coil wire?

[Davor]

Of course, centre tapped coil is simply reducing only the common mode signal, and when you observe a Tx coil as a strong source of a common mode signal through the distributed capacitive coupling, this differential setup makes sense. With unbalanced Tx coil a net potential seen fron capacitive coupling is half the Tx voltage. With either Rx, or Tx, or both coils in balanced mode this transfer is suppressed.

Maybe we should try to harmonise terminology a bit. When saying "differential" it can be regarding two completely different things. There is a differential mode of symmetrical signal feeding, and there is also a differential pickup that enhances the difference signal between coil parts and also suppresses the far field. Perhaps "differential mode" would be appropriate for a symmetrical signal feeding path, and "differential coil" for the figure 8 and top hat coils.

I expect differential mode front end to improve common mode behaviour of all coils, and especially those with centre tap, so much so that no shielding is necessary. It works for me on my IGSL. Point is that mutual inductance in case of a common mode effectively nulls the coil inductance, so most of the common mode signal ends up in a centre tap. Unlike shielding this is a very low impedance path.

I made some simulations of an IB coil system with and without shielding, and concluded that it hurts high resistivity materials response only by some small percentage, a dB or so. I also found out that placing shield exclusively on Rx (or Tx) is a bad idea. It seem that the bad frequency response effects of a shield are balancing out when it is applied on both coils.

[thehillfolk]

would this cable work OK for a Gary Pulse 1 detector ?



http://www.jaycar.com.au/productView.asp?ID=WB1534

[Midas]

Now this is something I find puzzling. I have this week built a differential coil of the Moodz type – twisted pair, bifilar wound, centre tap to ground etc., and modified a circuit board to have a differential front end. All works well, except that it is no quieter in terms of the pickup noise on the coil, than a similar size monocoil. All the pickup is sub 1Mhz noise as can be seen on the scope. We have high power transmitters in the UK from 100kHz through to 210kHz. There are also two French broadcast stations round about 190kHz. I have a communication receiver which covers 75kHz – 200kHz on its lowest frequency band, so I can hear what is going on. Hold the detector coil vertical and rotate it and you can get a null or a max, so you can use it to find the direction of a station. Even sometimes see the modulation and that it coincides with music or speech from the radio. Now this type of signal pickup is the magnetic component of the electromagnetic wave and appears as a differential signal on both a standard coil or a differential coil. Only the electric field would be common mode. Shielding the coil with copper fabric emc tape, or even lead foil makes little difference to this type of pickup. I shielded the differential coil with no difference in noise level. You could of course use solid copper tape, but then all you would see is the eddy current decay in the tape. The only ways I know of drastically reducing VLF noise is to use a figure 8 type coil or a differential coil with a second receiver stacked vertically above the lower TX/RX.

This is not to say that a differential coil does not have other positive benefits but I cannot see that it can reduce noise within the band that mostly affects PI detectors i.e say 1Mhz – 1Hz where the coupling to the coil is the magnetic component.

Eric.

I found the same thing, and it makes sense if you think about it. Moodz's Differential Coil (MDC) is really just a mono with the two coil halves occupying the same electrical and magnetic space. Both halves are still wound in the same direction so unlike the figure of 8 or the top hat, there is no cancelling going on for EMI. Its advantage is that it makes it very easy to create a coil with two identical halves that can create balanced capacitance with its environment so potentially no shielding. However apart from the extra work involved in fitting it I don't see anything really wrong with having a shield. Also MDC has has quite a large disadvantage for a mono in its massive inter-wire capacitance.

Midas

[Davor]

Capacitance is larger with bifilar wound coils, but so is with a shield. What goes around...
Yet unlike shielded bipolar mono, centre tap enables you to drive your coil in bipolar fashion.

BTW, I made some sims with idealised ferrous targets against idealised non-Fe targets with idealised pulse drive, separate Rx coil, and in log scale. Funny thing that happened was that at the moment field collapses the non-Fe targets produce a spike response, while such spike is not there with Fe. Combining these targets produce a ski-like response with a spike, same as the Fe response of the real world PI. My conclusion is that the real world coil is responsible for the spike and the ski response by means of eddy currents in coil wire. The idealised coil would not produce such a spike with Fe, yet the spike is there and it says "I'm non-Fe".

IMHO, the really fast coil would not only have low inductance and resistance, but will also be wound with Litz wire to reduce the coil eddies that reflect to the coil response.

[Ferric Toes]

The fastest coil I have made is 0.25mm wire wrap wire, teflon insulated, 2 ohms resistance and 30uH inductance. The coil resistance is of secondary importance as there was 27 ohms in series with it. Coil TC is therefore about 1uS. The TX rep rate was 16.6K pulses per sec. which compensates for the much lower TX current in the resulting S/N ratio. The important thing is that there was virtually no cable between the coil and the electronics, hence the resonant frequency was high (can't remember what).

The spike you see with non-ferrous may be due to the initial surface current induced in such an object. It's all down to skin effect, but the very early time is a very fast high voltage transient in a very thin surface layer which is trying to maintain the initial field just prior to switch off. There is then what is analogous to a diffusion process as the eddy currents move inwards until the field is uniform within the object. This is why a non-ferrous solid object can be represented by a sum of exponentials during the "diffusion" period until only the fundamental tau remains. Then it is only the one exponential decay from then on. The exception is a ring or thin cylinder as there is no metal to diffuse into. Hence just a single exponential. This is something to bear in mind when doing object simulations. Sampling is nearly always done while still in the early time period from the S/N consideration, and the true fundamental tau has not yet been reached.

Eric.

[Davor]

So in fact a Litz wire of parallel running 0.25mm wires should maintain a 1us response time. Guess if I make a flat braid to resemble a ribbon I can use it to wind a concentric coil and obtain extremely fast coil. Wonder what would happen if I (ab)use the very same braid as a basis of a bifilar coil.

[RING-N-THINGS]

I know this is an old post,Im looking to make a coil like your built..but can you tell me how did you make you coil ..I mean what did you do to get the coil wire inside,I see you use microphone wire from the base but like to know about wires inside the coil.Would really appreciated your time and help.MIke

[green]

The fastest coil I have made is 0.25mm wire wrap wire, teflon insulated, 2 ohms resistance and 30uH inductance. The coil resistance is of secondary importance as there was 27 ohms in series with it. Coil TC is therefore about 1uS. The TX rep rate was 16.6K pulses per sec. which compensates for the much lower TX current in the resulting S/N ratio. The important thing is that there was virtually no cable between the coil and the electronics, hence the resonant frequency was high (can't remember what).

The spike you see with non-ferrous may be due to the initial surface current induced in such an object. It's all down to skin effect, but the very early time is a very fast high voltage transient in a very thin surface layer which is trying to maintain the initial field just prior to switch off. There is then what is analogous to a diffusion process as the eddy currents move inwards until the field is uniform within the object. This is why a non-ferrous solid object can be represented by a sum of exponentials during the "diffusion" period until only the fundamental tau remains. Then it is only the one exponential decay from then on. The exception is a ring or thin cylinder as there is no metal to diffuse into. Hence just a single exponential. This is something to bear in mind when doing object simulations. Sampling is nearly always done while still in the early time period from the S/N consideration, and the true fundamental tau has not yet been reached.

Eric.

I have suggested faster decay of non-ferrous targets at start of decay is due to skin effect, was told due to something else. Missed this thread, before I joined. Think Eric is suggesting faster decay at start of decay is due to skin effect. Don't know which is correct. Eric states a ring has a single exponent, not what I get with thicker wire. A test to see if I can tell what is causing faster decay at start of decay. Think decay current travels around wire with ring(gap not soldered). Travels around ring(butted and soldered). Possible it could travel around wire and around the ring? Don't think so. AWG19 ring butted and soldered, AWG10 not soldered and AWG10 butted and soldered decay faster at start of decay. AWG19 not soldered is straight line decay. Thought 60us constant rate Tx might have something to due with it. Repeated test with 2ms constant current Tx. Time for decay to straighten appears to be the same, constant rate or constant current. Time for soldered ring to straighten is longer than time for not soldered ring to straighten. Not positive test equipment isn't causing some of the faster decay at start. Maybe someone can explain why faster decay at start or what should happen.