shielding for a fast coil

[Tinkerer]

Looking for different ways of shielding a PI coil. Would a carbon composition coating work? Or a Zinc-epoxy coating? How resistive or conductive must the shielding be? Which way to obtain minimal capacitive loading? Should the shield be very close to the windings or at a distance? How far?
Help would be much apreciated.
Tinkerer

[Reg]

Hi Tinkerer,

I see you jumped over here. That is probably a good idea since this site is more geared to building a detector.

Now, to answer your questions, first you can use several different types of shielding including the graphite or carbon type paint. One that is available is a conductive picture tube coating. Using a conductive graphite or carbon paint is a common technique used on VLF coils for shielding. The general concensus is to use just enough coating to provide the proper faraday shielding. This could be as high as 10K per sq inch resistance. Get the shielding on too thick and it can be detected on a VLF, but might not be as critical on a PI.

The objective of the shielding on a PI coil is twofold. First, it is to provide the faraday shielding and the second is to shield against noise, a major concern on a PI, depending upon where it is used.

Eric Foster uses a a very thin lead foil shielding to reduce the noise.

To reduce the capacitance between the shielding and the windings, there should be some distance between the shielding and the windings. Again, Eric uses a layer of spiral wrap (spirawrap) which spaces the shielding maybe .050 inch or so. some people have taken something as simple as rubber tubing and spit it to use as a spacer.

Does the material used as a spacer make a difference. Probably, but would be hard to measure. I have used plastic tubing, spirawrap, rubber tubing, and simple tape. My best results have been when I used a double layer of teflon spiralwrap, but other spiralwraps seemed to work about as well.

I presently use a 3M shielding tape that is about impossible to get right now. What I use is a polyester tape that is plated with a copper coating, almost like a paint. I will be switching to a slightly different shielding tape shortly. 3M has come out with a replacement that is still a polyester tape that is plated, but the plating is a copper/nickel plating. Unfortunately, this general type I use is somewhat expensive and hard to find. However, this tape is used to provide shielding against noise as well as a decent Faraday shield.

Several types of materials can be used as shielding and will work. I have used a conductive tape made for splicing high voltage splices. I have also used something as simple as a layer of conductive ribbon I got at Walmart. Some simple aluminized colored ribbon found in the materials section is conductive with a layer of conductive material. I just taped the bare leads of my shielding wire to the conductive layer and all worked fine.

Now, to your question about a fast coil. I would recommend you use a stranded wire, preferebly with Teflon insulation. wrap the wire tightly with spiral wrap, maybe two layers to assure a space between the shielding and the wiring. If you use the special shielding tape I use, then you should be able to build a coil that works at your goal of 10 usec. It will be very tough but it can be done.

As I said on the other forum, there are other factors such as the preamp design, the FET used and a few other things that will influence the final result. Right now, I have coils that will work in the 6 usec to 7 usec range using this technique on one of my PI's and less than 10 usec on my HH, so I know it works ok.

As I mentioned on the other forum, it is much easier to get a DD coil to work at the shorter delays. So, it is something to think about.

Finally, did you ever find a copy of the schematic of your PI? If so, can you post it, or even send me a copy via email so I can see what might be needed to speed it up.

Reg

[bbsailor]

shielding for fast coils

Tinkerer and Reg

A while back Eric Foster stimulated me, with his discussions on his forum, to do some critical thinking about coils design and response speed.

I wound many coils and found some interesting things that may not be obvious at slower speeds but are very important when trying to get below 10uS. Please let me know if you disagree with any of my conclusions.

1. Use wire with insulation and shielding spacers with a good dialectric constant. The shield adds about 100pf to 300 pf to the coil's self resonant performance. Measure the seld-resonance first without the shield and then with the shield and then calculate the capacitance of the shield. This simple experiment will sensitize you to the effect of capacitance on coil performance.

2. Use coil driver MOSFETS with low output capacitance. My crude calculations revealed that for each 100 pf you remove from the TX circuit, you can reduce the delay by about 1 uS. The first few hundered pfs of a slow coil are easy to remove by careful dsesign but it gets harder as the coil gets faster. This means that this 100pf per 1 us may not be linear but it is a good mental model to get you moving in the right direction.

3. DD coils are faster because the TX circuit is isolated from the RX circuit along with the associated loading by each circuit. A mono coil TX circuit is loaded by the RX circuit. The RX circuit of a mono coil is loaded by the TX MOSFET capacitance. Separate TX and RX also allows the RX circuit to have more turns for more sensitivity to small targets while maintaining a fast speed. DD coils also help eliminate noise and some ground effects in certain areas. Each coil and be optimized with a damping resistor separately.

4. The self-resonance frequency and the value of the damping resistor are very good predictors of a coil's potential performance. Induce a square wave pulse into the coil under test (CUT) with the pulse being fed into another coil. Examine the CUT using the 10X probe setting to minimize capacitance loading of the probe. Switch the probe (or change probes) between 10X and 1X to see the effect of a few pfs on the self resonant frequency. Obesrve the tapered ringing pulse train on the scope and calculate the self-resonant frequency. Increase frequency and amplitude on the coil attached to the pulse generator until you see a good pulse train from the CUT on the scope. If self resonance is near 1 MHz, you are reaching the "sweet spot" for a fast coil. The value of the damping resistor is very telling about the capacitance of the coil. The higher the capacitance of the coil and the associated parts of the TX circuit (for a DD coil) and TX and RX circuits (for a mono coil), the lower value resistance is needed to damp the stored energy in that capacitance. Therefore, the goal should be to wind a coil with low capacitance and use low capacitance components in the TX circuit so your damping resistance is about 1K ohm (higher is better).

When the coils reach their speed limit, it is time to look at circuit components and the metal detector design to get a faster response. Low power (thin wire or series resistance), fast PPS coils are just as effective as high power (thick wire), lower PPS coil speeds. This means that it take less time for lower power coils to recover before sampling can occur. When you are at or below 10 uS you are in a new territory and must be looking for low conductiveity targets like gold. Here, the coils must match the PI design for optimum performance.

I hope my own coil experience can be useful to you.

bbsailor

[Tinkerer]

coil shielding

Hi Reg and bbsailor,
thanks for the info. I plan to make a fiberglass mold and coil form. Since I found a can of graphite powder and a can of zinc powder in my old stuff, I will try these materials for the shielding.
first lay up a coat of shielding mix, (do I need to leave an air gap on the coil?) then incorporate a thin stranded wire for later connection to the ground, then lay up a layer of fiberglass for strenght.
Your info on spacing will come in handy.
Later cast the coil into the mold, using epoxy resin filled with microballoons to reduce the weight.
bbsailor, your info on capacitance is just what I was looking for, thanks.
Tinkerer

[bbsailor]

coil shielding

Tinkerer

Go to the link below and look at premade coil forms.
http://www.hayselectronics.com/parts.htm

You can make your coil and shield to fit the coil housing then use the microballoons to lock it in place inside the housing.

Make a wood coil form that will match the diameter of the coil that will fit your coil housing. Cut a top and botton about 1/2 inch larger in diamenter that can be removed once the coil is wound. Place slits about .5" long spaced each inch around the circumference pointing into the center. Once the sides are removed use these slits to tie dental floos around the coil to secure it prior to removing it from the coil form. Then use the spiral wrap or what ever spacer you choose. Add the shield. Leave a narrow gap. Connect the shield to the coax cable braid. Attach two coil leads to the coax. Mount in the coil housing. Add filler. Coax wire should have about 20pf per foot of capacitance. Here less is better.

Good luck. After you make your coil, let us know its critical parameners and how it works.

Here are the critical parameters to take notes about.

Coil diameter
Coil wire AWG size
Insulation thickness
Number of turns
Calculated inductance
Actual measured inductance
Self resonant frequency without shield (no coax cable)
Self Resonant frequency with shield (no coax cable)
Total self resonant frequency including coax cable
Damping resistance value once coil is mounted and damping is adjusted
Coax cable type used and length
Frequency range of your PI machine
Peak power in the coil
MOSFET device number

You are on your way to getting a good understanding of PI coil design.

bbsailor

[Largesarge]

Tinkerer:

Take a look at carl's coil winder in the projects section, that's what I'm going to whip up with a few mods of my own possibly. I got a nice little counter, with no reset though so I'll do a little addition each time I wind one.

I think this is a super thing your'e doing and if shipping wasn't so high to where you are I'd volunteer to assemble and initially test, ie. it don't blow up when power is applied. It would be pro bono of course, being a retired/disabled vet all I have is time to kill each day.

Good luck it'll will be interesting to see how things develop.

[Tinkerer]

Thanks for all the help

Hi Reg, bbsailor anf largesarge,
many thanks for all the help. The link for the DD coilbuilding and connecting to the HH detector is very helpful. I am impressed by the Hays coil molds and other parts. They really look good and are afordable.
The extensive explanations and specially the list of critical parameters from bbsailor are most useful.
The available database on rectangular coils seems to be restricted to the results of a search on the PI Forum. Anyway Eric has posted quite a few interesting comments about it.
It will take a few weeks until I will have another Oscilloscope, so in the meantime I tinker on the hardware. Maybe you could give me some good advise on that also.
Whenever I use the classical detector design, (crutch design with the handle and the arm support), my back aches after a relative short time. I have seen pictures of detectors that do not have an arm rest, but have a sort of forward curved handle that seems to be better balanced.
What is your opinion?
Tinkerer

[Reg]

Hi Tinkerer,


First, one of the best ways to minimize the weight problem of a metal detector is to use some form of a suspension system to hold the weight. This puts the weight on the neck and shoulders rather than the arm. The easiest is to use a bungee cord for the suspension. Selection of the diameter of bungie depends upon the weight of what you are trying to suspend. Larger coils do require a heavier cord.

The easiest and simplest is to tie a bungee cord to a wide strap that is loosely suspended around the neck. The bungee cord connects to the detector rod at the junction of the lower and upper rods on my setup. A more sophisticated setup will consist of a harness worn around the main torso that has a point where the bungee can connect. Here is a link to Doc's site. Just scroll down to see his "swingy thingy" to get an idea of what I am talking about. ON Doc's, he connects to the detector just below the control box. I prefer to connect at the junction of the lower and upper rod. This allows for a longer bungee when then allows for a greater travel of the rod before things get tight.

http://www.kamakazi.com/

I use this type of a suspension setup for hours at a time. Without it, I am tired in just a couple of hours. My harness is simple and home made. My dad is 79 and simply uses a shoulder strap he hangs around his neck like a loose necktie instead of a harness. The bungee connects to it and to the detector.

As for the upper metal detector rod, I use the conventional S handle design. This is the common design today. I use a Whites S rod and their lower rods. The setup is cheaper than most others and works very well. I also suspend either the control box with batteries under the forearm, or I will have the battery pack in a separate box located under the forearm. The battery weight in the back balances the coil weight in the front, making the unit better balanced.

On a totally different note, I haven't looked that long at the design you have for a PI but I can see where it would be difficult to get down to 10 usec. My guess is it will take some work to obtain a large improvement. As an example, I would increase R 5, the input resistor from 1 k to a 2.2 or 3.3K. This will reduce the gain but increase the speed of the preamp. The resistor in the + input should be changed also to match the input for the best response. The shorter delay should make up for any other losses.

Another thing that can be done is to increase R3 from 10 to 20 ohms. Again, this reduces the power into the coil, but will allow a faster recovery at the input. Again, it is a tradeoff but if smaller nuggets are the object, then getting the delay down is the only solution.

Again, keep your coil inductance between 200 uh and 300 uh to speed things up. This will allow your damping resistor to be increased.

If you decide to try a P channel FET, then R2 has to go away. One can simply jumper for test purposes, but only if a P channel FET is used. If a P channel FET is tried, I would recommend a avalance voltage rating of 250 to 400V if possible. the higer voltage is better to keep the speed up.

I would probably try a P Mosfet to see how much of an improvement it makes. The problem is they are a lot less common than the N channel types.

The timing of the circuit is severely limited. If you reduce the delay, you will reduce the pulse width. It appears the pulse width is equal to the delay, and to the pulse sampling. One could change this by cutting and jumpers if desired. A 10 usec delay and 10 usec sampling is fine, but a 10 usec pulse is a little short. One could try it initially, but I suspect you would have to "and" some of the stages to increase pulse to a more desired level. As a quick test, one could simply change the timing factors and use a 10 usec pulse for everything to see just how it all works. However, in the long run, I suspect a longer pulse will be needed. On the other hand, I doubt a longer pulse will do much for nuggets up to a few grams in weight.

For testing purposes, all of this could be done, but I think the HH is a better choice overall. Even the HH needs a few changes including a voltage regulator for the best results. When done, one should be able to approach the 10 usec delay. If not 10 usec, then one should be able to get to 12 to 13 usec which is still good. That is about where the Minelabs are now.

Reg

[bbsailor]

Bungee cord for a coil support system

I find that a doubled 1/8 inch diameter bungee cord works well for my 15" WOT coil or any other coil about that same size and weight. This cord stretches pretty well and can be set by adjusting the bungee cord length to lift the weight of the coil, shaft and part of your arm weight off the ground. With my set-up I can hunt for 8 hours or more without getting a sore arm. West Marine stocks this 1/8" bungee cord. I can guide and glide the coil over the ground with ease.

bbsailor