DD and Concentric

[plboy]

I should imagine that it was fairly tricky to calculate the parameters for this coil. How good was the nulling? i.e. what was the residual voltage in the RX with no target present?

I've also written a calculator for concentric coils, and that was difficult enough to get an accurate result. How many prototype boards did you build, before you got one that worked properly?

what about concentric

Some pictures on my album are too big to be displayed correctly.

Here is a compressed file, containing more pictures:

http://ibcoils.tripod.com/wien2.rar

Aziz

I looked at the pictures at ibcoils.tripod. This is very interesting, specially the pictures of the magnetic fields. Did you use the "Maxwell"software by Anvil for that? I find it very hard to make precise coils for IB. Your system seems to be the perfect answer.
How do you calculate the exact number of turns needed to achieve balance?
Is there a way to make final adjustments once the PCB has been etched?
Tinkerer

The software for this purpose is developed by me for more than 10 years. It is based on simple wire-by-wire calculation using the biot-savart law. I implemented an optimizer for this purpose and can find the optimum balance parameters (geometry, number of turns, etc.). There are many ways to find an optimum balanced IB-coil. Often, I decide to set the number of turns for the tx-coils and then calculating the rx-coil (number of turns, radius start, radius end, etc.). And sometimes, if I can not find a good coil, I change the number of turns for tx-coils. Only the limitation is the surface area (diameter) and the thickness and gap of the wire on the pcb.
On the picture shown, you can see, that almost all of the area of the coil is used. Especially the rx-coil has a transformation function (like transformer) and is built-in hardware amplifier.
All the integrated area of the rx-coil will induce no signals (due to balance). This balance will be disturbed by the environment (metal, ground, mechanical shock, etc.). The rx-coil is kept very thin to get more windings (higher transformation). The tx-coils are thicker, to provide more current and therefore higher magnetic flux.

For the fine and ground adjustment, there is one turn coil between two tx-coils. This will compensate all the effects, caused by ground and manufacturing tolerance. So, you need two signal feeds: one for the tx-coils (connected serial) and one low-power for the compensation coil. With this technique, I can achive a very good balance and get more sensitivy. The place for compensation coil is not important. But the gap between the two tx-coils is a good place for this because this gap is physically needed.

Here comes: I realized this metal detector with only a Laptop + 24Bit/96kHz sound card and very simple amplifiers.
The complexity of the detector is really high and you need very good software techniques. Especially to compute a real-time signal for the compensation coil. You need a big number crusher for this. It is possible to realize a good and very cheap detector usung a Laptop.

You all should start working now. I gave you too many hints.

[Tinkerer]

IB coils

DD Coils are very easy to build and balance. But they are due to geometric variant type of coils, any mechanical distortion will cause balance asymmetry. Especially, when temperature effects will occur. Further on, any mechanical load will cause asymmetry. To overcome this, a motion-mode is more convenient.

On the other side, concentric coils are very difficult to build. But they have many advantages.

Look to the following concentric coil: (PCB-IB-Coil)

http://ibcoils.tripod.com/album/index.album/induction-balance-type-pcb-coil?i=0&s=1

Have you ever seen such a coil?

I looked at the pictures at ibcoils.tripod. This is very interesting, specially the pictures of the magnetic fields. Did you use the "Maxwell"software by Anvil for that? I find it very hard to make precise coils for IB. Your system seems to be the perfect answer.
How do you calculate the exact number of turns needed to achieve balance?
Is there a way to make final adjustments once the PCB has been etched?
Tinkerer

[Fred]

Exactly, its not a smart game.

Yes, very tricky and complex.
My work was interrupted since my notebook gone broken. Maybe I will continue next year. Especially, on new dual-core notebooks, this very time consuming technology (much processing power is needed) became interesting and I can achive more measurements per seconds (>100-200). New tablet pc´s or small notebooks could make this metal detector possible. There is very simple hardware needed. All the detector parts are made via software.

this should be appliable on conventional method of making concentric coil? why not?
pcb coils are not so "fresh" idea. there are dozen patents on this one.
but you already explained some drawbacks.
better question is how to balance concentric coils more accurately. how to lower residual voltage more. overbalnce,underbalance?
without engaging specific md design to achieve proper method of scoping is coil underbalnced or overbalanced?quiaozhi's earlier posts on this and other threads are subjecting this pretty close,yet pretty academic,without practical examples.
turning tablet pc into md is interesting idea! why not some symbian cellphone?N95?
to many questions - so few answers.
i am positive here about your ideas.
regards

Yes, very tricky and complex.
My work was interrupted since my notebook gone broken. Maybe I will continue next year. Especially, on new dual-core notebooks, this very time consuming technology (much processing power is needed) became interesting and I can achive more measurements per seconds (>100-200). New tablet pc´s or small notebooks could make this metal detector possible. There is very simple hardware needed. All the detector parts are made via software.

[Qiaozhi]

I made four small ones to test the software. Than I simulated normal coils and found the geometric positions for a balanced state and tested this on the oscilloscope.

Than I made a big (20 cm) one, that was working perfect. The balance is dependent on the tolerance of the printer for layout (DIN A4) and of course the accuracy of your hands. The induced non-balanced signal was some fractions of mV.

To overcome the tolerance and ground signals, the balance is made additionally by a second coil (1 turn). So I achived a best balanced IB-coil and on this mode, you get the surprise!

Much more sensitivity and depth.

The coil is computed thru biot-savart law. Integration of small peaces of wire. Very time consuming but very accurate.

Very nice ... sounds like some excellent work.

I made four small ones to test the software. Than I simulated normal coils and found the geometric positions for a balanced state and tested this on the oscilloscope.

Than I made a big (20 cm) one, that was working perfect. The balance is dependent on the tolerance of the printer for layout (DIN A4) and of course the accuracy of your hands. The induced non-balanced signal was some fractions of mV.

To overcome the tolerance and ground signals, the balance is made additionally by a second coil (1 turn). So I achived a best balanced IB-coil and on this mode, you get the surprise!

Much more sensitivity and depth.

The coil is computed thru biot-savart law. Integration of small peaces of wire. Very time consuming but very accurate.

[Qiaozhi]

Inductance is relatively low. But I made double sided PCB-Coils, to increase the inductance. Due to relative thin copper wire for tx-coils on pcb-board, the resistance is too high (Coil Q is bad). It is not suited for high-current-PI´s. (the shown coil).

This coil, mentioned above, is driven in a capacitiveless current mode. So, not used in a resonant circuit (LC-circuit). Therefore the inductance is becoming not really important.

I used this coil on a Notebook with sound card and a simple amplifier (driver + signal). A continious changing operation frequency signal was sent and the received signal was processed at the same time (with some latency).

I made a high-current version of balanced IB coil, which have few number of turns for tx-coil. The balance is made via geometry and resistance of the two tx-coils. (tx-coils had paralell windings)

I should imagine that it was fairly tricky to calculate the parameters for this coil. How good was the nulling? i.e. what was the residual voltage in the RX with no target present?

I've also written a calculator for concentric coils, and that was difficult enough to get an accurate result. How many prototype boards did you build, before you got one that worked properly?

Inductance is relatively low. But I made double sided PCB-Coils, to increase the inductance. Due to relative thin copper wire for tx-coils on pcb-board, the resistance is too high (Coil Q is bad). It is not suited for high-current-PI´s. (the shown coil).

This coil, mentioned above, is driven in a capacitiveless current mode. So, not used in a resonant circuit (LC-circuit). Therefore the inductance is becoming not really important.

I used this coil on a Notebook with sound card and a simple amplifier (driver + signal). A continious changing operation frequency signal was sent and the received signal was processed at the same time (with some latency).

I made a high-current version of balanced IB coil, which have few number of turns for tx-coil. The balance is made via geometry and resistance of the two tx-coils. (tx-coils had paralell windings)

[Qiaozhi]

The number of turns and geometry of concentric coil is very critical. Therefore I made a software, where such coils can be computed and designed.
The link above shows such a balanced concentric coil. But you can not use such a coil in traditional metal detectors. A very new design is needed.

What value of inductance can you achieve for this PCB coil? I expect it is quite low, and probably most suited to a PI detector. Is that true?

The number of turns and geometry of concentric coil is very critical. Therefore I made a software, where such coils can be computed and designed.
The link above shows such a balanced concentric coil. But you can not use such a coil in traditional metal detectors. A very new design is needed.

[Qiaozhi]

thx Qiaoshi
I am made some DD coils and put the gap in a convenient place - no problem...but I want to make a good coil..... which type of coil will be recommendation

DD coil is the easiest coil to make. OO coil is also easy, but difficult to null properly. For best overall performance and good pinpointing, the concentric is the most popular. However, the problem here is calculating the correct number of turns for the nulling coil, remembering that the TX and nulling coils are connected in series, and their combined inductance must be a specified value. The concentric is the most difficult to null properly of all types. As you have already seen from some of the replies to your original question, each type has its own advantages and disadvantages.

My advice would be to buy a commercial Tesoro coil. Once your detector is working ok with that coil, then try to build a copy to achieve the same results. You can learn a lot from doing this.

DD Coils are very easy to build and balance. But they are due to geometric variant type of coils, any mechanical distortion will cause balance asymmetry. Especially, when temperature effects will occur. Further on, any mechanical load will cause asymmetry. To overcome this, a motion-mode is more convenient.

On the other side, concentric coils are very difficult to build. But they have many advantages.

Look to the following concentric coil: (PCB-IB-Coil)

http://ibcoils.tripod.com/album/index.album/induction-balance-type-pcb-coil?i=0&s=1

Have you ever seen such a coil?