Suggestions for DIY 18" coil for GP Extreme?

[bklein]

I didn't get the stock 18" DD coil with my GP Extreme so I'm interested in making an 18" coil but not sure what to try first. I'd like your suggestions. I see guys using litz wire, wire from USB cables, varnish-coated bare copper, coax center conductor spiral...
DD vs Mono?

I likely will be searching Arizona soils which typically are high iron oxide accompanied by hot rocks and 50 cal. bullet casings...

What confuses me is that with all the talk about interwinding capacitance and teflon vs air vs kynar vs whatever, stranded wire vs solid, there are several guys that use solid varnish-coated copper and do ok...

I will have to depend on a Hays housing - so mono/dd/oval?

Barry

[Mechanic]

Hi Barry,

Solid copper will not work ok with a minelab due to the early sampling times. All eddy currents in the wire must be dead before the first sample is taken.

You have 2 options: tin coated copper wire or litz wire.

I have some old litz here from an old ML coil. It has 35 strands of 0.21mm or .008" DIAMETER not awg, you will have to convert it, I am a mechanic after all!!
I am not sure what the length is. I think you will need 17~22 turns roughly.
You would want about the same specs for the tinned wire. Try a marine supply shop for tinned wire. You should be able to get some teflon coated stuff.

It is important to keep the capacitance as low as possible other wise the coil may not settle properly before the first sample is ready to be taken, which will have a drastic effect on target response and stability.

The above is for a mono coil. I would try and tackle this before attempting a DD. A mono coil will give better depth and it should not be too noisy as far as the ground goes. Larger coils are not as prone to picking up hotrocks. I only use mono coils unless I need to run it cancel due to bad interference. Run the ground tracking in fixed and ground balance often.

Cheers Mick

[bklein]

Thanks Mick,
The main source for marine supplies around here is West Marine:
http://www.westmarine.com/webapp/wcs...classNum=50561
Do you see the wire you are thinking of listed here? It doesn't list insulation material.

With Litz wire, I don't see anyone writing about concern for interwinding capacitance as isn't it coated just with cotton or something? Why is that?

I also have access to a lot of USB2.0/3.0 cables from work. I imagine they have awesome insulation but not sure if the wires are tin or silver coated. I haven't come across wire specs for them yet.

[Mechanic]

Hi Barry,

Not too sure why it is coated with "cotton" it is most likely a teflon or polypropylene thread or something along those lines. It is used as a spacer to keep each turn apart from the next. The more space the better.

Not sure on the usb cables. is it in a continuous length? You don't want to be joining it every meter or so.

You could try sending them an email to ask them what the insulation is made out of, they should have this information on hand.

IBgold or Stefan would be much better suited to helping you out with coil building. I have only ever built experimental coils. It is very difficult to get a coil right.

See here for more info on coils>>> http://australianelectronicgoldprosp...cseen#msg20812

Cheers Mick

[bklein]

Litz is interesting stuff - tends to stretch the limits of my intellectual capability. One of my earliest posts on this site was about using it and Carl had a lot of insight. I get confused when I think of the interwinding capacitance and also interstrand capacitance it should have. As several have said to just use tinned stranded I suspect they are right.
In regards to the USB cable material, yes I have some cables that are probably 12 feet long or so - so splicing would be minimal. No pain to just try it I guess.

[bbsailor]

Litz is interesting stuff - tends to stretch the limits of my intellectual capability. One of my earliest posts on this site was about using it and Carl had a lot of insight. I get confused when I think of the interwinding capacitance and also interstrand capacitance it should have. As several have said to just use tinned stranded I suspect they are right.
In regards to the USB cable material, yes I have some cables that are probably 12 feet long or so - so splicing would be minimal. No pain to just try it I guess.

bklein,

The main benefit of Litz wire is to minimize the effect of what happens when eddy currents are actually generated and detected in the coil wire itself. Solid copper wire would have the worse case. Tin plated is a better choice as the strand-to-strand resistance is higher than the strand-to-strand resistance of silver plated wire, thus lower eddy currents than the silver plated wire but still some at very low delays. When you want to be using low delays with large coils some designers have been using wire with individual strands equal to or less than AWG 32 with a variety of strand amount to account for the desired final coil resistance. Litz wire has insulated individual strands. However, the wire is twisted to allow all the strands to occupy all the spaces in the magnetic field to have an even current flow in each strand. The litz wire would allow the preservation of higher frequency harmonics compared to wire types where some amounts of eddy currents could be found.

Litz wire allows the higher frequency harmonics to keep the pulse turn-off time constant as fast as possible to better stimulate and detect smaller targets. Choosing the right wire gauge, stranding, stranding insulation (Litz), wire metal composition and shielding techniques allows coil designers to have many ways to balance all the variables to build specialized coils for a variety of machines.

Check out the Litz wire at "Surplus Sales of Nebraska". The 40/44 is 40 strands of AWG 44 which equals about AWG 28. This wire is "Served" meaning is has a fabric-like wrap to reduce the turn-to turn wire capacitance within the your PI coil winding. This would make a, 18 to 20 turn (depending on the wire turn-to-turn capacitance), 300uH coil, 11" diameter in the resistance range of about 3.5 ohms. If you want a lower resistance coil look for thicker strands (up to AWG 32) and/or more strands to equal a lower resistance wire. Minelab PI machines seem to need coils with more critical design choices related to resistance, inductance and coil Q. Coil Q is mostly controlled by using the right size (strand gauge and the number of strands) Litz wire.

Bottom line:

If you cut a piece of AWG 18 solid wire, tin-plated stranded sizes, silver-plated stranded sizes and Litz wire sizes, about .25" long (about the size of a grain of rice) and are able to detect them you should be able to detect the soild wire just like being able to detect a gold nugget of the same size. However, if you are using a coil with solid AWG 18 wire or even strands larger than AWG 32 you may not get the delay low enough to detect that grain of rice sized target. Other strand types and sizes represent targets with different time constants and may require progressively faster coils to detect those wire samples with lower time constants. Litz wire allows the fine tuning of some of the more subtle coil characteristics.


bbsailor

[bklein]

Hi bbsailor,
Could you expand on the self-detection issue with solid wire?
With a PI detector can't we get away with some self-detection as we are more interested in and designing for an amplitude envelope integration as we move the coil over the target?
What happens - does a self-detection somehow extend the decay time constant?

[bbsailor]

Hi bbsailor,
Could you expand on the self-detection issue with solid wire?
With a PI detector can't we get away with some self-detection as we are more interested in and designing for an amplitude envelope integration as we move the coil over the target?
What happens - does a self-detection somehow extend the decay time constant?

Bklein,

The short answer is Yes! This even includes the time it takes for the eddy currents to decay in the coil wire itself. This solid wire in the AWG 30 to 32 range should pose little problem at delays above about 10 us. Thicker solid wire can be more and more of a problem the thicker it is.

However, picture it like this. When the TX pulse turns off you want the TX pulse turn-off time constant to be at least 5 times faster than the time constant of your desired target to fully stimulate it. That is why some PI machines detect small gold nuggets, nickels or small gold rings better than others. One of two microseconds can mean alot when trying to detect smaller gold targets or when comparing PI machines.

The TX pulse turn off time contant is governed by the size of the damping resistor calculated as the coil inductance divided by the damping resistor value. The fastest that a coil pould possibly turn off is almost perfectly verticle, but we all know that the slope of the pulse cannot be any faster than the speed of electrons but slope can sure look slower due the invisible things that occur in coils that slow it down.

The invisible thing is total coil capacitance as seen by the coil in the TX circuit and upon which the value of the damping resistor is based. The value of the damping resistor tells you a lot about the coil, its construction and the things that slow the coil down. All the following things contribute coil TX circuit capacitance1) MOSFET COSS, (2) Coax wire type and length, (3) coil winding technique, (4) coil wire type, (5) coil wire size, (6) Wire insulation thickness, (7) wire insulation Dielectric Constant, ( Shielding material, (9) coil to shield spacer. Higher value damping resistors result in steeper TX turn off curves that can be damped earlier to allow a little less delay, thus being more sensitive to a little smaller target.

It does no good to make a fast coil if the RX preamp is still in saturation while the fast coil is ready to sample early. Fast coils need fast RX op amps usually in two low gain (33) stages for a total gain of about 1000. The lower gain op amp in two stages comes out of saturation a few microseconds sooner.

The more 100 pfs of capacitance you remove, reduce, eliminate or avoid in making a coil (from the above list), the less capacitive energy you need to damp resulting in a higher damping resistor value, thus a coil with a faster TX turn-off current. My own crude tests indicate that about 100 pfs of extra coil/TX circuit capacitance results in about a 1us extended delay. Keeping the coil and cable self-resonance between 600Khz and 750 Khz is a good target zone for faster sampling. Remember, higher is better!

bbsailor

PS The smiley faces are the result of my use of the wrong character combinations in my answer. Enjoy.

[bklein]

Hi BBSailor,
I don't know if you saw my other 18" thread yet but I just purchased a used stock 18" extreme coil and find it unimpressive and now wonder why. So reflectng back to your comments above, I am thinking we need an overview on how to test and improve/repair a coil system. i suppose my coil could have excessive capacitance. I am not sure how best to test for this though when the coil is a DD like this. I initially tried driving the tx and rx coils separately using a surf PI clone board. The TX coil gives a typical flyback response but how do I judge it as does the GP really care much about capacitance caused delays of the tx coil? The RX response after the TX pulse is a decaying sinewave, perhaps because I did not terminate it. (?). The RX coil doesn't test well as a TX on the PI due to its high resistance.