Working on PCB-IB PI-Coil

[Anonymous]

I am now working on a new type of PI coil, that
- has low resistance (0.046 Ohm)
- low eddy current effects (coil wires 0.4 mm, connecting more wires paralles)
- low inductivity, enabling low time constant (tau) to achive fast high current switching thrugh the tx-coil
- seperate receive coil, placed on a balanced position (IB-Coil-Type)
- high transforming effect (rx-coil windings >> tx-coil winding)
- concentric, placed on a simple pcb (ca. 20 cm diameter)
Anybody interested?
Aziz

 

[Anonymous]

What is the reason why you chose to go to a printed circuit board for making the IB coil?
Could you also make a Double D coil for a PI detector using this technique?
What are the current IB printed circuit coil characteristics (meaning inductance) for the TX, RX and feedback coils?
Does the IB coil require a shield to minimnize noise and the ground effect?
I'm sure if you post the coil pattern, some people may want to etch their own coils for experimental purposes.
Thanks for the innovative work.
bbsailor

 

[Anonymous]

Hello,
my answers following in the same order of your questions:
>What is the reason why you chose to go to a printed circuit board for making the IB coil?
There are many reasons for. First of all, it is very easy to build such coils and much cheaper then others. It solves mechanical problems of balanced coils, because the position of the balanced coil (rx-coil) is computed thrugh the magnetic fields of the tx-coils. The coil is mounted on the pcb board tight and if you take the concentric one, this is size invariant, thus is good temperature compensated.
It makes the coil very thin and lightweight.
Of course, pcb-coils has many other disadvantages: low Q (Quality) of coils, high resistance and lower inductivity (restricted pcb-area). I found one way, to live with the restrictions.
>Could you also make a Double D coil for a PI detector using this technique?
Yes of course. In this case you need two side pcb-board or two one-side. The concentric coils (no other forms recommendend) would overlap, so one coil would be on top and the other one on bottom side. I do not recommend double D coil types, because they give different signal phases for the same object.
Double D is the easiest type of balanced ib-coils. If you take two same simple pcb-coils (spiral from center to outside), you can have more overlapping area than on circle-type coils.
>What are the current IB printed circuit coil characteristics (meaning inductance) for the TX, RX and feedback coils?
My current prototype is two sided concentric ib-pcb coil. I made this 1 1/2 years ago. The TX and feedback coils have round about 26 Ohms (due to thin 35

 

[Anonymous]

I add some additional comments on my previous comment here.
The most advantage of the pcb coils are, that they will have very small variance and equal quality and electrical characteristics. No more complex and expensive adjustments for each coil is needed.
They are extensively used on the HF area, where the coil is part of the pcb. Of course, today the chips also including coils on silizium.
For my PCB-IB PI coil, I took 16 single wires (0.4mm width, 35

 

[Anonymous]

If you have been following the various posts about making PI coils, you will notice that PI coils function in a completely different way compated to IB coils. IB coils can be considered to be a loosly coupled transformer where operating the receive coil in a continuous wave null provides the maximum sensitivity.
Mono PI coils on the other hand, function as both the TX and TX coil, alternating between each fucntion. The challenge in a PI coil is to make the coil have low capacitance so the receive function can be turned on as fast as possible to detect low conductive items like gold.
When using a DD coil style, null balance is not necessary. The DD style coil removes the RX coil from being directly connected to the MOSFET TX circuit. This has a few benefits: better ground balance; smaller pulse voltage to allow RX circuit to turn on faster and seeing less circuit capacitance from MOSFET. The last two of these tend to make the RX coil have a faster response time.
High power PI detectors depend on a high current pulse to create a highly induced voltage. Therefore, low coil resistance is a key parameter, something that thin printed circuit traces don't do as well as traditional thicker wire coils. However, there is a low power PI design that operates in about the 10K PPS range that uses higher resistance coils and even series resistance to minimize the current allowing early sampling. This type of PI detector is where the printed circuit traces could have some benefit. There is a commercial Tesoro pulse induction machine that uses printed circuit coils.
The challenge for any PI coil is keeping the total capacitance low so the total self-resonance of the coil, shield and coax cable is as high as possible.
Keeping the coil (coil, shield and coax wire) self-resonance above approximately 500KHz is a good figure of merit to help determine the potential speed of any PI coil.
In the PI mode, what is the self-resonance of your printed circuit coil when conncted to the shield and coax wire (no damping resistor)?
bbsailor

 

[Anonymous]

Tesoro used a PC coil on its Sand Shark, it was a concentric spiral pattern It was the limiting factor in it having much depth, the comment I saw from a White's engineer was that it was producing a large capacidence field with the earth. I notice no one else has used one! Don

 

[Anonymous]

Hi Aziz,
There is no real need to design a TX coil with a resistance of 0.046 ohms. Except for Minelab PI's, the TX winding could be 1.0 ohms for most PI's, and work fine. My PI's are even higher, with the Deepstar at 4.0 ohms and the Goldquest at 6.0 ohms. Minelab coils are 0.4 ohms and any significant departure from this does not work for some reason.
Eric.

 

[Anonymous]

For making mono homebrew PI coils, nothing beats the old-fashoned AWG 22 to 24 stranded hook up wire wound in a 10.5" diameter coil with 21 to 22 turns covered with a spiral wrap and 3M 1190 copper fabric tape to act as the shield and fit into a Hays Electronic coil housing. Printed circuit coils would seem more difficult to adequately shield.
bbsailor

 

[Anonymous]

GET ERIC TO SHOW YOU HIS TRIPLANAR-CONCENTRIC COIL HE BUILT USING THE OLD GARRETT DEEPSEEKER CO-AX HOUSING.

 

[Anonymous]

Hello bbsailor,
I know the technology of the PI

 

[Anonymous]

Hello Eric,
yes, you could also measure with low current flows thrugh the tx-coil. But the signal-to-noise ratio decreases (on rx-side), when using low current.
It depends on the amplification stage of the rx. To take low noise op-amp would bring more to enable more amplification, than driving the coil with more current power.
The very low resistance comes from circuiting all the 1 wire loop coils parallel. Maybe this too low, producing more heating power (on MOSFET and dumping resistor), getting battery quick empty.
Aziz

 

[Anonymous]

Hi Aziz,
I agree that having more current through the TX coil, will improve the signal/noise ratio in the receiver. However, you can also keep the current the same and increase the number of turns on the TX, to get the same result. Simply put, the magnetic field that the TX coil generates is the product of the Amps and Turns. You get the same field fron 10A in 1 turn, as you would from 1A in 10 turns. The inductance goes up as the square of the turns, so whatever the 1 turn coil was giving, the 10 turn coil will have 100 times the inductance. A 200mm coil would normally have about 25 turns to give around 300uH inductance. This number of turns for a pcb TX coil is difficult on a single side, without using thin tracks and getting a high resistance. However, I would aim for 1 ohm resistance and instead of paralleling tracks, make them into extra windings, so that you can reduce the current needed.
Eric.

 

[Anonymous]

Aziz
Could you provide functional block diagram of your virtual metal detector in terms of common pieces of test equipment functions, such as pulse generator, wide-band amplifier, phase comparitor, etc. Also indicate which 5% you had to physically make? This will help me visualize your design and better understand your comments.
If I understand you so far, you are saying that using a IB type coil on a PI detector with a separate TX and TX coil allows additional intelligence about the target to be extracted?
Thanks
bbsailor

 

[Anonymous]

Using a balanced searchcoil does indeed provide extra information about the target. The method is referred to today as "PIB" or "Pulse Induction Balance". Eric Foster produced the first such detector way back in the 1980's. The method is not problem free as the extra information is subject to being modified by ground mineral and sea water signals which can cause massive errors.
Many other problems will need to be addressed before the PIB has a chance of becoming popular. There are a number of people around the world (including myself) who are working on PIB's as well as some other PI methods. I tested a PIB in England last year and I will be taking one to Ganes Creek, Alaska in June to see how well it does at finding gold nuggets.

 

[Anonymous]

Dave,
Is the extra inteligence obtained from a target using a PIB coil in the form of phase information?
I assume this additional target information is in addition to the traditional information obtained from target decay time analysis.
Little had been said about this PIB coil technique in this forum before. If it has, I have been asleep.
Thanks for clarifiying this new type of coil.
It looks like the Wiltron 352 low frequency phase angle differential meter that I just ordered from e-bay may come in handy.
bbsailor

 

[Anonymous]

PIB is the name of the detector not the coil. PIB stands for "Pulse Induction Balance". This is a combination of a PI or Pulse Induction detector and an IB or Induction Balance detector. VLF and TR detectors are all IB's as they used induction balanced searchcoils. The extra information is obtained by sampling the receive signal during the detectors transmit pulse on time. I am afraid that your phase meter will be of no use at all for work on a PIB.

 

[Anonymous]

Hi Eric,
high resistance coils became like heating resistors. The heating power P = I^2 * R, where R the coil resistance and I the current thrugh the coil, would reduce the effectivnes of the system. More battery power is taken for heating the coil.
The more winding for tx coil, the more the inductivity, the more the time constant, to saturate the magnetic field and of course, the more capacitance of the coil.
I think, with a very low resistance tx-coil with lower inductivity, you can switch the coil very fast for a very short period. This would save some battery energy.
Todays MOSFET or other switching transistors, it should be possible to switch 10-100 A for a very short period. Almost energy would taken for magnetic field not for heating.
Of course, with one winding tx-coil, you should take another coil with more windings for rx-coil, to transform the eddy-current induced signals, to have a good signal-to-noise ratio.
I will show you my new IB-PI-PCB coil, that combines all the advantages, I previously mentioned.
Aziz

 

[Anonymous]

Hello bbsailor,
I am using very simple components:
- my search coil (IB-PCB) as sensor
- low noise amplifier for rx-part, 1 stage, gain factor 500-1000x, ac-coupled with 1 kHz high-pass filter (1. order)
- Hi power amplifier with low harmonic distortion, sending a pure sine-wave to the coil without capacitance connected to the tx-coil (not using in resonance-mode)
- my PC with data acquisition module (multichannel 16 bit A/D, D/A converters)
- rest of the complex part is software (analyzing the target information, doing FFT, filter, ground balancing, etc.)
I do the ground balancing realtime through the ground balancing coil (1 winding). Thats, what really is innovative and brings more sensitivity, regardles of ground effects. The software can make complex filters, to cancel all the unwanted effects. Doing almost in software, you do not need more hardware.
So I

 

[Anonymous]

Hi Aziz,
Something has to heat up somewhere. If not the coil, then the switching transistor and damping resistor. You can certainly pour lots of current into setting up the magnetic field, but where does that energy go when you switch the field off? Unless you are using an energy recovery system, then it is dissipated in the power transistor and damping circuit.
A one turn 20cm coil would have about 1uH inductance. If the dc resistance is 0.04 ohms, we end up with a time constant of 25uS. We also get the same TC for a 250uH coil with a resistance of 10 ohms. My Goldquest SS has a 300uH coil with a total TX circuit resistance of 33 ohms. Peak pulse current is 400mA with a repetition rate of 10,000 per second. The mean TX current is 70mA and I can detect a US nickel at 14-15in with an 11in coil. The only thing that gets warm is the 27 ohm resistor in series with the Mosfet drain and the coil. You could do away with this if the actual winding resistance was 33 ohms instead of 6 ohms. With such a large surface area, the coil would not get even noticeably warm.
Both high and low current, and high and low coil resistances and inductances work, with not much difference in the end result. It will be interesting to see how your scheme works out.
Eric.

 

[Anonymous]

Hello Dave,
You said: "The extra information is obtained by sampling the receive signal during the detectors transmit pulse on time".
I am not at all familiar with PIB detctors but your answer to my question could help me visualize how it works.
If you are sampling the RX signal during the TX on- pulse, what is the way you critically null the RX coil to the TX signal so as not to mask the desired RX signal by the high level of the TX pulse?
Thanks
bbsailor