Flyback Effects with Dual Coils

[MagnetMan]

Hi Everyone,

I'm new to metal detectors and Pulse Induction. I've built three variations of Gary's Pulse-1 & 2 and am quickly learning how they work, but I am getting to the point where I need to ask questions in order to obtain a more deep understanding.

I read that separating the TX and RX coils improves sensitivity by reducing the flyback voltage but I don't understand why a collapsing field does not induce a similar flyback voltage in the RX coil. Is it because it isn't trying to maintain the same flow as the TX coil? Does a RX coil still saturate the receiving circuit when the pulse is completed? Lastly, does reducing or eliminating the flyback significantly shorten the time before the signal can be sampled?

Thank you all for such a rich resource of information!

-Stefan

 

[MagnetMan]

n/t

 

[Dave J.]

The trick is to separate the transmit and receive coils using a geometry which achieves "induction balance" -- that is, the transmitter does not couple net magnetic flux into the receiver. The most common form of induction balance coil used in PI's is the double-D.

No matter how carefully you try to align the coils to achieve balance, you won't get a perfectly clean signal in the receiver, because of parasitic circuit elements (the things that really exist but which you didn't draw on the schematic, like stray capacitance).

In general, using a separate receiver permits little if any reduction in receiver turn-on pulse delay because of the additional LCR delays in the receiver coil circuit.

--Dave J.

 

[MagnetMan]

Thanks, Dave.

Is a double-D coil truly the ideal geometry? Seems that there is so much hinging on proper coil design that there might be additional, unexplored, layouts to better balance them. I have read that around 10

 

[Dave J.]

There is no "ideal" geometry, Stefan. The DD configuration is simple and relatively easy to work with, that's why it's popular.

Whether or not a separate receiver coil will allow shortening the delay time depends on the details of the coils, the details of the circuit, and how much noise and drift you're willing to listen to as a result of pushing the receiver timing too close to the flyback pulse. The only way to know is to actually design and build something and see what you've got, and then fiddle with it to find your best compromise.

--Dave J.

 

[MagnetMan]

I'm amazed that there are so many aspects of design that don't have one solution! I'm not suggesting that someone could draw me a diagram of the perfect coil, but an "ideal" in a theoretical sense. Has there been much work done on alternative coil designs? I recognize that there are some guidelines in terms of parasitic capacitance, higher "Q" value, etc, but apart from altering the materials and some winding techniques, have there been any successes in other approaches?

-Stefan

 

[Eric Foster]

Hi Dave and Stefan,

Coils are fascinating things and they often don't work out in the way you would expect. A while back I built some 11in DD coils where the two windings were identical in their construction. i.e same resistance and inductance. Both windings were screened with wrap around lead tape, and both screens were properly grounded. The correct overlap was made by observing the preamp output and adjusting so that there was a flat zero line during the transmitter period. I was expecting to be able to use a considerably shorter sample delay, but in fact there was little difference to that of a standard mono coil. The main benefit of a DD is that you get less near surface ground signal.

Eric.

 

[MagnetMan]

Thanks for your insight, Eric. Since part of our goal is to reduce the sample delay, is it possible to use a RX coil with fewer turns so that the spike is a lower voltage? Or would that reduce target signals so much that they were unrecognizable? I've seen a few posts that mention asymetrical coils but they don't seem to go into any significant detail.

-Stefan

 

[Eric Foster]

You could reduce the number of turns, but there are better ways of doing it. Both coils and targets are basically acting as a transformer. In the case of metal targets it is a huge step down, simply a one turn loop with both very small inductance and capacitance. An adjacent coil can have more, or less turns than the TX, so can step up or down. Then in both cases, there is the coupling factor to take into account. Switching off the TX current can generate a huge voltage across the TX coil (the flyback), which is the energy stored in the magnetic field being dissipated as quickly as possible. This flyback voltage is usually limited by other circuit elements which have the effect of stretching out the switchoff time, which then delays the point when you can start sampling. For short delays and improved sensitivity to small low conductor targets, I use much lower current pulses but repeat them more often. Using this method, sampling delays of 5uS are quite easily achieved with flyback voltages of less than 100V. Primarily delays as short as this are used in industrial applications as there are no ground effects to cope with, which escalate dramatically as the delay is reduced. For example, I once tried a 5uS unit on a wet beach and it was almost impossible to use, while readjusting to 10uS was OK.

Eric.

 

[MagnetMan]

Eric, you certainly were thinking along the same lines when you mentioned transformers. If you reduce the pulse current in order to favor sensitivity to small targets won't that potentially reduce the coil's range? I was wondering if a "step down" from TX to RX would help lower flyback while having a lesser effect to coil range. For that matter, what about an additional coil that remains in a closed circuit from the switch off time to perhaps 3-4

 

[Eric Foster]

Hi Stefan,

One of the nice things about PI is that you can use a simple single coil for both transmit and receive. Also PI circuits are generally non critical regarding the coil parameters, but all that changes when you want to find very small targets, minimise ground signal, or rf pickup, or optimise to give the best range on a particular target. A lot of work has been done on coils by various detector manufacturers and aftermarket coil suppliers, but much is kept under the table for commercial reasons.

Reducing the pulse current does generally reduce range but if you can't detect a small object at 15uS delay but can detect it at 10uS with a lower pulse amplitude, then you have gained range. Also metal detection is all about signal /noise ratio. Having a small pulse more frequently can give a better signal to noise than a large pulse less frequently. This is particular so, as there is a circuit in a PI receiver that takes a running average of a sequence of sampled receive signals. Typically a "normal" PI will have a few hundred pulses per second, whereas low pulse current high frequency PI's may run as high as 10,000 pps. Superimposed noise will average out better at the higher sampling rate.

Having another coil to extract energy from the flyback pulse would actually cause the delay to be longer. In other words, the extra coil acts like a damping circuit with its own eddy currents.

Eric.

 

[MagnetMan]

Thanks, Eric. I'm certainly getting a better idea of the the appropriate characteristics of PI coils!

-Stefan