Chris Chapman's Lehman Design


Prototype

    Although this is a horizontal instrument, the damping- and sensor-system designs work equally well with a vertical instrument.

Damping Magnet System

    Oil damping was ditched by the professionals somewhere in the late 1940s. It is relatively difficult to adjust, potentially messy, damps the higher frequencies selectively, is VERY temperature dependant and drowns bugs to give 'bug quakes'.
    Electromagnetic damping with quad NdFeB magnets is not expensive, not significantly temperature dependant, has a linear response, is clean, dead easy to adjust and very effective. Let me know if you want further advice / help.  Why accept operational problems / headaches which you do not need to have? 

    There seems to be a shortage of information on magnetic damping methods and on high sensitivity, high linearity coil sensors. The damping force required depends on the mass and the period. The force increases and as the mass increases and also as the period is decreased.

    Some amateur designs still use unnecessarily large coils, massive magnets, and oil damping. It is unnecessarily difficult to construct such sensors and it is much more expensive.

    Considering magnetic construction materials, I can buy 2" wide by 0.25" thick bright rolled mild steel strip, which is adequately straight and flat. It seems to be readily available. I can also buy 1/4" zinc plated mild steel bolts, nuts and washers.

    I dip the cut, filed and drilled backing plates in Phosphoric Acid supplied for car repairs for a few minutes and then wipe and dry them - I do NOT rinse them. I coat most of the plate with Hammerite rust protecting paint, but give the magnet and bolt contact areas just a smear of CL chassis grease, which has rust inhibitor in it.

    I've checked for any magnetic field penetration through the 1/4" mild steel plates and they seem fine with the magnets and separations specified, but 1/8" thick mild steel is definitely NOT adequate. If you wished to reduce the backing plate area with less than a 1/2" border between the magnets and the edge of the plate, or chose thicker magnets, you might need to increase the backing plate thickness.

    It is a definite advantage to place both the sensor and damping magnets on the seismometer frame. This avoids any magnetic interaction and gives the lowest sensitivity to magnetic interference from changes in the Earth's field and from pulses on the house utility power wiring.

 

Scale drawings of Lehman damping system.  The bolts are zinc plated mild steel and form a magnetic 'short circuit' between the mild steel backing plates.
      You can adjust the damping both by sliding the magnet system over the damping tongue, from clear of the magnets to overlapping them by 1/2" and also by varying the magnet spacing. Note the orientation of the rectangular damping magnets. In an application which required a very low stray field, I used lathe turned mild steel tubular spacers to separate the backing plates, fixed in place with mild steel screws - but the gap was then fixed. Copper damping tongues may be preferable. The mild steel backing plates were dipped in phosphoric acid and then dried, for corrosion protection. Tannic acid is preferable if you can get it. The exterior surfaces were painted. The magnet contact areas were given a film of corrosion inhibiting grease. Another option is to nickel plate all the mild steel components - OK for a production run, costly for a 1 off. You may find that thinner magnets will give you the damping performance required. This force decreases as the period is increased. Note that the thickness of the backing plates is chosen to completely contain the flux from the magnets - there should be no significant field penetration.
      To ensure precision alignment of the bolt holes, I marked out and centered the holes on one plate only, then clamped the two plates together and drilled through both of them. I then marked one edge of both plates with a centre punch for future orientation - there are too many ways of 'getting it wrong'!

Sensor Coil/Magnet System

Scale drawings of Lehman sensor system. The bolts are zinc plated mild steel and form a magnetic 'short circuit' between the mild steel backing plates.

My current sensor systems use a rectangular coil and pairs of 1" square x 1/8" thick NdFeB magnets. This allows about +/-5/8" movement with excellent linearity and a high sensitivity. The bolts are zinc plated mild steel and form a magnetic 'short circuit' between the mild steel backing plates. This gives both magnetic and electrostatic screening for the coil and a very low stray field.

The coil that I use came as a spare for a 220V water valve on a washing machine and it is about 3/8" thick by slightly over 1" OD, about 5,000 turns. 

This coil and magnet arrangement could also allow be used to make an effective force-feedback system.

    I note that some amateurs use huge (15,000 turns!) round unshielded coils and large open U Alnico magnets, presumably to increase the sensitivity. The 5,000 turn circular coil which I used initially with the 1" square x 1/8" quad NdFeB sense magnets tended to overload my amplifier with background noise on the low gain setting, so I halved the number of turns and designed a rectangular version to give a greater linear range. Using this box construction, with the mild steel backing plates connected together with mild steel bolts, has the additional advantage of providing excellent magnetic and electrostatic screening for the coil, while providing an intense central field for the coil.

Making the Coil

    The coil is as shown above. I cut the top plate (30x40 mm) and the bottom plate (40x80 mm) from old 1/16" glass board. I saw and file two matched slots in the top and bottom plates.  The center lines of the slots are 18.5 mm apart.  I then stick two straight strips of circuit board (8 mm x 30 mm) into the slots with two component acrylic glue, scrape off excess glue before it sets hard and later smooth the faces with a file. I chamfer off all the sharp edges, and round off the internal corners.  I drill a hole through the center of the coil form, mount it on a mandrel and fit it in a hand drill clamped in a vice. I thread the wire through a 1/16" plastic tube (ex spray can) to locate the wire position and start counting turns! It does not take all that long to put on 2,500 turns. I use polyurethane coated 0.1 mm Cu wire - 38 AWG, 660 Ohms / 1000 ft. The coils come out at about 650 Ohm. You use a hot soldering iron, which melts the  insulation - you don't have to strip clean the wire.

 

Back