designed a special circuit that amplifies noise less than
signal ? They're called FILTERS btw.

Try the simple inverting opamp.

As usual, you're showing how abyssal your lack of understanding is.

Noise... Noise?
Hem, that reminds me some past thread where your skills shined too.

F

#### Fred Bloggs

Jan 1, 1970
0
Paul said:

That confused idiot doesn't know what the hell he's talking about...When
it comes to linear electronics, and especially operational amplifier
circuits, you need to think about catalogs and directories of
standardized circuits, which can be used in various non-standard ways.
To make things more useful to the cookbook engineer, it makes sense to
refer the total circuit noise to a standardized input, the non-inverting
input in the case of operational amplifiers, and provide a gain function
from that input to the output, called the noise gain. Then the cookbook
engineer need only divide that noise gain function by his signal I/O
gain function to refer the noise to his signal input and thereby derive
a S/N for his particular application. By this reasoning, the noise gain
is by definition the ratio of total output noise to input noise referred
to a standardized input for the circuit topology under consideration. If
you don't like the idea of an assumed standardized input then certainly
you would have to take the standardized input as the one called out in
the manufacturer's specification sheet for the input voltage and current
noise density functions, and this is almost always the non-inverting
input. Call it the datasheet specified input if you don't like
standardized input terminology.

S

#### Stephan Goldstein

Jan 1, 1970
0
What is the "Official Definition of Noise Gain"?

I'm looking for a good clean analytical method I can add into PSpice.

Pease's article is, as usual, clear as mud ;-)

...Jim Thompson

Jim, do you have a copy of the "bible", a.k.a. Motchenbacher &
Fitchen? That's where I'd look first. Mine's on loan, sorry.

steve

A

#### Allan Herriman

Jan 1, 1970
0
That's what I'm concluding, also. Is that in general?

That's how I've been using it. I've not noticed any other
definitions.

Regards,
Allan

J

#### Jim Thompson

Jan 1, 1970
0
To- said:
Jim Thompson wrote:
What is the "Official Definition of Noise Gain"?

Definition: A gain that happens during the months before major
elections. Typically affects technical newsgroups the most ;-)

[...]

Sno-o-o-o-ort!

Just got back from "Bucket List" and dinner.

Great movie if you're old enough to have experienced (and appreciated)
some of the representations ;-)

...Jim Thompson
"Noise gain (1/B): The voltage response given by the
inverse of feedback-loop attenuation" --Walter Jung p.31

and, "The term that includes error factors due to nonideal
gain is the _true noise gain_ or _closed-loop gain_... "
on p.30

I'll pull out Walt's book here in a few minutes when I finish my
coffee, but I think that is true only from the non-inverting input.

Say you have a TIA, the noise-gain is dramatically different from
closed loop gain.

...Jim Thompson

J

#### Jim Thompson

Jan 1, 1970
0
That confused idiot doesn't know what the hell he's talking about...When
it comes to linear electronics, and especially operational amplifier
circuits, you need to think about catalogs and directories of
standardized circuits, which can be used in various non-standard ways.
To make things more useful to the cookbook engineer, it makes sense to
refer the total circuit noise to a standardized input, the non-inverting
input in the case of operational amplifiers, and provide a gain function
from that input to the output, called the noise gain. Then the cookbook
engineer need only divide that noise gain function by his signal I/O
gain function to refer the noise to his signal input and thereby derive
a S/N for his particular application. By this reasoning, the noise gain
is by definition the ratio of total output noise to input noise referred
to a standardized input for the circuit topology under consideration. If
you don't like the idea of an assumed standardized input then certainly
you would have to take the standardized input as the one called out in
the manufacturer's specification sheet for the input voltage and current
noise density functions, and this is almost always the non-inverting
input. Call it the datasheet specified input if you don't like
standardized input terminology.

Except that noise gain has more to do with loop stability than noise
;-)

...Jim Thompson

J

#### Jim Thompson

Jan 1, 1970
0
Jim, do you have a copy of the "bible", a.k.a. Motchenbacher &
Fitchen? That's where I'd look first. Mine's on loan, sorry.

steve

Somewhere in my over-loaded office. Thanks for the pointer, Steve!
That is indeed the "bible".

...Jim Thompson

J

#### Jim Thompson

Jan 1, 1970
0
Somewhere in my over-loaded office. Thanks for the pointer, Steve!
That is indeed the "bible".

But nary a mention of "noise gain" :-(

...Jim Thompson

D

#### Don Bowey

Jan 1, 1970
0
Except that noise gain has more to do with loop stability than noise
;-)

...Jim Thompson

As probable cause and effect, how does one separate them?

E

#### Eeyore

Jan 1, 1970
0
Fred said:
Eeyore a écrit :

Try the simple inverting opamp.

Where's the problem in analysing that ? Audio desingners do this kind of thing
every day.

As usual, you're showing how abyssal your lack of understanding is.

Projection.

Graham

F

#### Fred Bloggs

Jan 1, 1970
0
Jim said:
Except that noise gain has more to do with loop stability than noise
;-)

...Jim Thompson

All of the transfer functions for all the nodes that are possible inputs
have the same denominator, for a linear network, whether its active or
passive or whatever, right. This may be where the noise gain is again
useful for characterizing the circuit frequency response in a general way.

F

#### Fred Bloggs

Jan 1, 1970
0
Jim said:
But nary a mention of "noise gain" :-(

...Jim Thompson

It is called out in Dostal on pages 120 and 237, where he matter of
factly states it is the inverse of the feedback factor, 1/beta, and
often approaches the ideal closed loop gain.

J

#### Jim Thompson

Jan 1, 1970
0
It is called out in Dostal on pages 120 and 237, where he matter of
factly states it is the inverse of the feedback factor, 1/beta, and
often approaches the ideal closed loop gain.

From the NON-INVERTING Input.

...Jim Thompson

J

#### Jim Thompson

Jan 1, 1970
0
To- said:
Jim Thompson wrote:
What is the "Official Definition of Noise Gain"?

Definition: A gain that happens during the months before major
elections. Typically affects technical newsgroups the most ;-)

[...]

Sno-o-o-o-ort!

Just got back from "Bucket List" and dinner.

Great movie if you're old enough to have experienced (and appreciated)
some of the representations ;-)

...Jim Thompson
"Noise gain (1/B): The voltage response given by the
inverse of feedback-loop attenuation" --Walter Jung p.31

and, "The term that includes error factors due to nonideal
gain is the _true noise gain_ or _closed-loop gain_... "
on p.30

On pp 23-24 of my 3rd edition.

It boils down to closed-loop gain from the non-inverting input.

I currently have modified my LoopGain symbol/macro to get loop gain,
loop phase and closed-loop gain with data from a 3-pass AC analysis
macro 'd into PSpice Probe.

I thought I had enough data to also get noise gain, but it looks like
I need another pass... blink of an eye but it gets me everything
displayed at once... the customers like that ;-)

...Jim Thompson

P

#### Paul Hovnanian P.E.

Jan 1, 1970
0
Eeyore said:
I've been intruiged by this thread since I'm puzzled how anyone could think the
gain for noise would be any different to the gain for a signal (measuring
bandwidth issues aside) !

The method you mention Paul is of course exactly how audio designers do it. I
can't imagine why anyone would do it any other way since they'll simply be WRONG
!

There's some validity to considering a gain that is different for a
voltage (noise or otherwise) that is injected at some point other than
the circuit input. Resistor thermal noise 'sees' different gains
depending on where they are WRT the feedback circuits. And op-amp
spectral noise is spec'd WRT its input.

But once the noise sources are modeled and referred to the circuit input
(or output), a separate figure for 'noise gain' is something I'd throw
away rather than quoting in a system description.

P

#### Phil Hobbs

Jan 1, 1970
0
Paul said:
There's some validity to considering a gain that is different for a
voltage (noise or otherwise) that is injected at some point other than
the circuit input. Resistor thermal noise 'sees' different gains
depending on where they are WRT the feedback circuits. And op-amp
spectral noise is spec'd WRT its input.

But once the noise sources are modeled and referred to the circuit input
(or output), a separate figure for 'noise gain' is something I'd throw
away rather than quoting in a system description.

Noise gain is important in several instances I continually run into, the
leading one being transimpedance amps. I think most of us have an
inbuilt tendency to regard noise as flat with frequency, since all the
fundamental circuit noise sources are flat. Photodiode capacitance
hanging on the summing junction puts a big high-frequency peak on the
noise gain of a vanilla op-amp TIA. That multiplication only affects
the input voltage noise of the op amp--all the other noise current
sources appear in parallel with the signal current, so they see the same
gain.

So the noise gain is something one needs to keep a close eye on sometimes.

Cheers,

Phil Hobbs

P

#### Paul Hovnanian P.E.

Jan 1, 1970
0
Jim said:
Except that noise gain has more to do with loop stability than noise
;-)

But the problem with one figure for noise gain would seem to be that
noise sources at different points within the system 'see' different
gains. Lumping them all to one point (an op-amp input, for example) and
then applying a single gain figure to them, you would lose visibility of
the stability margin due to noise at each point.

Lumping stuff at one point might be useful as a rule of thumb to see how
close you are pushing things, but eventually you'll have to pull the
pieces apart again to see what contributes the most. If you are going
to model the system anyway, you've already got that.

J

#### Jim Thompson

Jan 1, 1970
0
Paul said:
Eeyore said:
[snip]

But once the noise sources are modeled and referred to the circuit input
(or output), a separate figure for 'noise gain' is something I'd throw
away rather than quoting in a system description.

Noise gain is important in several instances I continually run into, the
leading one being transimpedance amps. I think most of us have an
inbuilt tendency to regard noise as flat with frequency, since all the
fundamental circuit noise sources are flat. Photodiode capacitance
hanging on the summing junction puts a big high-frequency peak on the
noise gain of a vanilla op-amp TIA. That multiplication only affects
the input voltage noise of the op amp--all the other noise current
sources appear in parallel with the signal current, so they see the same
gain.

So the noise gain is something one needs to keep a close eye on sometimes.

Cheers,

Phil Hobbs

Yep. (Our buddies in Atlanta are back on the air ;-)

...Jim Thompson

J

#### Jim Thompson

Jan 1, 1970
0
But the problem with one figure for noise gain would seem to be that
noise sources at different points within the system 'see' different
gains. Lumping them all to one point (an op-amp input, for example) and
then applying a single gain figure to them, you would lose visibility of
the stability margin due to noise at each point.

Lumping stuff at one point might be useful as a rule of thumb to see how
close you are pushing things, but eventually you'll have to pull the
pieces apart again to see what contributes the most. If you are going
to model the system anyway, you've already got that.

With PSpice it's trivial to get a device-by-device listing of noise
contributions.

...Jim Thompson

Replies
1
Views
769
Replies
4
Views
1K
P
Replies
2
Views
3K
P
P
Replies
1
Views
1K
Sylvia Else
S
A
Replies
125
Views
6K
Joerg
J