Fast coil with teflon wrapping wire

[Carl-NC]

For example, I didn't think that having more resistance in the Tx circuit will do any good, but that is not true. As R directly affects the TC, the higher R is, the shorter TC is.

True, the turn-on time will be faster, but the final current will be more limited. I prefer a lower resistance and using the pulse width to determine the final current, sort of a ramped-current method. Some people disagree with this approach.

I wonder what would happen if we would use a constant current source to charge the coil (a constant current source has internal resistance very high, and based on this, TC = L/R would be very low).

If you suddenly push a constant current through a coil, you will get the same flyback voltage you get when you suddenly turn the current off.

- Carl

[nick_f]

True, the turn-on time will be faster, but the final current will be more limited. I prefer a lower resistance and using the pulse width to determine the final current, sort of a ramped-current method. Some people disagree with this approach.
- Carl

I consider that R reduces the efficiency of the system (reduces battery life), besides limiting the max current. In my oppinion, the normal approach would be to have this resistance as low as possible. If TC will be too high and the desired PPS can't be accomplished, then either add some resistance, or reduce the inductance.

[nick_f]

If you suddenly push a constant current through a coil, you will get the same flyback voltage you get when you suddenly turn the current off.
- Carl

I find this very interesting. What we want to do is to load the coil with a certain amount of energy. And by supplying a constant voltage to it, is not quite the best way, because initially the current is very low and it gradually increasing in time. If we charge a capacitor to a voltage equal to the flyback pulse, and calculate the value of the cap so can it store enough energy, we could discharge the capacitor into the coil, then interrupt the connection between cap and coil after a specific period of time. I think this would make the coil happier, because it will receive the energy more uniformly

[nick_f]

I was thinking about how to reduce the capacitance of the coax cable that links the circuit to the coil. We use a coax because we want to remove the electrical interference from the target signal. Coax cables have larger capacitances than bifilar cable, with wires some distance apart (like 300 ohm TV bifilar cable). If we use bifilar cable and the Rx circuit consists of a opamp with good CMMR, wouldn't that work pretty well?

[nick_f]

If we use bifilar cable and the Rx circuit consists of a opamp with good CMMR, wouldn't that work pretty well?

Sorry, I meant CMRR (common mode rejection ratio).

[porkluvr]

Fast coil

I decided to build a coil using wire-wrap insulated with teflon. The coil has 29 turns on 20cm and an inductance of 461uH. At this stage it doesn't have any shielding. Initially when I connected the new coil I was expecing to see some reduction in the width of the pulse. I was surprised to see that nothing changed!

Regards,
Nicolae

You left out some crucial information. Why not tell us the resistance of your coil and or the gauge? Not all wire wrap wires are the same.

I think you are incorrect in assuming a fast coil will result from a single strand of ww wire. It should be stranded wire to reduce skin effect.
I made the mistake of buying several rolls of solid 26AWG, thinking it was peaches because it was silver plated Teflon. Wow.

No, I did not take into account the high rise and fall times we are trying to achieve and the resulting skin effect. Stranded 18awg would be excellent, 20AWG might be good enough. I think bbsailor suggests using tin plated wire -not- silver plated, to reduce interaction between the strands.

Also, I think maybe the 29 windings and 461uH may be a little on the high side so you should look at that. I could be wrong. I am planning my detector around a 300uH coil, but I am not an expert and have almost no practical experience. It's hard to take all pieces of the puzzle in at one time.

[nick_f]

Hi Porkluvr,
I will add the information about resistance and gauge tonight.
I didn't assume that a single strand will make a fast coil. It was rather a relative comparison with my other coil that I built with magnet wire. Anyway, it is very good you pointed these matters, to avoid any confusion.
Have you compared the results of the coil made with solid silver plated wire, as opposed with the results of multistranded tin wire? I am curious if you noticed any difference in the performance. I am aware about the skin effect, which has more effect at higher frequencies. Multistranded wire has a total surface area larger than solid wire and that's why is better.
I already have been advised to reduce the inductance to 300-350uH, in order to have a fast coil and I built the teflon ww coil with 368uH (461uH is the magnet wire coil).

Regards,
Nicolae

[porkluvr]

Hi nick f.

I am still working on a prototype after revising the design. Last night I tried to thrash out a switching regulator copper layout, only later to decide that I should have the battery & switcher as a separate entity. Leaving out the power supply should make the PCB layout lots easier because I have a multi-tx-channel system to worry about.

I'm sorry about spouting information that you already know. It's been a rough week for me - lost sleep, and having to take pain medication (which works sporadically).

I seldom look in this section of forums and now realize there's a lot of new stuff since my last visit.

bbsailor has an excellent "Making a Fast Pulse Induction Mono Coil" article that is very good. He and Carl are the experts here.

At the current stage of development, my detector will use a DD coil. One side with 26AWG(?), solid, the other side 18ga. stranded. The solid winding will have a relatively long primary delay on a single long pulse. The stranded wire channel will transmit 3~8 short pulses, having short delays.

I've been doing this for years but have yet to complete numero uno because I do too much research and too little development. My lab consists mainly of LTspice. It's not much, but I would be nowhere without it. I'll need to acquire a scope pretty soon.

If you feed me some coil data I could make side-by-side comparisons, but there are many inaccuracies even with the best data. (How would I simulate skin effect, huh?) Maybe I should look into that.

I can only guess, but I think there could be a vast improvement in using stranded wire compared to solid -but- your FET drive and cutoff also needs be fast, and, that's not all... . (Owwww. I'm outa here.)

[nick_f]

Hi Porkluvr,
I measured the diameter of my teflon wire. It is very thin, about 0.25mm. Which means is AWG30 (0.255mm). With the teflon insulation, it has a diameter of 0.6mm.
It has a nice shiny aspect, I am pretty sure is silver plated. The coil resistance is 6.1 ohm. As I said, its inductance is 0.467uH, and measured with a capacimeter, the capacitance coil-shield is 248pF. Coax cable capacitance for the cable I used is 150pF (1.5m of 50 ohm cable). The MOS-FET I use is IRF740. The coil dumping resistance is 824 ohm and at the moment, I get a pulse of 10us (I have no idea why the pulse width has increased from 8us to 10us. I tried adjusting R3 to make the pulse lower, but it didn't help.
I was also thinking of a modular approach, where to build each basic module on its separate board, then connect them on a tiny mainboard via connectors. In my case, each module would have at most 5 or 6 connections to the mainboard. So, if I decide to try a different preamplifier, I would just build that on a separate piece of PCB and plug it in. I am not yet at that stage, but I want to do it. I know bbsailor and Carl are the experts, I already got a lot of great feedback from them. Tinkerer is also very helpful, he provided me with many ideas already. I decided to build a metal detector just two or three months ago and the HH is the first I've done. I already made another PCB for Tesoro Golden Sabre, the PCB is posted on this site and some people built it and it works. I have to buy some parts and populate the board, but I prefer to finish this one before I start the second one.
I also use LTSpice, but I am not good with creating subcircuits or importing other components not listed in its default library.
I think you could come up with a model for skin effect if you have already some real data and see how close the simulations matches the real world. I don't know if skin effect is significant enough in the metal detector world. Of course, anybody should use multistranded wire rather than solid single core. I am more curious if silver plating indeed has a negative impact, and how significant it is.
Some comparative test would be excellent, if somebody who has both silver plated and tin plated teflon wire with the same gauge of wire.

Regards,
Nicolae

[porkluvr]

The "Magnet Wire Dimension Chart" from Allem Inc is a very useful tool. Starting at 6/0 ending at 60AWG, lists SWG, BWG, MM, in incremental diameter. It's great but it does not include resistance per unit length, so, I made an addendum, showing the basic resistivity formula. With that (and a simple calculator) we can do the math ourselves.
http://www.allem.com/wireengr.html

Your 6Ω 30AWG coil has got to go. You should use 26AWG or better. For a really good PI coil stranded 18 would be better. Tin plating is preferred for stranded wire because you do not want low resistance between the individual strands. Think about why we'd prefer Litz wire.

You are attempting to create a flyback pulse waveform of ZERO duration. I think that zero duration equates to infinite frequency. It's unobtainable, but you get closer to that goal as you minimize coil resistance and coil capacitance.

You should also try to find some lower capacitance coax. Try RG6.
I may be able to help you with LTspice in the future but I don't feel up to it today. I will entertain specific requests, just don't get impatient.

Oh, I edited the equations shown here for better readability. If anybody sees a problem, please let me know.

[Tinkerer]

Inter-wire capacitance for PI coils

Hi Porkluvr,
I measured the diameter of my teflon wire. It is very thin, about 0.25mm. Which means is AWG30 (0.255mm). With the teflon insulation, it has a diameter of 0.6mm.
It has a nice shiny aspect, I am pretty sure is silver plated. The coil resistance is 6.1 ohm. As I said, its inductance is 0.467uH, and measured with a capacimeter, the capacitance coil-shield is 248pF. Coax cable capacitance for the cable I used is 150pF (1.5m of 50 ohm cable). The MOS-FET I use is IRF740. The coil dumping resistance is 824 ohm and at the moment, I get a pulse of 10us (I have no idea why the pulse width has increased from 8us to 10us. I tried adjusting R3 to make the pulse lower, but it didn't help.
I was also thinking of a modular approach, where to build each basic module on its separate board, then connect them on a tiny mainboard via connectors. In my case, each module would have at most 5 or 6 connections to the mainboard. So, if I decide to try a different preamplifier, I would just build that on a separate piece of PCB and plug it in. I am not yet at that stage, but I want to do it. I know bbsailor and Carl are the experts, I already got a lot of great feedback from them. Tinkerer is also very helpful, he provided me with many ideas already. I decided to build a metal detector just two or three months ago and the HH is the first I've done. I already made another PCB for Tesoro Golden Sabre, the PCB is posted on this site and some people built it and it works. I have to buy some parts and populate the board, but I prefer to finish this one before I start the second one.
I also use LTSpice, but I am not good with creating subcircuits or importing other components not listed in its default library.
I think you could come up with a model for skin effect if you have already some real data and see how close the simulations matches the real world. I don't know if skin effect is significant enough in the metal detector world. Of course, anybody should use multistranded wire rather than solid single core. I am more curious if silver plating indeed has a negative impact, and how significant it is.
Some comparative test would be excellent, if somebody who has both silver plated and tin plated teflon wire with the same gauge of wire.

Regards,
Nicolae

The wire to wire capacitance for your Teflon coated wire is about 35pf/meter. Now, your windings are not perfectly parallel so you have also some air between the windings. This reduces the capacitance because air is a better dielectric.
If you were to suspend the wires in the air, so as to have air instead of Teflon as insulation, the capacitance would only be 16pf/meter.

Another dielectric that is nearly as good as Teflon, is PP or Polypropylene. Now, if you were to take air and PP and make a foam with it that is 50% air, the dielectric would be somewhere between air and PP and even better than Teflon. They used to make Coax cable with such kind of foam.

So how can you use this information for coil making?
I have some 0.95mm diameter magnet wire and want to wind a new coil.
To get 35pf/m inter-winding capacitance, I would need to have about 2.5mm TFE (Teflon) insulation. That must be difficult to find.
So my solution is going to be PP+air, about 2mm, I think I will manage to get the capacitance down to about 25pf/m.

The coil diameter will be 280mm
Coil inductance about 300uH
Coil DC resistance about 0.4 Ohm

Then I hope to be able to sample at 3uS, with a coil current of about 1.2A.

Lets see if I can accomplish that.
Then I would like to have somebody to come up with signal amplitude measurements for 1" square alu foil to compare with.

All the best

Tinkerer

[nick_f]

Hi Tinkerer,

That sounds good, I think I will copy your coil specs (same diameter, inductance and resistance - as close as possible). I will use multistranded wire insulated with PP (not sure about what material is actually). I found some PP spiral wrap, hope it will fit the coil. For the shielding, I will try to use shield from video cable of a monitor, I am very curios of the results.
What mos-fet transistor are you using?

Regards,
Nicolae

[Tinkerer]

Hi Tinkerer,

That sounds good, I think I will copy your coil specs (same diameter, inductance and resistance - as close as possible). I will use multistranded wire insulated with PP (not sure about what material is actually). I found some PP spiral wrap, hope it will fit the coil. For the shielding, I will try to use shield from video cable of a monitor, I am very curios of the results.
What mos-fet transistor are you using?

Regards,
Nicolae

Hi Nicolae,

I did not mean for you to copy what I am doing. Mine is an experiment, I don't know the outcome myself.
Insulated multi strand wire is good for you to try. You could also try to use about 5 strands of your magnet wire in parallel. It would be like Litz wire. If you feel like trying that, I suggest you use 5 strands of magnet wire and 3 strands of polyester twine and twist them lightly together. You will need about 20 to 21 meters of each strand.

One more thing about insulated multi strand wire. PVC insulation is not too bad as dielectric, but it's dielectric quality deteriorates quite a bit when hot. This might be a problem when working with the detector coil in the hot sun. Also make sure your coil is of light color so it does not get too hot.

I use the IRF731. It is far from ideal, but it is what I have. For yourself, use a low COSS and low resistance Mosfet with the voltage you need for the flyback not to discharge through the reverse diode. This slows down the decay and gets the Mosfet hot.

Tinkerer

[porkluvr]

The coil diameter will be 280mm
Coil inductance about 300uH
Coil DC resistance about 0.4 Ohm

Tinkerer

I know you didn't intend to leave out the number of turns. Curious: which coil calculator do you use? Spiral wound or toroid coil?

Although it is incomplete that looks like the start of a very good coil. 0.95mm is aproximately equivalent to 18AWG which is part of my plan, but anything close whould be good enough for me. I was stubbornly hanging on to a 330uH number but I've decided to knock it back.

I tried to look up data on my "Magnet Wire Dimension Chart" but it did not list 0.95mm. However it is listed in "Magwire Dimensions Chart", also available at the hyperlink I posted earlier. I calculate that 0.95mm is about 7.4ohms/1000' compared to 18AWG at 6.4ohms/1000'. I hope I did not make an embarassing mistake in my calculations and, yes I rounded those numbers. Considering the relatively small length of wire needed for a 300uH coil - the variance should not be a big deal.

The bigger part of the design error seems to me to be the coil calculator output. Even the "good" ones seem to come without good documentation or be totally without documentation.

[Tinkerer]

I know you didn't intend to leave out the number of turns. Curious: which coil calculator do you use? Spiral wound or toroid coil?

Although it is incomplete that looks like the start of a very good coil. 0.95mm is aproximately equivalent to 18AWG which is part of my plan, but anything close whould be good enough for me. I was stubbornly hanging on to a 330uH number but I've decided to knock it back.

I tried to look up data on my "Magnet Wire Dimension Chart" but it did not list 0.95mm. However it is listed in "Magwire Dimensions Chart", also available at the hyperlink I posted earlier. I calculate that 0.95mm is about 7.4ohms/1000' compared to 18AWG at 6.4ohms/1000'. I hope I did not make an embarassing mistake in my calculations and, yes I rounded those numbers. Considering the relatively small length of wire needed for a 300uH coil - the variance should not be a big deal.

The bigger part of the design error seems to me to be the coil calculator output. Even the "good" ones seem to come without good documentation or be totally without documentation.

Hi porkluvr,

Ok, I wound the coil. It did not come out exactly as I wanted, bud close enough to give an idea if it works.
The turns are 23. I used 21 meters of wire.
I use Misc.el for the calculations. There is always some guess work about the fill factor.
So what I wanted is a fill factor (copper/insulation) of about 16%, but the coil came out with about 19%. Like usually I used what I had and the string I had was a bit thin.
Now I will measure the inductance and resistance approximately, with the tools on hand.

Tinkerer