Hi Chris and all,

Thanks Chris for the extensive reply below.  Actually my impression with the
stacked magnets came from
trying to utilize a old surplus Texas Instruments rectangular coil form, and
at the time, the stacked
magnets approach seemed quite impressive at the time via a table top and
voltmeter experiment.

The problem with the coil is it has relatively thin walls, and I'am not so
sure that a quad (4) magnet
setup is the best or only approach.  Its possible to use 3 magnets and
essentially get a similar "quad" (4 pole)
arrangement, where the fields go through the sides of the coil walls.

The coil form shape is quite similar to transformer coil forms, and in a
sense there maybe 2 ways to
use magnets with them.  Esterline-Angus also made similar coils back in the
1950's, of which
I also have some of those.

I put up a quick web page with 2 photos on the TI coil; along with my
"notes/ramblings" related to such:

http://seismometer.googlepages.com/tirectiwritercoilconversion

Take care, Meredith Lamb


On Fri, May 16, 2008 at 2:56 PM,  wrote:

> In a message dated 2008/05/16, paleoartifact@......... writes:
>
> Am getting to the (shall we say a unqualified) opinion that with a choice
> of round or square / rectangular coils with a magnet, that it seems like the
> round coil approach is the more efficient route to take.
>
>
>
> Hi Meredith,
>
>        If you want an accurately linear output from a Lehman, while
> allowing for the large sensor drifts that you always get, you definitely
> need to use a quad arrangement of square NdFeB magnets and a rectangular
> coil.
>
>        Round coils will DEFINITELY NOT give you a constant sensitivity over
> the drift range of a Lehman sensor.
>        If you are using a vertical pendulum, you can use a quad arrangement
> of rectangular magnets with the width of a magnet ~ the ID of the coil.
> Alternatively, you can make up a magnet block and attach it to the pendulum
> to form part or all of the mass. But you will then still get a small
> sensitivity to environmental magnetic fields.
>
>        You make the ID of the actual windings about 3/4 of the width of a
> magnet and the internal length about 1/5 greater. This allows for some
> position error either as the coil swings in an arc or for any setup position
> error.  The flux change for a given magnet movement is then CONSTANT over
> maybe 3/4" drift with 1" square magnets.
>
>        Mount the magnets on 1/4" thick mild steel backing plates held apart
> with zinc plated mild steel bolts. You need to give the mild steel some etch
> primer and a coat of anti rust paint. I use Hermetite. You paint the whole
> surface of the plates, then leave them a couple of days for the paint to
> harden. I make up a cardboard cut-out, clamp it to the plate and slide the
> magnets into position over the cardboard. This is 'the easy way to do it'!
>
>        One advantage of this construction is that you only need maybe 2,500
> turns on you coil. The voltage output is maybe 10x what you get with a U
> Alnico magnet.
>
>        This arrangement is shown on the drawings at
> http://www.jclahr.com/science/psn/chapman/lehman/index.html
>
>        It is quite easy to make a rectangular coil former out of thin glass
> circuit board / sheet glass fibre. DigiKey sell both. You can either use the
> construction shown, or you can make a wood / plastic inner coil former and
> screw / glue it between two sheets to form the ends of the coil. I suggest
> winding it using a bolt through the centre gripped by a hand drill, held in
> a vice. Use some fine plastic tube to guide the wire onto the coil. Spray
> cans usually have suitable tube if you open one up. I use two part Acrylic
> glue - Devcon Plastic Welder or similar.
>
>        I defiitely avoid using 'coil dope' - I find that I get too many
> shorted turns. You can use coil dope OK if you very carefully wind a single
> layer of wire at a time and then use tissue paper to interleave the winding
> layers.
>
>        With luck, you are only going to ever make one coil per seismometer,
> so it REALL IS worth a bit of extra effort to get the best out of your
> instrument.
>
>        Regards,
>
>        Chris Chapman
Hi Chris and all,

Thanks Chris for the extensive reply below.  Actually my impression with the stacked magnets came from
trying to utilize a old surplus Texas Instruments rectangular coil form, and at the time, the stacked

magnets approach seemed quite impressive at the time via a table top and voltmeter experiment.

The problem with the coil is it has relatively thin walls, and I'am not so sure that a quad (4) magnet
setup is the best or only approach.  Its possible to use 3 magnets and essentially get a similar "quad" (4 pole) 

arrangement, where the fields go through the sides of the coil walls.

The coil form shape is quite similar to transformer coil forms, and in a sense there maybe 2 ways to
use magnets with them.  Esterline-Angus also made similar coils back in the 1950's, of which

I also have some of those.

I put up a quick web page with 2 photos on the TI coil; along with my "notes/ramblings" related to such:

http://seismometer.googlepages.com/tirectiwritercoilconversion


Take care, Meredith Lamb


On Fri, May 16, 2008 at 2:56 PM,  <ChrisAtUpw@.......> wrote:

In a message dated 2008/05/16, paleoartifact@......... writes:



Am getting to the (shall we say a unqualified) opinion that with a choice of round or square / rectangular coils with a magnet, that it seems like the round coil approach is the more efficient route to take. 




Hi Meredith,



       If you want an accurately linear output from a Lehman, while allowing for the large sensor drifts that you always get, you definitely need to use a quad arrangement of square NdFeB magnets and a rectangular coil.




       Round coils will DEFINITELY NOT give you a constant sensitivity over the drift range of a Lehman sensor. 

       If you are using a vertical pendulum, you can use a quad arrangement of rectangular magnets with the width of a magnet ~ the ID of the coil. Alternatively, you can make up a magnet block and attach it to the pendulum to form part or all of the mass. But you will then still get a small sensitivity to environmental magnetic fields.




       You make the ID of the actual windings about 3/4 of the width of a magnet and the internal length about 1/5 greater. This allows for some position error either as the coil swings in an arc or for any setup position error.  The flux change for a given magnet movement is then CONSTANT over maybe 3/4" drift with 1" square magnets.




       Mount the magnets on 1/4" thick mild steel backing plates held apart with zinc plated mild steel bolts. You need to give the mild steel some etch primer and a coat of anti rust paint. I use Hermetite. You paint the whole surface of the plates, then leave them a couple of days for the paint to harden. I make up a cardboard cut-out, clamp it to the plate and slide the magnets into position over the cardboard. This is 'the easy way to do it'!




       One advantage of this construction is that you only need maybe 2,500 turns on you coil. The voltage output is maybe 10x what you get with a U Alnico magnet.



       This arrangement is shown on the drawings at http://www.jclahr.com/science/psn/chapman/lehman/index.html



       It is quite easy to make a rectangular coil former out of thin glass circuit board / sheet glass fibre. DigiKey sell both. You can either use the construction shown, or you can make a wood / plastic inner coil former and screw / glue it between two sheets to form the ends of the coil. I suggest winding it using a bolt through the centre gripped by a hand drill, held in a vice. Use some fine plastic tube to guide the wire onto the coil. Spray cans usually have suitable tube if you open one up. I use two part Acrylic glue - Devcon Plastic Welder or similar. 




       I defiitely avoid using 'coil dope' - I find that I get too many shorted turns. You can use coil dope OK if you very carefully wind a single layer of wire at a time and then use tissue paper to interleave the winding layers.




       With luck, you are only going to ever make one coil per seismometer, so it REALL IS worth a bit of extra effort to get the best out of your instrument.



       Regards,



       Chris Chapman