Transformer Induction coil (COIL CONSTRUCTION)

[vk6kro]

So aside from you designs, I can't simply place a primary coil followed by a secondary coil in sequence to generate a signal? (NO EXICITATION WINDING. See figure)

You might get a small signal if the steel ball happened to have some residual magnetism, but generally, no. There would have to be a magnet there somewhere.

This is sounding a little like a metal detector design project.

I saw a metal detector coil in a magazine. It had coils about the size of compact disks (ie about 12 cm) but they were flattened on one side to make a "D" shape. The two coils were then laid over each other with the curved parts facing away from each other.

like this:
http://dl.dropbox.com/u/4222062/metal%20detector%20coils.PNG

The idea was that the vertical part of each "D" would get opposing signals from the vertical and curved parts of the other "D" shape and you got a cancellation.

The rectangular coils at right would probably give the same effect. You would feed the signals in where the gaps are in the vertical sections. You would only feed the ends of the windings, of course.

When you had cancellation, bringing any metal into the vicinity would cause the signal radiated from one coil to be reradiated into the other coil.

I have never tried it but it seems like a good idea.

 

[Bob S]

Mike's coil-3.jpg in Post 22 is the right way to do it. The two pickup coils, if properly balanced (bucking mode), will give a null signal unless a conducting ball rolls inside the pipe. It is important that the two pickup coils be symmetric about the excited coil. If necessary, the coils could be moved back and forth on the PVC pipe to get a null.

Bob S

 

[Mike_In_Plano]

Hey VK,

Yep, the twin D is common for metal detectors. With the sensing coil overlapping the excitation coil's loop, you can get rejection. It's just a matter of tweaking.

What's really interesting about these detectors is that you can amplify / limit the output of the sensing coil and run that through synchronous detectors. The end result is that you get the phase relationship between the excitation signal and the return signal. This tells you natural time constant (L/R) of the target. Hence you get a good idea what kind of metal is present ;)

 

[Bob S]

It is useful to look at the circuit design used in the coin pass/ fail detection units in vending machines. They use half a pulsed ferrite pot core on one side of the coin, and a L/R sensing pickup (the other half of the pot core) on the other. It is very fast.

Bob S

 

[jegues]

Mike's coil-3.jpg in Post 22 is the right way to do it. The two pickup coils, if properly balanced (bucking mode), will give a null signal unless a conducting ball rolls inside the pipe. It is important that the two pickup coils be symmetric about the excited coil. If necessary, the coils could be moved back and forth on the PVC pipe to get a null.

Bob S

So this is the design I should be aiming for in order to obtain the best consistent signal?

 

[Bob S]

Mike's coil-3.jpg in Post 22 is the right way to do it. The two pickup coils, if properly balanced (bucking mode), will give a null signal unless a conducting ball rolls inside the pipe. It is important that the two pickup coils be symmetric about the excited coil. If necessary, the coils could be moved back and forth on the PVC pipe to get a null.

So this is the design I should be aiming for in order to obtain the best consistent signal?

Yes, IMHO.

Bob S

 

[jegues]

Also, how crucial is it that the two pick up coils are exactly identical? Will it make a huge difference if one pick up coil has more turns than the other?

Remember, the two pick up coils would be consisting of one 15m strand of very thin wire, it might be difficult to have the two coils be symmetric, no?

 

[Mike_In_Plano]

If your coils are off, you can scoot them closer or further with your thumb nail until you get a null. Or, you can use the orthogonal coil. I gaurantees a null simply by scotting it from left to right.

Mike

 

[Bob S]

If your coils are off, you can scoot them closer or further with your thumb nail until you get a null. Or, you can use the orthogonal coil. I gaurantees a null simply by scooting it from left to right.

The optimum design is probably a ~100-turn primary coil and two symmetric ~50 turn bucking pickup coils. The ideal location is with each pickup coil probably centered over each end of the excited coil, to maximally intercept the dipole field from a conducting (or magnetic) ball. This would require some electrostatic shielding between the excited and pickup coils, however, so probably starting out with Mike's end-to-end coil layout in coil-3.jpg is best. I suspect that the excitation frequency will be in the 20 kHz range. The optimum frequency will be when the skin depth is ~equal to the radius of the balls. If it is too high, you will not be able to discriminate between aluminum and steel balls, for example. Do you know whether the excitation is pulsed or CW?

Bob S.

 

[jegues]

The optimum design is probably a ~100-turn primary coil and two symmetric ~50 turn bucking pickup coils. The ideal location is with each pickup coil probably centered over each end of the excited coil, to maximally intercept the dipole field from a conducting (or magnetic) ball. This would require some electrostatic shielding between the excited and pickup coils, however, so probably starting out with Mike's end-to-end coil layout in coil-3.jpg is best. I suspect that the excitation frequency will be in the 20 kHz range. The optimum frequency will be when the skin depth is ~equal to the radius of the balls. If it is too high, you will not be able to discriminate between aluminum and steel balls, for example. Do you know whether the excitation is pulsed or CW?

Bob S.

Thanks again for your input Bob.

I'm not sure whether the excitation is pulsed or CW, I'm sure if I asked he would tell me but as of now we aren't given any detailed infromation about the circuit, just that it resembles a transformer.

We're going to be hopefully testing out coils out today so I'll let you know how our trial and error goes.