Coil shielding and overshoot

[Gavinius]

Hi to all,
In my PI project, I've noticed a very strange effect that I can't explain.
This is a picture of coil decay signal just after being amplified about 1000 times.

vert: 2V/div hor:5uS/div

Instead of decaying to 0 the signal presents a positive "overshoot". Does not depend by capacitance cause the overshoot doesn't change if I replace damping resistor with a smaller one, doesn't depend by the opamp cause I tried many parts and this overshoot is always visible even with faster opamp models.

The only way to eliminate this effect is to remove coil shielding. Actually the shield is made with kitchen aluminium.

The question is: why the shield can cause this?

[bbsailor]

Overshoot

Gavinius,

The kitchen foil acts like a target. As you reduce delay, shields made of less conductivity are actually detected and could extend the RX zero point sampling window. Kitchen foil is pretty thick and eddy currents can form on it.

Try using a silver (color) mylar baloon, the kind that is filled with helium, and is used for various celebrations. This mylar coating is conductive and is very thin and should act like a good shield and not be detected. Just use an ohm meter to check for and measure continuity on the surface of the mylar baloon. You should get a low ohm reading with the probes an inch or two apart.

When you look at the scope signal at the preamp for the coil with the mylar shield, the overshoot should be gone.

I hope this helps.

bbsailor

[Gavinius]

Thank you bbsailor, when I have some time I'll try with mylar shield.

[Ferric Toes]

As bbsailor says, kitchen foil is generally a bit too drastic for shielding. However, looking at your scope display, it is well on its way to recovery at about 20uS, except that it overshoots, so you will still get a reasonable sensitivity to coin size targets.

I have seen this effect many times, and it is due mainly to coil/shield capacitance. The fact that it disappears when you remove the shield indicates this, as the response is the opposite polarity to what a metal target would give. Try spacing the shield further out from the coil with some polythene spiral wrap. This will give some improvement. It is almost impossible to get rid of completely, except by severe over damping.

Try this little trick. Get some brass washers of different sizes, M2 - M5 range, and lay one M5 horizontally next to the coil. You will see the overshoot lessen somewhat. Try two washers and note the effect. By selecting a washer with the same eddy current decay as the overshoot, you can virtually cancel it out, and end up with a flat response. You may need two or three washers of the right size to get sufficient amplitude. Once you have got the response right, stick them down with epoxy, so they can't move relative to the coil.

I use lead tape for shielding which is very effective, or copper braid tape, which is not easy to get. Aluminium, will limit the sensitivity, unless you use something thinner than kitchen foil. I've not tried aluminium coated Mylar myself, but provided you can get a good ground connection, it should be OK.

Eric.

[Gavinius]

Thank you Eric, I'll try your ideas too, but I have some quesitions about tape you mentioned.
Where do you find the lead tape? I can get some copper tape, the one used to shield EMI in electronic devices, but I never heard before of lead tape. How thin is? Is it heavy?