Searching \ for 'Low Noise, High Gain Amplifier' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=low+noise+high+gain
Search entire site for: 'Low Noise, High Gain Amplifier'.

Truncated match.
PICList Thread
'Low Noise, High Gain Amplifier'
2008\07\26@013152 by Rich

picon face
The more stages you add the greater will be the noise problem.  The first
thing you need to do is create a "budget;" that is, an error budget that
includes the highest level of tolerable noise together with the other
factors that contribute to noise, such as operating temperature, operating
voltages and so on.  If you can accomplish what you want with a single FET
or bipolar, you may be better off than with an op amp.  That being said,
there are some good low noise op amps available.  The gain you require may
lead to your choice.  Are there any other parameters, like slewing rate and
settling time, for example?  It is always a good idea to write a
specification before you begin a design.  What percentage distortion do you
require as a minimum? There are many considerations.  Perhaps you can think
about it and put your specs up on the list.  There are some excellent design
engineers on this list.  Some of them post regularly.  I am sure that you
will get excellent guidance.  : - )



{Original Message removed}

2008\07\26@023255 by Sean Breheny

face picon face
Hi Fred,

First off, what is this for? That may help to answer your question.

Your bandwidth is fairly small - I think that your rough noise "goal"
could be met with most decent op-amps. The best low-noise op-amp I
have used is the LMV772. It is about a factor of 2 better than most
decent op-amps (like the LMC6482). You would have to add some
filtering before the comparator to eliminate noise outside of your
desired bandwidth.

Taking a step back, you want to turn a 2kHz sinewave into a square
wave. How much do you care if the edges move around? Do you care if
the output is random rising and falling edges when there is no signal
input?

Since you are going to end up clipping the sine wave to a fixed
amplitude, you do not need to worry about amplitude noise (except in
terms of what happens when no signal is present) but you do have to
worry about phase noise. This will show up as jitter in your
squarewave edges. If you Google for jitter and phase noise, you will
find some info. I think there is a simple formula relating average
phase noise power and rms jitter.

As long as you design your amplifier chain properly, and as long as
the first stage has sufficient gain to boost your signal out of the
noise of the next stage, then the first stage is the only one which
really matters for noise level. A simple discrete transistor
amplifier, if carefully designed, could easily yield noise in the 1uV
range in your bandwidth. Selecting low-noise transistors and using
good low-noise design methods, something like 0.05uV is attainable in
your bandwidth. That would not be easy, though.

However, with your inexperience, I think that a low-noise op-amp would
be a better choice. You'll still have to pay attention to the values
of resistors used in the circuit to prevent the addition of extra
noise. Also, 60Hz hum could really be a plague if you aren't careful
to prevent places where it can leak into your signal path.

What is the largest signal you need to be able to handle?

Finally, you will have to place filtering at the appropriate places to
make sure that the signal and noise stay in the linear region of your
amplifiers until AFTER the filter, otherwise noise could mix with
stray signals and produce a spurious signal within your bandwidth.

Sean


On Sat, Jul 26, 2008 at 1:31 AM, Rich <spam_OUTrgrazia1TakeThisOuTspamrochester.rr.com> wrote:
{Quote hidden}

> {Original Message removed}

2008\07\26@130222 by Mark Rages

face picon face
On Sat, Jul 26, 2008 at 12:31 AM, Rich <.....rgrazia1KILLspamspam@spam@rochester.rr.com> wrote:
> The more stages you add the greater will be the noise problem.

No, noise performance is determined by the first stage or two.

Since noise performance can sometimes be traded off against gain,
sometimes adding stages is helpful to build a low-noise amplifier.

Regards,
Mark
markrages@gmail
--
Mark Rages, Engineer
Midwest Telecine LLC
markragesspamKILLspammidwesttelecine.com

2008\07\28@072145 by Rich

picon face
My point is not to misinform, but to shed a little clarity on the subject.
It is a good engineering practice to begin with a specification that defines
your design criteria. Newbees especially can use such advice.  And what is
presented on the list is observed also by those who are just entering
engineering and learning.  What is considered an acceptable noise level in
one application may be considered excessive in another.  If a single stage
single transistor or FET will work it is probably a better solution with
respect to noise.  I don't have anything against "rule of thumb" design, it
has a valid place in engineering.  But to say that the noise contribution
does not present itself after the first two stages is a bit misleading to a
novice or a student. You may say that the majority of the noise is in the
first stage.  When you amplify, you amplify both the signal and the noise
which is passed on to additional stages.  It would be wonderful to get a
small signal out of the noise by amplifying it up but that just can't
happen. Every stage that has active components will contribute to the
overall noise.  There is no active electronic component that I have ever
been aware of that does not contribute to noise.  If electrons move there
will be noise.  Even resistors contribute to noise.  That is apparent when
one looks at the physics of the devices. In the final analysis the formal
design begins with a design specification, and that specification includes
an error budget which also accounts for the noise; stage by stage.  I have
not been convinced that such an approach is unnecessary or an unimportant
consideration in engineering.  Perhaps at the hobby level one can dispense
with the specification but not when you are designing product for some
company.


{Original Message removed}

More... (looser matching)
- Last day of these posts
- In 2008 , 2009 only
- Today
- New search...