PSN-L Email List Message
Subject: Re: Op amp front end noise
From: John Popelish jpopelish@........
Date: Tue, 15 Mar 2005 12:05:01 -0500
> ChrisAtUpw@....... wrote:
> > John Popelish wrote:
> > Something interesting to me that does not show up in the
> > list I posted, but in the graphs is the fact that the
> > overall signal to noise ratio climbs as the coil wire size
> > is reduced, even though it results in more resistive
> > noise. But some opamps have such low current noise and
> > the extra turns provide more signal voltage, so that there
> > is a steady climb in signal to noise ratio to coil
> > resistances around 100k, and then there is a second, even
> > bigger peak for coil resistances of 100 meg ohms, but I
> > wouldn't want to handle the wire.
> The opamps have a design impedance when the current and voltage
> noise levels are about equal. If the coil resistance is less than
> this, it pays to add turns.
Agreed. Processing the opamp data to take the turns count (in a given
winding volume) into account and displaying this graphically makes
this very clear. I need to get set up to post these pictures to the
> Are you taking the 1/f noise into account? This is usually
> fairly critical for seismic sensors, particularly when you are
> considering long period types.
Yes. I selected or extrapolated all device specs to 0.1 hertz for
this comparison. Amplifiers that have a low 1/f corner frequency or
chopper amps that actively correct 1/f noise do better in this range
than if you compare amplifiers at higher frequencies. I selected this
frequency as representative of what long period instruments measure.
> > But the graph does show that there is signal to noise
> > value in going with the smallest size wire you can deal with.
> It will usually pay to choose a fairly low amplifier impedance
> for inductive systems. The coils are much easier to make and
> physically smaller, which allows you to take full advantage of the
> very high fields that can be produced by 'modern' NdFeB magnet
My list is valid for any signal source, but the rising signal to noise
effect with impedance mentioned in that last paragraph refers to coils
that fit in a given volume. This implies that doubling the turns
count raises the DC resistance by a factor of 4.
> The larger the coil, the more difficult it is to screen it from
> environmental noise.
My comments referred to coils of similar dimensions.
> In general, most of us do not have the luxury
> of quiet seismic sites. The larger the inductance, the more
> susceptible is the wiring to picking up stray signals.
This is a separate problem.
I think it is a good idea to electrostatically shielded any pickup
And the inductive pickup problem can be addresses with a hum bucking
I am presently in the middle of building a dual voice coil pickup that
is an attempt to buck out the effect of any external magnetic field
changes, like those from power distribution, lightning strikes and
aurora. I finally got my 6 iron pole pieces from the machine shop and
assembled them with 20 ferocious NeFeB magnets. I am creating the
coil bobbin now. The Lehman pendulum is almost done and I may have
everything but the amplifier together by April fools day.
> It can pay to put a ceramic capacitor across the input to the opamp.
I concur about the capacitor, but you have to be careful with
ceramics. Some of them are quite microphonic, especially the larger
values in the high K class, like Y5V and Z5U. I think I prefer film
I also like a pair of non light sensitive diodes across the input to
protect the front end from static when wires are not connected or
excessive generated voltage when the coil is bumped.
> The use of
> screened cable with a large dielectric loss can be an advantage.
I think there is value in getting at least the first stage of
amplification close to the coil, to lower the chance of stray signals
getting into the input. I may mount the whole amplifier-filter right
beside the pendulum.
Public Seismic Network Mailing List (PSN-L)
[ Top ]
[ Back ]
[ Home Page ]