Parallell or Serial coil windings?

[tohu]

Hello,
This may be a newbie question but I'm wondering a little about the strength of the magnetic field generated by a coil (PI). I've read that the wire-gauge and number of turns affect the strength of the field.

But I'm curious if the coil necessarily has to be made of a single wire, would several parallell wires, in theory, produce the same effect?

[bbsailor]

You can study the characteristics of coils in the abstract and find some design rules that seem universal, but when you apply those universal rules to optimizing a coil for seeking a particular class of targets, some of these universal rules may clash.

Coils must be analyzed together with the following PI instrument circuit design parameters. (1) Pulse Per Second (PPS) Frequency, (2) Peak Coil Current, (3) Coil Resistance, (4) MOSFET on-resistance, (5) Peak Coil Flyback Voltage, (6) Main Sampling Delay (7) MOSFET voltage.

Large diameter coils, operating at high pulse peak current are operated at low PPS frequency, have long sampling delays (above and beyond 30uS) and do not need shielding. Now when you try to design a coil for small, low conductive objects like gold, you need a smaller coil (relative to the size of the target), a coil that can operate at low sampling delays (10uS or lower) and a coil that now must be shielded to minimize the ground effects. Shielding adds to the coil capacitance and tends to extend the minimal delay slightly. My own crude measurements indicate that each 100 pf reduction in capacitance gains me about 1 uS in potential coil speed. This is probably not a linear figure, but sort of a rule of thumb to appreciate the interrelationship of some abstract variables.

To make a good gold coil, do I follow the universal rules and make the coil more sensitive by winding more turns or design the PI circuit to sample sooner when the receive (RX) signal is stronger? If I wind more turns, I make the flyback voltage higher and I may exceed the MOSFET voltage capability. Also, higher flyback voltages take more time to settle down and tend to extend the minimum possible sampling delay. I also, by having more turns, make the coil have more capacitance, thus limiting the minimal delay I can achieve. The answer is to sample sooner with a coil that will have a low enough coil capacitance to operate at that delay. Now I must be creative with my coil design, and size (11 inches in diameter at 300 uh), possibly use Teflon insulated wire, or a basket weave or other low capacitance coil winding technique, a shield spacer to reduce the capacitance of the coil allowing it to operate at a lower delay. If I want to operate a very low delays at about 5uS, I may even need to make a coil that uses Litz wire (individually insulated strands wound in parallel) to reduce the eddy currents in the coil wire itself and keep the wire from being detected or extending the minimal possible delay.

So, with this example you can see that universal rules are good to understand, but you also must know how to apply these rules to making a coil that will operate well at the chosen PPS frequency, sampling delay, coil size, and peak current for the class of targets that you seek.

There is a wealth on information on this forum about coil design. Use the search feature on the following coil-related key words.
Coil Capacitance
Delay
Damping Resistance
Damping Resistor
Inductance
Coil Current
PPS (Pulses Per Second)
Frequency
Pulse Width
MOSFET
MOSFET Voltage
COSS (MOSFET output capacitance)
Shield
Shield Spacer
Shield Capacitance
Flyback Voltage


Also, read the latest Hammerhead article written by Carl Moreland. This article contains an excellent explaination of how a PI circuit operates.

bbsailor

[Sid]

Hi, let´s try another consideration. The response of detector is proportional to a derivation of magnetic field strenght/intesity/. Intesity of magnetic field in a coil depends on the number of turns and current, thus H ~ I n = U/R n , sorry, I cannot do a better fraction, voltage and coil resistance limits the current, = U/rn n ,resistance of the coil is a summ of resistive turns, so the intensity of a magnetic field depends only on U/r the ratio voltage across coil and and the resistance of single turn. Then derivation it is just a question of the coil resonant frequency and of the switch. Nice day, Sid.

[Kev.]

Been thinking bout this too!

Hi,
The magnetic field density equals the coil inductance times the current through the coil, divided by the number of turns. Therefore to amplify the field you either need to reduce the number of turns while keeping the inductance and current the same, or increase the current and or inductance while keeping the turns the same.

As bbsailor has pointed out the inductance and current have far reaching performance affects, depending upon what you are seeking to detect. Another factor is the time constant of the coil, which we need to turn on slowly but switch off quickly, this is one reason I think Minelab use Litz wire, a slow current build up in the coil reduces the back emf generated and so the settling time of the target and coil is likewise reduced, thus when the flyback occurs the maximum benefits of the peak current change produces the biggest response from the target and least from the ground minerals due to it's longer time constant.

OK my point. Increasing the inductance while keeping the number of turns the same. I'm wondering if small amounts of magnetite or other ferrite type minerals encased in the coil housing would improve the field strength?

tohu this doesn't answer your question, but raises others. I'm not sure what you mean by parallel wires, Litz wire is many wires running in parallel like the strands of twine twisted and woven making up a rope, each strand of wire is discrete, but also integral to the whole, if you get the drift. In this way a medium inductance coil with very minimal self resistance and maximum current capacity can be made.
Cheers
Kev.

[Carl-NC]

But I'm curious if the coil necessarily has to be made of a single wire, would several parallell wires, in theory, produce the same effect?

Durn good question... my gut feel is 'no'. Ferinstance, if you replace a 20-turn coil with a single turn of 20 parallel wires, the 20 parallel wires will look like a single turn of stranded wire. But then I thinks a bit more... if those 20 parallel turns are insulated, and they all miraculously carry exactly the same current, then would that not create the same flux density as a 20-turn coil? I need to think some more...

- Carl

[tohu]

Hehe, nice to see someone following my trail of thought.

My thought was that using 10 two-turn wires in parallel instead of one 20-turn wire would allow much more current (greater flux) and reduce the inductive-effects (phase shifting) and simplify problems in regards to capacitance and that "ringing" effect. :confused:

I haven't really thought about the "receiving" part...

[Sid]

Hi Carl, that´s it. The reduction of number turns is OK, because induction responds with field energy and limits resonant frequency. Low resistance coil is limited by internal resistance of a source and the switch. Nice day, Sid.

[ikarus.nu]

I tink reducing turns you reduce the strength of the magnetic field, like Gauss would say.

To increase the magnetic field in a sized coil, you must increase the current flowing through the coil and the DENSITY of TURNS, I will xplain better:

With a fixed number of turns and diameter the length of the coil affect: more length less magnetic field, if you maintain the number of turns and the diameter, and reduce the lenght of the coil, you will increase the strength of the magnetic field. So will be better to give a second layer increasing 0.7mm the diameter and giving more streng to the magnetic field and more inductance (so you can reduce the original number of turns), than continue winding the first layer increasing the length of the coil plus a 1cm.

I have my own inductance and capacitance meter with a 0.1% of accuracy and 0.1uH 0.1pF resolution, I can give interesting information about coils, so if you need information about a kind of coil, tell me and I'll response as soon as possible.

Ciao!

[ikarus.nu]

ahhh and increasing the inductance you difficult the frequency, with an oscilloscope you can see you need a large inductance to dificult the pass of a 15khz current, so there is no problem, the goal is to get a dense coil with few turns and a big current.

[allan.au]

20 times the current

from my understanding the 20 parallel turn coil would be ok for the transmit pulse as long as u dont mind needing 20 times more tx current. its really like one fat wire carrying 20A. though with the advantages of litz wire... better to use 1 A in a 20 turn series coil .at least for a mobile detector...i guess eric has been into all this but maybe where power consumtion is not an issue then with a low inductance hi current coil the flyback voltage would be lower which could mean earlier front end settling and sampling..and of course u would need a separate rx coil as the voltage induced would be only 1/20th...

[Huizer]

What about split coil windings??

Just a thought.
But what would happen if you needed lets say....30 turn coil of a certain gauge wire. You start with 5 turns of that wire then split into 2 separate wires of smaller gauge. wind this for another 5 turns and then split each of those into 4 wiresof even smaller gauge. wind for 10 turns and then split back to 2 wires for 5 more turns and then to 1 wire for the rest of the turns.
Like this:

[Sean Goddard]

Erm...Litz cable?? I thought this topic had been explored before, but hey, another slant on the idea.

what about Litz cable for a P.I.?? Any ideas anyone?

[Geo]

I do not understand why to use inductor with 20 turns parallel and no 1 turn. We use 20 turns only and only in order to achieve the desirable inductive resistance. So that we make all this combination of turns with likely placement of resistances in serious it is preferable I believe to use the required coils. :confused: :confused:

[Sid]

Theoretically the number of turns makes no difference, practically, I thing, that it is just a problem of switch resistance. When the resistance of FET is one ohm, it makes efficiency fifty percent with a coil of the same resistance. Some days, we will get an ideal switch with zero resistance and the efficiency of a single turn coil will be limited by the cable resistance.