Perhaps I should have applied for a patent?

[Ferric Toes]

Some interesting specifications that I have dug up:
polyester core material plated with a layer of nickel, a heavy layer of copper, and a layer of nickel
200 threads per inch
impedance 0 - 100 kHz < 0.5
attenuation >60 db 10 MHz - 3 GHz (80 db@10 Mhz, 93 db@100MHz, 65 db@3 GHz)
resistance 0.04 (specified as ohms/(square symbol), assumption is per sq mm)
aggressively conductive acrylic adhesive

Ohms per square can be any size of a thin conductive sheet. It could be per square mm but it is easier to measure on a larger area. Using 25mm wide tape, and in our case, the nickel coated copper tape, it is easier to measure a 300mm length, or even longer such as 600mm to avoid getting a very low reading. The ohm-meter has to connect to each end, and it's useful to have a bit of overlap of tape, by a test electrode that spans the 25mm width. Say we get a reading of one 1ohm. You then divide the tape length of 300 by the width of 25mm = 12. To get the value in ohms/sq (that is a 25mm square), simply divide the resistance by 12. i.e. 0.08ohm. This value will hold true for any area of that same material. e.g if the same tape was 50mm wide then a 300mm length would measure 0.5 ohms. We then divide 0.5ohms by the number of 50mm squares in the 300mm total length = 6. 0.5/6 = 0.08, the same result as before. 1.00 sq.mm would give the same figure as would 1 sq.m. of that material.

Eric.

[Ferric Toes]

That would be nice considering all the weeks of simulation in spice and yet nobody (not even Eric) has actually shown the design to work in the real world, at all, and isnt just the equivalent of a dummy load.

So my only question is the same as back on page 1: Does it actually work?

Yes, it does 'work' in the same way that a new rocket engine works on a test rig, but has not yet lifted anything off the ground. There is little point in guessing what it might detect or at what range in the real world, until the basic TX/RX development has been done. The receiver preamp is critical as would be the sampling and integrator performance. Only then will it be worth adding further amplification/filtering/DSP etc. to see what ranges we get on different targets. If we can detect 1mm diameter ferrous and non-ferrous spheres in the plane of the coil, I shall be happy. Meanwhile, no rush, no deadlines to meet; just an interesting sideline to investigate, and a big thankyou for those who are interested to do Spice simulations and put forward technical knowhow.

Eric.

[green]

Yes, it does 'work' in the same way that a new rocket engine works on a test rig, but has not yet lifted anything off the ground. There is little point in guessing what it might detect or at what range in the real world, until the basic TX/RX development has been done. The receiver preamp is critical as would be the sampling and integrator performance. Only then will it be worth adding further amplification/filtering/DSP etc. to see what ranges we get on different targets. If we can detect 1mm diameter ferrous and non-ferrous spheres in the plane of the coil, I shall be happy. Meanwhile, no rush, no deadlines to meet; just an interesting sideline to investigate, and a big thankyou for those who are interested to do Spice simulations and put forward technical knowhow.

Eric.

Could someone calculate the time constant for a 1mm diameter lead ball? I copied a formula from another thread awhile back and calculate .155usec but don't know if it's right.

[Qiaozhi]

Could someone calculate the time constant for a 1mm diameter lead ball? I copied a formula from another thread awhile back and calculate .155usec but don't know if it's right.

I get 0.145us, so we're in close agreement.

[green]

I get 0.145us, so we're in close agreement.

Thanks Qiaozhi

Question for Eric. If a 1mm diameter lead sphere has a time constant around .15usec a 1mm diameter gold sphere would be around 1.25usec. A 1.25usec target would be a lot easier to detect with spice. Are you thinking of a specific targets, material or time constant? For spice I need a time constant. Searched lead shot size for 1mm diameter, 1.02mm(dust) was the smallest they listed. A 2x2mm square or 1x2mm piece folded once(1x1mm, 2 layers) cut from regular strength aluminum foil should have a time constant close to the 1mm lead sphere if you are thinking lead. A 20x20mm square of the foil would have a time constant close to the gold sphere. A larger target that could be detected at some distance but at least the same time constant.

[Ferric Toes]

Thanks Qiaozhi

Question for Eric. If a 1mm diameter lead sphere has a time constant around .15usec a 1mm diameter gold sphere would be around 1.25usec. A 1.25usec target would be a lot easier to detect with spice. Are you thinking of a specific targets, material or time constant? For spice I need a time constant. Searched lead shot size for 1mm diameter, 1.02mm(dust) was the smallest they listed. A 2x2mm square or 1x2mm piece folded once(1x1mm, 2 layers) cut from regular strength aluminum foil should have a time constant close to the 1mm lead sphere if you are thinking lead. A 20x20mm square of the foil would have a time constant close to the gold sphere. A larger target that could be detected at some distance but at least the same time constant.

I should have been more specific. When I was involved in industrial applications of metal detectors, I had a set of test pieces which consisted of high density polythene rods with metal spheres embedded. These spheres range from 0.5mm to 5.0mm diameter and were labelled non-Fe or Fe. The non-Fe spheres were bronze and the Fe were steel. These test sticks were supplied by Goring Kerr, a company involved in metal detectors for the food and pharmaceutical industries. Unfortunately, I don't have these any longer. The aim was to detect the 0.5mm spheres in an aperture type coil. i.e. the product passed through the coil. The best we could do with a PI was to detect the 1.0mm spheres, both Fe and non-Fe. Other less conductive or non-magnetic metals such as lead or stainless steel were out of the question. Spheres are an idealised test piece as there is no orientation problems when passed through a coil. In the real world metal fragments can be irregular shapes, as in a broken off need tip, which happens in the textile and clothing industry.

For these tiny metal spheres the industry generally uses high frequency balanced coil systems operating at 100kHz or more, but P.I. detectors are used in certain cases, but must have a high sensitivity to smaller fragments. Coil design is obviously critical and that is one reason for my interest in pursuing this subject.

Eric.

[green]

I should have been more specific. When I was involved in industrial applications of metal detectors, I had a set of test pieces which consisted of high density polythene rods with metal spheres embedded. These spheres range from 0.5mm to 5.0mm diameter and were labelled non-Fe or Fe. The non-Fe spheres were bronze and the Fe were steel. These test sticks were supplied by Goring Kerr, a company involved in metal detectors for the food and pharmaceutical industries. Unfortunately, I don't have these any longer. The aim was to detect the 0.5mm spheres in an aperture type coil. i.e. the product passed through the coil. The best we could do with a PI was to detect the 1.0mm spheres, both Fe and non-Fe. Other less conductive or non-magnetic metals such as lead or stainless steel were out of the question. Spheres are an idealised test piece as there is no orientation problems when passed through a coil. In the real world metal fragments can be irregular shapes, as in a broken off need tip, which happens in the textile and clothing industry.

For these tiny metal spheres the industry generally uses high frequency balanced coil systems operating at 100kHz or more, but P.I. detectors are used in certain cases, but must have a high sensitivity to smaller fragments. Coil design is obviously critical and that is one reason for my interest in pursuing this subject.

Eric.

Just curious. If it's a secret I apologize for asking. When or why might a PI be a better choice over high frequency balanced coil systems in an industrial application?

Thanks for the answer about the targets. Will wait until you have done some testing.

If wanting to detect a 1mm sphere of lead(TC=.15usec) with a high frequency balanced coil system what frequency would be best?

[Sean_Goddard]

Eric, I used a current sensing circuit and when the di/dt was no longer above a certain level (via an integrator) I used this to cut off the Tx pulse as this indicated peak current had been reached. This meant the best use of batteries, achieving peak Tx power whilst nont having the Tx on for too long and over heating anything.

Any use to you, mines a Bottle of Bushmills ;-)

[Ferric Toes]

Just curious. If it's a secret I apologize for asking. When or why might a PI be a better choice over high frequency balanced coil systems in an industrial application?

Thanks for the answer about the targets. Will wait until you have done some testing.

If wanting to detect a 1mm sphere of lead(TC=.15usec) with a high frequency balanced coil system what frequency would be best?

No secrets here. It often boils down to the very basic fact that a PI works with an untuned mono coil and no balancing is required. Applications that I have designed for are a handheld detector for locating broken off needle tips in the manufacture of clothing: particularly baby clothes. This uses 1/2 of a large ferrite pot core so that the field only comes out on one face and the electronics is mounted on an aluminium plate on the opposite face of the pot core. That way we have no connecting lead to speak of and minimum capacitance. Then for basic textile manufacture, often several metres in width I used modular rectangular figure 8 multilayer printed circuit coils with a printed comb shield on the operating face. Electronics was multi-channel with one master and up to ten slaves so that all was sychronous. Then we have drop through detectors where a product comes down a pipe and if metal is detected a diverter flap shoots that portion into a reject bin. This is easy to do with a mono coil around non metallic section of the pipe, but the system has to react very fast.

In the late 1990's balanced coil detectors ran up to 1MHz, but today may go rather higher as quality standards in the food and pharmaceutical markets become stricter.

Eric.