[volt-nuts] How to keep voltage stable in the sub-100nV range?

David davidwhess at gmail.com
Wed Nov 2 10:23:51 EDT 2016


On Wed, 2 Nov 2016 00:33:32 +0100, you wrote:

>On Tue, 01 Nov 2016 09:50:38 -0500
>David <davidwhess at gmail.com> wrote:
>
>> 1. Thermocouples and temperature gradients - this is a huge problem
>> and special attention will need to be directed toward the layout and
>> maintaining an isothermal environment.  Careful design is required to
>> get the specified drift performance out of chopper stabilized (10nV/C)
>> and low drift operational amplifiers (100nV/C before trimming or
>> grading).
>
>Hmm... I guess I will have to put everything into a machined block of
>aluminium or even copper/brass/bronze to keep the temperature gradients
>low and temperature variations slow.

It is both more complicated and simpler than that.  A thermal baffle
(Dixie cup) placed over sensitive circuits to prevent air flow and
convection can be enough.  Check out Linear Technology application
note 9 written by Jim Williams of course for a discussion on this:

<http://www.linear.com/docs/4105>

>> 5. Pink Noise - 1/f noise increases as frequency decreases.  Chopper
>> amplifiers have flat 1/f noise so are invaluable below about 1 Hz.
>
>I was thinking about using low noise opamps (probably LT1128 or LT1677)
>with offset compensation using a LT2057 on each of them. This
>should at least kill the 1/f noise and I would guess also most of
>the opamp induced temperature variation. Does that make sense?
>Or should I skip the low noise opamp and use the LT2057 directly
>with some second order LP filter with 100Hz-1kHz bandwidth?
>
>			Attila Kinali

The LTC2057 will cancel the 1/f noise, drift, and offset but I think a
composite amplifier is only needed if wideband noise needs to be lower
than an LTC2057 will provide.  Simple filtering will remove the
wideband noise on all except for the last stage.

The LT1028/LT1128 and LT1677 have high input bias current with its
associated high input current noise so are not suitable for buffering
high impedance (long time constant) filters.  I might consider the
LT1793 if an LTC2057 by itself was not good enough.

I wonder if a pair of LTC2057's could be used in parallel for lower
noise.  The LTC2057 has no clock pin though so intermodulation might
be the result.


More information about the volt-nuts mailing list