Split coil drive (PI mono coil)

[Teleno]

A mono coil can be spit in two equal parts, driven separately and it's faster than a single coil.
This is a conjecture based on my understanding and tested in LT-Spice, not yet on the bench.

One 380uH coil is equivalent to two 100uH coils connected in series (coupling coefficient of 0.9. Used this calculator).
If we drive the two 100uH separately with the same total current as the 380uH coil and add the received signals, we can sample earlier without loss of gain.

In the simulation the damping resistors have been adjusted manually. The upper half of the circuit is the split driver followed by two separate preamplifiers whose outputs are passed to an adder/level-shifter to refer the output to ground. The lower half is the mono coil. I assumed the coupling between the target and each split coil is the same as the coupling between the target and the mono coil.

The improvement is a bit over 2us.

LT-Spice file attached below.

[Carl-NC]

I would expect the coupling to be better than 0.9. What happens if you increase the coupling to 0.95? 0.99?

[Teleno]

I would expect the coupling to be better than 0.9. What happens if you increase the coupling to 0.95? 0.99?

Pretty much the same thing. You can try it in the attached LT-Spice file. Remember to adjust L4 to the new coefficient. It should be 400uH for a coefficient of 1.

An added bonus is that the peak transient voltage is less than half.

[Carl-NC]

I've never that particular technique, but did something similar:


This uses the center-tapped coil in a bipolar pulsing scheme. Because each half-coil is 1/4 the total inductance you get a higher current and faster flyback while getting the entire L on the RX side.

[moodz]

What's old is new again ... https://www.geotech1.com/forums/foru...091#post104091

:-)

[Carl-NC]

I knew we had talked about this before. Hard to remember all the discussions.

[Teleno]

I've never that particular technique, but did something similar:


This uses the center-tapped coil in a bipolar pulsing scheme. Because each half-coil is 1/4 the total inductance you get a higher current and faster flyback while getting the entire L on the RX side.

Thank you and moodz for the old posts. They confirn thatl my conjecture is true.

The center-tapped technique works the same but it's limited to N = 2 splits. My variation allows splitting the coil further. For example, for N = 3 in the circuit attached L = 44uH and the timing gain would be above 3us.

[moodz]

Thank you and moodz for the old posts. They confirn thatl my conjecture is true.

The center-tapped technique works the same but it's limited to N = 2 splits. My variation allows splitting the coil further. For example, for N = 3 in the circuit attached L = 44uH and the timing gain would be above 3us.

I did not get to the bottom of this research ... there appear to be some benefits ... you should pursue them.

moodz

[Teleno]

I'm thinking of placing the split coils in an IB configuration, coupling as low as feasable. The simulation allows even earlier sampling but I guess the resulting field be greatly reduced, also the coupling with the target. Could be useful for small gold.

[lucifer]

Hi Teleno,
To me your split coils look connected in parallel rather than in series. If two coils are connected in series the current flowing through them both is always the same, even if the coils have different resistances, inductances, etc. In your case, if you change the second coil to have slightly different resistance you will see that the currents in both coils differ. When constructed in real, you may get slightly different coils hence slightly different decay curves. When they are amplified and summed you may get noise issues.

[Teleno]

Hi Teleno,
To me your split coils look connected in parallel rather than in series. If two coils are connected in series the current flowing through them both is always the same, even if the coils have different resistances, inductances, etc. In your case, if you change the second coil to have slightly different resistance you will see that the currents in both coils differ. When constructed in real, you may get slightly different coils hence slightly different decay curves. When they are amplified and summed you may get noise issues.

A point to consider though I expect no big issues with that.

Another possible advantage is a much reduced coil capacitance. Take for example a spiral coil wound with two parallel conductors (one for each split coil). The distance between two turns of the same coil is double. Also, since two adjacent turns of different coils are always at the same potential the capacitance between them has no effect.

[lucifer]

Generally I like spiral coils, they have less capacitance compared to regular coils. Not sure about this one though.. You still have different potentials between adjacent turns.

[Teleno]

Not sure about this one though.. You still have different potentials between adjacent turns.

Look carefuly at the drawing. The length of wire to reach two adjacent points is almost the same in the split spiral. In a mono spiral there's a difference of one turn between adjacent points.

In the drawing the gap is exaggerated, in practice the wires are touching side by side, the difference in length is minimal.

[lucifer]

Look carefuly at the drawing. The length of wire to reach two adjacent points is almost the same in the split spiral. In a mono spiral there's a difference of one turn between adjacent points.

In the drawing the gap is exaggerated, in practice the wires are touching side by side, the difference in length is minimal.

Yes, that's true when you're close to the center. As you go several turns to the outer side the difference gets bigger. If you have a wire 1mm in diameter (including insulation) and the coil radius gets to 10 cm, the difference between two adjacent turns is 6-7mm even if the turns are touching side by side.

[Teleno]

Yes, that's true when you're close to the center. As you go several turns to the outer side the difference gets bigger. If you have a wire 1mm in diameter (including insulation) and the coil radius gets to 10 cm, the difference between two adjacent turns is 6-7mm even if the turns are touching side by side.

That's correct. The maximum difference is half a turn at the outer edge, in the mono coil it's a full turn along all the wire length. Not ideal but still an improvement relative to the mono coil. The peak voltage being lower also helps reduce the difference.