Coil Wire Size for Fast PI Coil

[Bob_T]

In BBSailor uses 30 AWG wire in his example in his Fast PI Coil Building paper. It seems everyone else is using from 24 AWG (Smitty II) through 26 AWG, but I haven't seen many other references to 30 AWG. I know that the wire size has to be matched to the electronic circuit and in the end everything is a compromise on one level or another.

It seems that 30 AWG will have higher resistance and will limit the pulse current, but have a shorter decay period. 24 AWG on the other hand will conduct more current and have a slower decay, so then why does Smitty II recommend it for his fast PI design? I keep changing my mind on the wire size and I'm in danger of buying so many spools of wire that I could just go out and buy a new $4500 MineLab PI detector for the same money (ha).

As soon as I decide on what LCR meter to buy, I can wind a bunch of different coils and test them and learn something, but in the meantime, can someone please explain the issues that determine the optimum wire size for a fast PI coil. I know this has been covered before but the information I find is often conflicting.

Thank you!

Bob

[bbsailor]

In BBSailor uses 30 AWG wire in his example in his Fast PI Coil Building paper. It seems everyone else is using from 24 AWG (Smitty II) through 26 AWG, but I haven't seen many other references to 30 AWG. I know that the wire size has to be matched to the electronic circuit and in the end everything is a compromise on one level or another.

It seems that 30 AWG will have higher resistance and will limit the pulse current, but have a shorter decay period. 24 AWG on the other hand will conduct more current and have a slower decay, so then why does Smitty II recommend it for his fast PI design? I keep changing my mind on the wire size and I'm in danger of buying so many spools of wire that I could just go out and buy a new $4500 MineLab PI detector for the same money (ha).

As soon as I decide on what LCR meter to buy, I can wind a bunch of different coils and test them and learn something, but in the meantime, can someone please explain the issues that determine the optimum wire size for a fast PI coil. I know this has been covered before but the information I find is often conflicting.

Thank you!

Bob

Bob,

The purpose of using AWG 30 is because I was using the Hammerhead PI in its high frequency region with over a 100 us TX pulse width. Remember, the coil charge TC is based on the coil inductance divided by the total TX circuit resistance including the coil wire, MOSFET on-resistance, and lead resistance. By keeping the TX pulse width near about 130 us I can get near 3 TCs on the coil charge TC or about 95 percent of max current. Also, remember the flyback pulse the needs to be damped and is based on the coil current so faster TX pulses, lower capacitance in the coil circuit, lower COSS MOSFET rating, better shielding and shorter/lower coax capacitance all contribute to finding the critical damping resistor value. In this case a higher value of damping means a lower overall capacitance and flyback to damp. All of this is a trade off and my article was my attempt to address this. When you are integrating a few thousand pulses you somewhat make up for less coil current by the increased sampling. Eric Foster's high frequency PI designs seemes to allow faster sampling because of less TX current build up. Most lower frequency PI machines in the 100 to 400 PPS range are higher power machines that typically use about 200 to 300 us TX pulse widths but also use lower value damping resistors to damp higher flyback voltages and make lower delays more difficult. It all depends what your primary targets are! Using about AWG 26 stranded will make a 3 ohm total resistance coil circuit (about 2.5 ohms for the coil .5 ohm for the MOSFET) will produce a 100us TC so a 300us pulse width will allow the current to rise 3TCs to about 95 percent of max current.

My AWG 30 (.024" OD) was a single strand Terlon insulated wire that minimized the amount of eddy currents being generated in the wire itself. The rule of thumb is to use AWG 32 or thinner strands in stranded wire but use tin plated stranded Teflon to help minimize the eddy current build up or you will develop eddy currents and extend the minimum delay you can set your machine at. Repeat! ...Silver coated stranded wire will make the eddy current problem worse at lower delays.

Remember, optimize your TX pulse current and pulse width based on the nature of the targets time constants that you seek. Nickels and mens gold wedding rings are good targets to optimize on for beach hunting.

bbsailor

[bklein]

My intent is to make a mono 18" for a GP Extreme. I'm having a hard time finding the recommended 36 strands of 32 AWG Litz. IBGold says that 660/46 has been used in smaller coils. Should I go to a wire manufacturer and pay the big bucks for custom 36/32 or use something else that gives the same ohms/ft.? Does the frequency range specified for Litz wire have any importance when used in a PI design? Is Litz always the wire of choice for Minelab GP's or will single strand (or stranded) be ok in some coils?

[Bob_T]

bbsailor,

Thanks for taking the time to post a detailed reply to my questions. It'll take me awhile to assimilate it all and if it doesn't answer all my questions, it will at least allow me to ask better questions...

Bob

[Carl-NC]

My intent is to make a mono 18" for a GP Extreme. I'm having a hard time finding the recommended 36 strands of 32 AWG Litz. IBGold says that 660/46 has been used in smaller coils. Should I go to a wire manufacturer and pay the big bucks for custom 36/32 or use something else that gives the same ohms/ft.? Does the frequency range specified for Litz wire have any importance when used in a PI design? Is Litz always the wire of choice for Minelab GP's or will single strand (or stranded) be ok in some coils?

I think you'd be hard-pressed to see a difference between Litz and stranded in this case. You do want low R, so I'd use 18awg or such.

[bklein]

Since I bought an 8" mono Minelab coil already, and it works great, I'll give it a try using stranded instead of Litz on this one. Say I use the technique of adding turns of wire until optimized. What determines optimization? 300uH? .4 ohms? Best air distance? (could I really tell?)

[bbsailor]

bbsailor,

Thanks for taking the time to post a detailed reply to my questions. It'll take me awhile to assimilate it all and if it doesn't answer all my questions, it will at least allow me to ask better questions...

Bob

Bob,

Let me ask you a few questions so I can give you a better answer.

1. What PI design are you making?
2. What is the Pulse Per Second (PPS) TX pulse range?
3. What is minimum delay that your PI design can be set at?
4. What MOSFET part number are you using to drive the coil?
5. What coil resistance and inductance is recommended for your PI machine?
6. What are your primary targets?
7. Do you have an oscilloscope?

With answers to the above, I can give you some better recommendations.

bbsailor

[Carl-NC]

Since I bought an 8" mono Minelab coil already, and it works great, I'll give it a try using stranded instead of Litz on this one. Say I use the technique of adding turns of wire until optimized. What determines optimization? 300uH? .4 ohms? Best air distance? (could I really tell?)

You want ~300uH and (I think) <= 0.6 ohms.

[Bob_T]

Bob,

Let me ask you a few questions so I can give you a better answer.

1. What PI design are you making?
It's an attempt at building the HH variation designed by Smitty II over on the Findmall forum. Here is his schematic.

2. What is the Pulse Per Second (PPS) TX pulse range?
From about 900 to 1100
3. What is minimum delay that your PI design can be set at?
10uS
4. What MOSFET part number are you using to drive the coil?
NE555
5. What coil resistance and inductance is recommended for your PI machine?
Smitty II recommends a 300uH inductance and I could not find any reference to coil resistance in his threads. It was his suggested use of 24 AWG stranded wire for the coil that was partly behind my question about coil wire size. I sure don't want to be construed as second guessing him or anyone else. It's just something I'm trying to understand so I can make my own decision, and hopefully an informed decision, on coil wire size.
6. What are your primary targets?
The intent is to find gold nuggets down to 3 to 5 grains at a depth of one to 3 inches. The smallest size would be about the size of a 12 font "o" at an inch or so. Of course larger and deeper nuggets are acceptable!
7. Do you have an oscilloscope?
Yes. I also have a function generator but don't as yet have an inductance meter. I plan on measuring coil inductance with this method:

http://technologyinterface.nmsu.edu/...ct/induct.html

With answers to the above, I can give you some better recommendations.

Thanks!

[maikl]

4. What MOSFET part number are you using to drive the coil?
NE555
Thanks!

MOSFET is not NE555, but the IRF740...

[Bob_T]

MOSFET is not NE555, but the IRF740...

Thank you. I wasn't sure and that's why I attached the schematic.

Regards,

Bob

[mikebg]

FAST CIRCUIT

People, the fast coil is not enough for small sample delay.
You need FAST NETWORK.
To avoid cable capacitance, you should place the end stage of TX (IRF740 and energy capacitors) near to search head or inside it:

http://australianelectronicgoldprosp...2462/#msg12462

[Bob_T]

People, the fast coil is not enough for small sample delay.
You need FAST NETWORK.
To avoid cable capacitance, you should place the end stage of TX (IRF740 and energy capacitors) near to search head or inside it:

http://australianelectronicgoldprosp...2462/#msg12462

Yes, Mike, Thank You and it's understood that everything works together as a system and a fast coil will not reach it's capability without the equivalent fast network and vice versa: a fast network has no value with a slow coil.

There are hundred of threads on network design. This particular thread is intended to be about the parameters involved in selecting the optimum wire diameter and other parameters for a fast coil. For the purpose of the discussion here can we just assume that we will have the corresponding fast network?

Bob

[bbsailor]

People, the fast coil is not enough for small sample delay.
You need FAST NETWORK.
To avoid cable capacitance, you should place the end stage of TX (IRF740 and energy capacitors) near to search head or inside it:

http://australianelectronicgoldprosp...2462/#msg12462

Mikebg,

You are correct and better way would have been to broaden the concept to making a fast network, however I did mention those other factors in the article including:
1. Minimize MOSFET COSS
2. Minimixe coax cable length
3. Minimize coax capacitance per foot rating
4. Minimizing capacitance between coil turns
5. Minimizing capacitance between the coil and the shield
6. Minimize the capacitance and detectability of the shield material

Collectively, they all contribute to making a fast network of which the coil is just a part. The active clevis-located active circuit is an interesting concept. I also believe that if we can put the first preamp circuit right in the coil (as is done in the Minelab Sovereign coils) to do an (1) impedance connversion and (2) low noise close to source amplification that a few microseconds can be taken off the delay. That all depends on making the circuit small enough and located in the right place in or near the coil to not be subject to the interference of the TX pulse. Maybe with better coil modeling, location of reliable null spots in milti-coil designs and the miniaturization of preamp chip circuits that, this might be possible and even a leap forward. However, once you get the delay low enough, you start to pickup the eddy currents in the coil wire itself and more importantly, the solder blobs that make the joint between the coil and the coax. Maybe, we need a new breed of coil-shaft combinations where the coil wire is suspended inside the shaft/clevis to eliminate the solder blob connetion in the coil housing or at least partly up so as to be beyond the detection range. The preamp can be placed at the top of the clevis along with the connector that secures the clevis to the shaft about 1 foot above the coil?

These types of discussions are what this forum is all about. The next question is what is fast enough to be practical? On wet beaches, coils start to react to the wet sand at and below 10 microseconds.

bbsailor

[bbsailor]

Yes, Mike, Thank You and it's understood that everything works together as a system and a fast coil will not reach it's capability without the equivalent fast network and vice versa: a fast network has no value with a slow coil.

There are hundred of threads on network design. This particular thread is intended to be about the parameters involved in selecting the optimum wire diameter and other parameters for a fast coil. For the purpose of the discussion here can we just assume that we will have the corresponding fast network?

Bob

Bob,

Since your primary targets are small gold nuggets, I would recommend a 6" diameter coil with 28 turns of stranded AWG24 which works out to 44 ft at .02567 ohms per foot or 1.13 ohms plus about .55 ohms for the MOSFET on resistance and about .1 ohms for the coax/connector resistance for a total of 1.78 ohms with a coil inductane of 297 micro-henries. You may even want to go down to stranded AWG26 for a little higher resistance and slightly higher inductance of about 310 microhenries if you are using a very short TX pulse width.

bbsailor

PS Use this web site to calculate your own coils. http://wetnet.net/rf_design/airind.main.cgi For cable length use the diameter of the coil wire bundle typically 5 X the wire OD for up to 19 turns and 6X the wire OD for up to 31 turns.