Hi Geoff,

At 07:30 AM 7/18/2010, you wrote:
>Hello Mr. Nordgren;
>
>The question I have is thus:
>
>Is that coil right for the magnets or,
>Should it be rectangular in shape with the
>two vertical sides of the rectangle outside the magnetic flux ?

Ideally you want your sensor to be linear.  If you move the coil at a 
constant speed over its mechanical range you would like to see a 
relatively constant voltage out.  You can achieve that in two 
ways.  Have the magnet pole faces large enough that all the coil 
wires stay well within the magnetic region, or alternatively, make 
them small enough that all the magnetic lines stay within the region 
filled with coil wires.  When you have the edges of the magnetic 
field moving across the outside or inside edges of the coil, the 
linearity suffers, though perhaps not enough to worry about too much.

>I figure I may need like 1206 feet of 36 AWG(B&S) copper enameled wire
>for a 2100 turns 500 ohm coil ?
>
>The 2100 turns are of enamled wire without a heavy coat of enamel.
>
>Is 2100 turns enough ?

I guess the question would have to be, enough for what?  It all 
depends on how sensitive you plan to make your signal detection 
circuit--that is, how much amplifier gain do you plan to have and if 
you are connecting to an A/D device, what is its sensitivity?  In 
general, I would try to start with the smallest wire and the largest 
number of turns I could easily manage.  Coil resistance of several K 
Ohms wouldn't be unreasonable.  The coil Larry sells I believe has 
10,000 turns and is 9,000 Ohms.

>Would it be better if the coil had a copper or iron core ?

Definitely no.  Iron would "suck in" the magnetic lines, away from 
the wires where you want them to be.  Copper, unless it made a 
complete loop, wouldn't do much since it wouldn't be in the magnetic 
field, certainly nothing particularly helpful.

>This would mean a custom coil rectangular instead of circular.

Possibly you'd get slightly better linearity with a rectangular coil 
and rectangular magnet pole pieces, but either shape should work 
reasonably well for what you are trying to do.

>I think it may be possible to have two identical
>coils center tapped in the middle with a single
>rare earth magnet in the middle between the two
>coils then you have the right setup for
>a proper op amp differential circuit.

I may be wrong, but I am suspecting that you are wanting to connect 
one output to each of the two inputs of an op amp.  If that is what 
you are thinking about, the problem will be too much gain.  Typical 
op amps have voltage gains of 100's of thousands, or more often, 
millions.  Generally for approximate analyses designers assume that 
their gain is infinite.  That means that extremely tiny input signals 
(noise) would have the op amp output bouncing between its voltage 
limits--not very useful.

Usually op amps are connected up as single ended amplifiers using a 
two-resistor feedback circuit, which makes a very nice voltage amplifier.

See: 
http://www.analog.com/static/imported-files/tutorials/MT-032.pdf   Figure 3.

If you do that, one coil will work fine, and by changing the resistor 
values you can choose the amplifier gain to complement your coil 
sensitivity.  The only advantage I can think of for some kind of 
differential coil setup is that it might not be as sensitive to 60 Hz 
hum.  Though it probably wouldn't be that hard to connect a 
differential coil to a single-ended amplifier, sort of like a guitar 
"hum-bucking" pickup.

An instrumentation amplifier *would* allow for independent 
connections to a pair of coils, but they tend to be a lot more 
expensive and their gain usually can't be adjusted as precisely.

See: 
http://www.analog.com/static/imported-files/tutorials/MT-032.pdf  Figure 2.

>+COIL-  NmagnetS +COIL-
>
>THE left coil - is connected to right coil +
>which is then the ground.
>
>Left Coil + goes to op amp +
>and
>Right coil - goes to op amp -
>
>The relative motion is magnet fixed to ground
>and coil fixed to device.
>
>Coil is stable and magnet moves right and left
>between the coils.
>
>Possibly +/- 2mm of range of motion.
>
>Each coil having 1050 to 2100 turns
>custom wound to be exactly the same.

Hope that's a start,
Brett



__________________________________________________________

Public Seismic Network Mailing List (PSN-L)