Maker Pro
Maker Pro

Differential Pair Input Noise Issue

M

MooseFET

Here is picture:

http://www.flickr.com/photos/26914086@N05/2772737582/sizes/o/

I didn't figure out why "run a lot of current ot reduce the voltage
noise". And how to pick the circled resistor?

I suggest you look at the LSK170 if you want lower noise levels and
can stand the input capacitance.

The basic rule for resistors in low noise circuits is to make them
either as large as you can or as small as you can. The one on the
gate of the JFET should be small or perhaps zero.
 
Here is picture:

http://www.flickr.com/photos/26914086@N05/2772737582/sizes/o/

I didn't figure out why "run a lot of current ot reduce the voltage
noise".

Noise current is proportional to the square root of the total current,
and resistor (Johnson) noise is proportional to the square root of the
resistance value; by going for a higher current and lower resistors at
that point, the designer has minimised the noise contribution atthat
particular point.
And how to pick the circled resistor?

Small enough the op amp doesn't swing to the rails (and saturate its
internal gain paths) under normal operating conditions.
 
thanks, your explanation is the best even though I read your post for
a couple of times.
2k ohm is a big value based on my understanding of thermal noise
that's why I posted this question here. Right now I understand the
author's words-he thinks 2k ohm is small enough to minimize thermal
noise. In stead I think the author should pick a op-amp running under
lower supply voltages then pick a smaller resistor.
 
P

Phil Allison

thanks, your explanation is the best even though I read your post for
a couple of times.

** Do not top post !!!!!!!!!!

2k ohm is a big value based on my understanding of thermal noise


** But the voltage noise of those 2N5486 fets is much higher - and it rises
sharply at low frequencies.


that's why I posted this question here. Right now I understand the
author's words-he thinks 2k ohm is small enough to minimize thermal
noise. In stead I think the author should pick a op-amp running under
lower supply voltages then pick a smaller resistor.


** Did you come here to ask questions or give lectures ?




..... Phil
 
** Do not top post  !!!!!!!!!!


** But the voltage noise of those 2N5486 fets is much higher -  and it rises
sharply at low frequencies.

It's usually not as simple as that; You need enough gain in the first
stage to make sure that the op amp input noise is lower than the noise
from your long-tailed pair of FETs; these are two independent noise
sources so the total noise is the square root of the sum of their
squares, which means that it s rarely worth getting the lesser of the
two noise sources below half the bigger one.
**  Did you come here to ask questions or  give lectures ?

Don't mind Phil - he's obnoxious, but he does know his electronics.
 
Top