Induction balanced spider coil development

[Aziz]

I looked for a price list for a big IB-PCB-Coil (ca. 30 cm diameter) manufacturing and found horrible prototype prices for such coils. Nevertheless the Q-Factor on PCB-Coils are very bad. So I decided not to develop an IB-PCB-Coil anymore.

But I found some very interesting tips on spider coils: Less capacitance, low impedance, high Q and very easy to build.

On next days, I will simulate some spider coils with my software and you will be surpriced about the sensitivy and power. I will do a big detection area for rx-coil with much more windings (analog amplification). Spider coils are relatively easy to build, even if they are an induction balanced types.

to be continued..

[Tim Williams]

Looking forward to your findings.

Tim

[Aziz]

Now implemented the spider coil modelling in the software. Also revised the code for some speed enhancements using multi-core-CPUs capability of the Personal Computer. Calculating the magnetic field can be easly parallelised and it can be done on seperate cores in the CPU. Thus the simulation will be faster in the future.
;-)

http://imgbox.de/users/zet/Spidercoilmodel.gif



On the picture above, you see the TX-Coil 1, which is the main transmit coil. TX-Coil 2 is for canceling the field in the inner loop of the coil. RX-Coil is shown in a region from radius 1 to radius 2, enabling a big area of detection region.
To minimize the resistance, eddy currents and skin effect of the TX-Coils, I will use a HF wire (with many seperate isolated thin wires). The idea is from the spider-coil for simple LW/MW radios.
For the RX-Coil, a very thin wire with lots of compact windings between radius 1 and radius 2 is used. Resistance is trivial and can be neglected (voltage driven, not current driven).

to be continued..
Aziz

[Tinkerer]

Aziz,
I have experimented with IB-PI spider coils. I agree with you that the capacitance is much reduced if the right compromise is realized.
I found 2 problems.
I have difficulties in winding several coils to exactly the same standard manually. Since the induction is extremely sensitive in its balance, any slight difference in the tightness of the windings produces a difference in balance. With the bucking coil this can be compensated, but I find it imperative to include an electronic adjustment possibility.
The second problem, is the reduced magnetic field strength as compared with a traditional PI coil of similar coil current throughput. The design of the coil is critical to obtain best performance.
Tinkerer

[Aziz]

Hi,

Aziz,
I have difficulties in winding several coils to exactly the same standard manually. Since the induction is extremely sensitive in its balance, any slight difference in the tightness of the windings produces a difference in balance. With the bucking coil this can be compensated, but I find it imperative to include an electronic adjustment possibility.

Yes of course, a simple adjustment electronic networks will be used for the nulling of the RX-coil. Or the number of turns for the RX-coil will be slightly chanced (some more or less windings, radius 2 of the RX-coil will be used for balance). Or both of them. Additionally, a low current driven third TX-coil (one winding) is used on the front end for automatic fine balance adjustment and ground balance.

If you build an IB-coil, you can not avoid tolerances. As you can see, the RX-coil arrangement is an integrated field of radius area (ring from radius 1 to 2, similar to my IB-PCB-coil). So the number of turns gets irrelevant, as you meet the radius specification for RX-coil. If you take a very thin wire, you can realize hundreds or thousands number of windings. The more RX-windings, the more difficult to balance. Of course, the RX-coil can have a constant height to enable more windings. I do not want to count the number of RX windings. I want do this with a coil frame, which has the starting radius and will stop at the end of the radius 2 and will give some more windings for balance adjustments.

I can calculate the geometry for an ideal balanced coil arrangement and the user must keep to the given data as accuracy as possible.
That was the main reason, why I spent a lot of time with IB-PCB-coils. There are no balance problems with IB-PCB-coils but many electrical disadvantages (high resistance, low Q, less windings, high costs for the PCB and small coil dimension).

Aziz,
The second problem, is the reduced magnetic field strength as compared with a traditional PI coil of similar coil current throughput. The design of the coil is critical to obtain best performance.

One should keep the coil width (radius 2 - radius 1) as small as possible. The more the coil width, the more magnetic fields are cancelled. Especially inner windings will cancel the magnetic fields generated of outer ones. Therefore a wider coil is less effective than a compact one. So, a reasonable (big enough) hole in the TX-coil is neccecary. I saw some horrible spiral PI coils here or on another forum with small holes!

I will do a compact one with flat spider TX-coils as shown above. So the width will be kept as small as possible and reasonable (depends on coil impedance, wire diameter, etc). I intent to use a thick HF wire (0.6 - 1.0 mm) for the TX-coil. I do not know, where to get such a coil. I have to look for a good one. The reason for HF wire is because of the thickness of the TX-wire (causing more eddy currents but lowering the impedance).

Fortunately, the TX-coils can be build easily and can also meet the geometric specification (radius 1, radius 2) by winding them one wire nearby another without any gap on crossing places. The gap between other places can be made also easy by putting a small piece of TX-wire between the windings. After fixing the TX-coils, the small pieces can be removed. To meet the TX-coil geometry, a special spider coil frame is necessary. There are many instructions on the internet how to build a spider coil. Old style radios for instance use such a spider coil.

I will test the environment on the PC using a sound card and simple amplifiers (TX-Driver and RX-amplifier). I still want to develop a Laptop MD with simple electronics and state of the art software.
;-)

Aziz

[Aziz]

Finally, I got a HF wire today. 29 thin seperately isolated wires.
Total ca. 0.7 mm diameter, including silk coat.

So the development of the coil can go on.
;-)

[Tinkerer]

PI-IB coil configuration

Aziz,
the attached picture could represent a TX coil with a bucking coil above it. If we add a RX coil inside, we would have maximum use of space/diameter for it. Unless we put it below the TX coil.
How would that work out?
The attainable dept of detection depends on the diameter of the RX coil.
The minimum detectable size target depends on the shortest delay possible.
The shortest delay depends a lot on the capacitance of the coil, shield and cable assembly.
If we could build a coil assembly with a self resonating frequency of 1Mhz, we would be able to detect very small targets.
Increasing the diameter of the coil rapidly increases the capacitance.
High coil current increases dept and sensitivity.
The questions now are:
What is the minimum size target that we want to find?
What compromise do we want to reach between minimum size target and maximum dept?
Where do we want to set the limit of power consumption? This limit can probably pushed quite a bit higher today, with the lightweight Li-Ion batteries.
I am trying to look at the above limitations from the angle of a commercial design.
The magnetic field visualization above has been made with the MAXWELL SV software that can be downloaded free at http://www.ansoft.com/membership/login.cfm?campaign=26

[Aziz]

By the way, my HF wire has 30 thin isolated wires (I made a mistake during counting). I have bougth 44,5 m and it has ca. 3,2 Ohm (0.070836 Ohm/m).
It should be enough for both TX coils. I will try to match the impedance to 2-3 Ohms.

HF-wire summary (for TX coils):
wire: 30 x 0.16 mm (AWG 4
diameter: 0.7 mm including the silk coat.
Length: 44,5 m
Ohms: 3,2 Ohm (0.070836 Ohm/m)


I have seen your coil above and I assume that they are same size and characteristics (winding, impedance, etc.).
Would you give me the diameter and the distance of the coils? I need the mean coil diameter and the mean distances from each other (from center to center of the coil windings). Then I will start the simulation of the coil arrangement and will optimize some RX-coil dimensions and positions. I am sure, you will be eager to the results of analysis.
;-)

Aziz

[Aziz]

Without running my coil application, I can tell you, that in this coil configuration above, only rx coil places above the bottom tx coil makes sense. Below the coil configuration, there is no canceling magnetic field. Best rx coil height position will be on the same top tx coil.

[joecoin]

Tinkerer,

I would not let power consumption be too large of a constraining factor. I could see LARGE coils with great depth being mounted on wheels. I would like to have such a creature.

[Tinkerer]

Hi Joecoin,
Without going as large as wheel mounted coils, what do you think of the ergonomics of brush cutters. I often see people run these machines 8 hours per day, yet they are of considerable weight. The operators also are able to move the cutting head at great precision, but mostly swing it from side to side just like a metal detector.
Tinkerer

[Tinkerer]

Aziz,
The above coils are just examples of models created by the MAXWELL SV software. they are equal coils. For a PI I think it would be preferable to have a bucking coil that generates the lowest possible field, just enough to cancel the field in the RX coil.
What I dont quite understand, is how we are able to receive the eddy currents, without the RX and bucking coil fields canceling out.
Tinkerer

[joecoin]

Hi Joecoin,
Without going as large as wheel mounted coils, what do you think of the ergonomics of brush cutters. I often see people run these machines 8 hours per day, yet they are of considerable weight. The operators also are able to move the cutting head at great precision, but mostly swing it from side to side just like a metal detector.
Tinkerer

Funny you should ask, I own a commercial grade brush cutter (aka Weed Eater). It has a harness over the shoulder and around the back. Now that you mention it, it does balance rather nicely.
There are several adaptive aids for metal detectors, one is called the swingy thingy.

[Aziz]

The above coils are just examples of models created by the MAXWELL SV software. they are equal coils.

If the coils are really equal, then different current flow through the coils must be feeded. If they are connected in serial (same current), then tx-coil must have nearly twice more windings than the bucking coil. I modeled this configuration with my software and found this.

Aziz,
For a PI I think it would be preferable to have a bucking coil that generates the lowest possible field, just enough to cancel the field in the RX coil.

Ok, in this case, the bucking coil should be relative smaller than the tx coil. Half to 2/3 diameter of the tx-coil would be reasonable. The bucking coil must have enough radial distance from the tx coil. Also the rx-coil should be very close to the bucking coil (same radius, however above or below the bucking coil). Then this could be realised very easy. All of the VLF-TR detectors working the same way.

Aziz,
What I dont quite understand, is how we are able to receive the eddy currents, without the RX and bucking coil fields canceling out.
Tinkerer

In the induction balance mode, you will receive the eddy currents. Yes. Remember, eddy currents on target cause secondary magnetic field and this deforms the primary stationary magnetic field. Induction balance means, that the coil is acting as a "zero-coupled" transformator, when the stationary magnetic field is not changed. However, this primary magnetic field is also changed by the material below the search coil (paramagnetic, diamagnetic, materials etc. are changing the permeability of the coil arrangement). This effect also disturbs the primary magnetic field and deforms localy. Then our "zero-coupled" transformator gets working. So you get more information on target as you won't get in a standard PI MD (one search coil).

Standard PI MD just analyses the decay characterictics of one or more window time slices. Unfortunatelly, some important signals are destroyed via integration or average operation (low pass in frequency domain). Also no phase information of the induced signals are processed. I think, the standard PI technology can not give more results with only one coil.

I know the real secrect of how getting a very good MD and depth:
1. Signal analysis (Where are the signals? And how they can be detected and processed?)
2. Increase the SNR of your MD.
3. Cancel all instability (instability reduces the SNR).


In ALL stages of your development, you should think about, how to achive these simple three points! Point 1 is really well known and you can not change the physics. But on point 2 and 3, you can make a lot of.

Aziz,
more on next time..

By the way: The search coil size defines the target object size. If you want to search smaller objects with big coils, you just have to increase your SNR.

[Tinkerer]

Aziz,
I agree with you on the above, except on the quote below, when applied to the PI.

By the way: The search coil size defines the target object size. If you want to search smaller objects with big coils, you just have to increase your SNR.
On the PI, the target response is defined/limited by the target's TC curve. This means that most of the target's eddy currents are gone after 2 TC's. Take a small piece of gold like a 0.1 gram nugget or a hollow gold ear ring, and the TC is about 3us. That means that with a delay of 10us to the first sample, very little signal amplitude is left.
Now, I have never worked with VLF detectors, but the few experiments I made showed that a VLF frequency of 30Khz or so is high enough to detect such targets.
Why?
One more question about the IB coil arrangement. I can attest that the coil arrangement you describe above does indeed work. I am using it with the TX and bucking coil in series. However, I get considerably less amplitude in target response with this IB arrangement than when I use the same coil, with the same coil current in the traditional way, with only TX and RX coils. This is why I believe that it is very important and useful to analyze the interacting fields to find the best possible combination of coil disposition, coil characteristics and coil currents.
Please continue analyzing.
Tinkerer