Crude coil sheilding paint

[maikl]

We could divide shield discussions into VLF shields and PI shields. Quote from Eric Foster refers to the PI detectors. Shield geometry being parallel to the magnetic field lines can be accomplished with enough thin strips lead, copper (suitable for soldered) ... Paint is not suitable for this.

[simonbaker]

Link to Kev has set in post # 49 http://www.findmall.com/read.php?34,...370#msg-134370 provides answers for this topic.
Quote from Eric Foster:
"Shielding is essential for PI detectors if they are to have good sensitivity for small metal targets. Careful choice of the material for the shield, and also the positioning of the shield with respect to the coil, will result in no loss of sensitivity, in fact I cant think of any negative effects that proper shielding has.
As has been said, if the shield is parallel to the lines of force generated during the transmit period, and the material is thin enough, then there will be no measurable eddy currents induced in the shield. The other important factor is the shield materials conductivity. The lower this is, the faster any eddy currents will decay, and provided they decay before the sampling period of the detector, then they wont be seen.
A flat sheet of shielding, either above or below the coil, is not the best, although it is often used for convenience. A field plot of the coil will show that only immediately under the winding is the field parallel to the shield. Most of the shield therefore has eddy currents generated in it and it has to be made of a relatively poorly conducting material. This is why graphite and nickel paints are often used here. Again, it depends on how thick a coat you put on as to the decay time of these currents.
If you wrap the coil with a metallic tape, then all points on the tape are parallel to the field. The field being coaxial with the coil cross section when close to it. You can then use a more conductive material, and as it is completely enclosing the coil, the shielding efficiency is much greater. You do, of course have to leave a small gap between the start of the wrapping and the end, otherwise the shield will form a complete ring of the same diameter as the coil. This would give a very strong signal as the flux is totally threading this ring.
Aluminium tape, or copper tape will certainly work, provided it is thin enough so that cross sectional eddy currents are not developed. I prefer lead tape as it has a lower conductivity for a given thickness. Copper and lead also have the advantage that the shield grounding wire is readily soldered to it.
Why is shielding needed? The primary reason is to prevent the capacitance between the ground surface and the coil giving false signals. For an unshielded coil, this effect is particularly severe on a wet salt water beach. Touching the wet surface, seaweed, and on land wet grass, can all cause problems without the shield, particularly with sample delays that are less the 25uS. With a shield that is connected to the electronics ground, the capacitance that the coil sees is only that of the shield, and is therefore constant. PI detectors such as the Pulstar and Superscan that are designed for finding large objects at depth, do not necessarily need a shielded coil. This is because the minimum pulse delay is greater than 25uS where the effect becomes much less noticeable.
The second purpose of the shield is to attenuate r.f. interference from various broadcast and other transmitters. It doesnt get rid of it completely, as the shield is nowhere near as efficient as an aluminium enclosure, as it has to be thin enough to not cause attenuation of the wanted object signals. Also, low frequency r.f. signals from about 500kHz downwards, will not be attenuated to any degree. If they were, then you would start to lose sensitivity. Power line interference will also not be attenuated, and be just the same as for an unshielded coil.
A third purpose of the shield, which has a greater importance in these days of EMC, is to prevent any spurious emissions from the detector electronics causing interference with other electronic equipment. In the European Union, radiated emissions are measured from 30mHz upwards, to at least a 1GHz. With an unshielded coil a standard PI will likely fail this test, not because the transmitter itself is generating frequency components of this order, but other parts of the circuit, such as the clock generator, do have fast edges that leak into the coil circuit. I had the problem once where a 555 timer on the board was radiating sufficient energy via some over long pcb tracks on the output pin, to cause the detector to fail the test. A properly shielded coil and a well laid out pcb should have no trouble passing existing emissions tests, even when using a fairly high power pulse transmitter.
Eric."

This is a nice catalog of reasons which we can examine.

I think his points are good, but there is still a tendency to want to "have it both ways". We want our shields to let through our TX signals and RX signals, but also to block outside EMI.

I don't think you can have it both ways. I suspect that to transmit and receive signals, you also have to receive EMI of certain types.

He talks about two main characteristics of the shields that determine how they affect signals:

1. Shield geometry being parallel to signal "lines of force". (I think he means parallel to the "magnetic field lines", which are not what I think of as lines of force).

2. Frequency of the signals we are affecting. Blocked EMI is much higher frequency than the TX frequency.

Regarding (1), I suspect that if the shield's geometry does not block our TX signal, then it will not block EMI of a similar electromagnetic field orientation, and vice versa.

Regarding (2), the frequency argument (shields block much higher frequencies) sounds valid, but our tuned coils and MD electronics may block those frequencies anyway to a very high degree. Of course VLF and PI coils are tuned very differently, so his statements seem more accurate for PI coils.

Similarly, a tuned coil would not be expected to radiate any significant high frequency signal. Even a PI coil probably "chokes" high frequencies quite a bit. His radiation problem may have been directly from a circuit board trace, not the coil.

His argument about creating a constant capacitance to ground seems convincing, as different couplings to ground would probably modulate the TX frequency and amplitude somewhat. It is most convincing for the case of actually touching objects with the coils. Of course a Synchronous Detector is supposed to be somewhat impervious to phase/frequency changes in the TX signal, but clearly MDs are super-sensitive detectors, so it makes sense to suppress any disturbing influences.

If we are really serious about expecting shields to block EMI, then they should be highly conductive, because that is how shielding works -- induced currents canceling the electro-magnetic field of the EMI. However, if the EMI has a similar polarization to our TX signal, I wouldn't expect our shield to block it -- because the gap in our shield will prevent the needed current to block the EMI; except if the EMI is very high frequency, in which the gap is like a small capacitor and we get the current we need. But again, is high-frequency EMI ever a real problem?

Because high-resistance shields (graphite, etc) seem to be actually used in commercial coils, it implies to me that EMI blocking is not the real objective of coil shields. But perhaps high-resistance shields are only useful to one type of MD (PI or VLF).

It is probably advisable to divide shield discussions into VLF shields and PI shields until it is proven that both technologies benefit the same from any shield.

-SB

[DOOLEY]

hello vortxrex ,

not sure if lead paint is still available ,anywhere , something about it killing people , but i like your train of thought .

not sure that lead paint was ever conductive enough , amount of lead content was quite low .

[Vortxrex]

If Lead works so well, why not use Lead Paint?

[maikl]

Link to Kev has set in post # 49 http://www.findmall.com/read.php?34,...370#msg-134370 provides answers for this topic.
Quote from Eric Foster:
"Shielding is essential for PI detectors if they are to have good sensitivity for small metal targets. Careful choice of the material for the shield, and also the positioning of the shield with respect to the coil, will result in no loss of sensitivity, in fact I cant think of any negative effects that proper shielding has.
As has been said, if the shield is parallel to the lines of force generated during the transmit period, and the material is thin enough, then there will be no measurable eddy currents induced in the shield. The other important factor is the shield materials conductivity. The lower this is, the faster any eddy currents will decay, and provided they decay before the sampling period of the detector, then they wont be seen.
A flat sheet of shielding, either above or below the coil, is not the best, although it is often used for convenience. A field plot of the coil will show that only immediately under the winding is the field parallel to the shield. Most of the shield therefore has eddy currents generated in it and it has to be made of a relatively poorly conducting material. This is why graphite and nickel paints are often used here. Again, it depends on how thick a coat you put on as to the decay time of these currents.
If you wrap the coil with a metallic tape, then all points on the tape are parallel to the field. The field being coaxial with the coil cross section when close to it. You can then use a more conductive material, and as it is completely enclosing the coil, the shielding efficiency is much greater. You do, of course have to leave a small gap between the start of the wrapping and the end, otherwise the shield will form a complete ring of the same diameter as the coil. This would give a very strong signal as the flux is totally threading this ring.
Aluminium tape, or copper tape will certainly work, provided it is thin enough so that cross sectional eddy currents are not developed. I prefer lead tape as it has a lower conductivity for a given thickness. Copper and lead also have the advantage that the shield grounding wire is readily soldered to it.
Why is shielding needed? The primary reason is to prevent the capacitance between the ground surface and the coil giving false signals. For an unshielded coil, this effect is particularly severe on a wet salt water beach. Touching the wet surface, seaweed, and on land wet grass, can all cause problems without the shield, particularly with sample delays that are less the 25uS. With a shield that is connected to the electronics ground, the capacitance that the coil sees is only that of the shield, and is therefore constant. PI detectors such as the Pulstar and Superscan that are designed for finding large objects at depth, do not necessarily need a shielded coil. This is because the minimum pulse delay is greater than 25uS where the effect becomes much less noticeable.
The second purpose of the shield is to attenuate r.f. interference from various broadcast and other transmitters. It doesnt get rid of it completely, as the shield is nowhere near as efficient as an aluminium enclosure, as it has to be thin enough to not cause attenuation of the wanted object signals. Also, low frequency r.f. signals from about 500kHz downwards, will not be attenuated to any degree. If they were, then you would start to lose sensitivity. Power line interference will also not be attenuated, and be just the same as for an unshielded coil.
A third purpose of the shield, which has a greater importance in these days of EMC, is to prevent any spurious emissions from the detector electronics causing interference with other electronic equipment. In the European Union, radiated emissions are measured from 30mHz upwards, to at least a 1GHz. With an unshielded coil a standard PI will likely fail this test, not because the transmitter itself is generating frequency components of this order, but other parts of the circuit, such as the clock generator, do have fast edges that leak into the coil circuit. I had the problem once where a 555 timer on the board was radiating sufficient energy via some over long pcb tracks on the output pin, to cause the detector to fail the test. A properly shielded coil and a well laid out pcb should have no trouble passing existing emissions tests, even when using a fairly high power pulse transmitter.
Eric."

[DOOLEY]

hello all ,

sb , the combination of all the testing of different materials and on tx/rx or both would take a long time , but i'm sure among the collective on here most things have been tried , so we could simply post all our own views and go from there.thrn poss edit into a usefull list.

incidentaly , un-shielded PI coil , based on own experiance isn't affected that much from the ground , my signal goes slightly lower ( quieter ) as i aproach the ground , but it all depends on sample delay ,

long sample delay ( 30-50 uS) minor reduction in audio "click" / threshold setting

short delay (below 20 uS) big reduction in audio "click" / threshold setting ,slow click (5/sec) goes dead.

[Comp]

I know I don't fully understand the effects of our shields... especially since there seem to be quite different types of shields, from the high-conductive foils to the lower conductive sprays and graphite mixtures. So anything that sheds light is interesting.

Of course, even engineers may not really know all that a shield is doing -- the bottom line is someone noticed that things go better with a shield. I'd like to see actual situations (different soils, EMI environments, humidity, soil surfaces, etc.) demonstrating an unshielded coil vs a shielded coil, for each type of shield material.

And then there are other variations -- such as just shielding the RX coil and just shielding the TX coil -- how would that compare too?

Also variations in how you ground the shield and coil leads -- dfbowers noted some interesting variations with the TGSL coils.

All these variations are part of shielding.

Regards,

-SB

My take is the purpose of the shield is twofold,........ A, to screen the coil from random electronic signals, (Radio waves, static noise, E.T. {LOL} ) which could cause a 'false
' signal, and B, to place an 'earthed screen', in between the coil/winding, and ground,... because as you swing the search coil near the deck, variations in ground capacitance will again, cause false signalling,...... {I personally suspect this is less of a problem with PI,.. but dont bet on it, I'm a newbie!!!}

Of course, if the Guru's can enlighten me otherwise with their wisdom, I will happily stand corrected.......

Cheers, Fred

[simonbaker]

I am relatively new to this Coil section of the forum so have not read much of what is contained here. So forgive if this has been discussed to death already but it sounds like
there is continuing discussion of the actual purpose of shielding??
I was told 25 years ago by an engineer in Forest Grove that helped with designing my pinpointer coil that the purpose of the sprayed on shield was to dissipate Static Electrity generated in the plastic from rubbing it against things. The use of Graphite prevented eddy currents from forming in the shield. At least that was the intent of the shields from the 1980's.
I suspect you guys are talking of something else though huh?
Phil M

I know I don't fully understand the effects of our shields... especially since there seem to be quite different types of shields, from the high-conductive foils to the lower conductive sprays and graphite mixtures. So anything that sheds light is interesting.

Of course, even engineers may not really know all that a shield is doing -- the bottom line is someone noticed that things go better with a shield. I'd like to see actual situations (different soils, EMI environments, humidity, soil surfaces, etc.) demonstrating an unshielded coil vs a shielded coil, for each type of shield material.

And then there are other variations -- such as just shielding the RX coil and just shielding the TX coil -- how would that compare too?

Also variations in how you ground the shield and coil leads -- dfbowers noted some interesting variations with the TGSL coils.

All these variations are part of shielding.

Regards,

-SB

[turtlebowl]

I am relatively new to this Coil section of the forum so have not read much of what is contained here. So forgive if this has been discussed to death already but it sounds like
there is continuing discussion of the actual purpose of shielding??
I was told 25 years ago by an engineer in Forest Grove that helped with designing my pinpointer coil that the purpose of the sprayed on shield was to dissipate Static Electrity generated in the plastic from rubbing it against things. The use of Graphite prevented eddy currents from forming in the shield. At least that was the intent of the shields from the 1980's.
I suspect you guys are talking of something else though huh?
Phil M

[turtlebowl]

Found this at Home Depoe. A product made by Blaster and cost about $7 per can. It is quick, uniform and works really nice. I found it in the key making area.

http://www.blasterchemical.com/graphite-dry-lube.html

Ohm value very much depends on how thick the application sprayed on. Single medium coat reads about 2-3K, thin coats up to about 8K per square inch. 3 coats about 600 ohm.
I did not buy a can but happened to have a ohm meter in my pocket and the rep let me try the spray in the store. I sprayed a paper sack and brought it home for more testing.
I think one would need to spray a sealer over everything when finished as the graphite will rub off with medium abrasion.
hope this helps someone. Please let us know of any further experimenting with this product.
Phil M

[simonbaker]

You said it: flux changes.

The TX and RX, are of magnetic nature. Changes in the magnetic flux (B field?) of the TX and RX coils.

We want to shield against EMI (E field?) that is generated outside the metal detector, often far away, although I found a good test for the shield to be the EMI that the hand emits.

I looked at the FFT of the EMI that my hand introduced into the un-shielded coil, at the output of the preamp. It was of a very wide spectrum, with the upper limit rolling off above the coil's self resonant frequency (300kHz).

Then I tinkered with the shielding, until I could not see any "hand noise" anymore. Good enough.

An exact definition of coil shielding is related to the sensitivity of the coil. A thumb rule would be, the detector should not pick up the shielding if presented to the coil.

Once I got the detector sensitive enough to detect gold leaf (1/250,000 of an inch thickness), at several inches distance, I found that the shielding with graphite compound was still working fine. Why?

The shield should not sustain eddy currents.

Tinkerer

The physics of shielding is complicated to me, but your experiments indicate it works to reduce EMI (although for non-PI MDs the Synchronous Detector may effectively kill EMI more than a few hundred Hertz from the TX frequency because it is an extremely sharp filter), so I will believe they work, I'm just not sure to what extent and how.

In E&M theory, there really is no way to separate changing magnetic and electric fields -- one always goes with the other. You can't just have a changing magnetic field. However, there are some "near field" magnetic field components that are non-radiative and perhaps that is what we rely on with our MD coils, and so we shield against the radiative components that are small from our TX coil but large from EMI due to frequency differences. I think conductive shields work more effectively as the frequency increases -- to a point. After that point, they don't work well I think. So I would think maybe for non-PI MDs, our shields improve the S/N ratio for high frequency noise (> 20 kHz), but maybe they make it worse for low frequency noise (< 10 kHz, e.g. 60 Hz, 50 Hz) because they block our TX signal more effectively than the noise????

The way a conductive material shields against electromagnetic waves requires precise currents in the shields. I guess by "eddy currents" you mean extra currents that slop over that are not precisely the ones that cancel the field. So yes, we don't want eddy currents, but we do want some currents if we want to cancel EMI. If we just want to dissipate static charge, I would think we don't need much conductivity.

I'm just thinking out loud, I don't have a good model in my head for how shields work, especially ones that wrap around the wire bundle as opposed to a "box". So I'll rely on experiments such as yours for guidance for now. Also, I tend to think in terms of non-PI MDs, and I think you're focusing on shielding PI coils, so maybe some differences there.

Regards,

-SB

[Tinkerer]

Hi Tinkerer,
With Minelabs you can only shield to eliminate HF EMI and not LF or else the detector would become dumb, and no better than a home brew machine, throw away US$5K.

It is the capacitance of other objects especially the ground that must be completely eliminated. Eric says that if you do this properly then sensitivity is actually improved contrary to how we would think. Your hand would've been introducing a capacitive load to the coil and the EMI was using this to migrate to the coil.

Seems to me that a Faraday shield is specific to each detector and search head design and there are no set values other than rules that work for all. What is good for one design may dumb down another, hence the need to experiment with your own setup, as you cleverly did to arrive at the best compromise.

Cheers
Kev.

There is an important factor that I forgot to mention. The coil to shield capacitance. This can slow down a coil dramatically and make it useless for small targets.

A spacer of 4mm to 6mm between the coil winding and the shield helps a lot.

Tinkerer

[Kev]

You said it: flux changes.

Then I tinkered with the shielding, until I could not see any "hand noise" anymore. Good enough.

Tinkerer

Hi Tinkerer,
With Minelabs you can only shield to eliminate HF EMI and not LF or else the detector would become dumb, and no better than a home brew machine, throw away US$5K.

It is the capacitance of other objects especially the ground that must be completely eliminated. Eric says that if you do this properly then sensitivity is actually improved contrary to how we would think. Your hand would've been introducing a capacitive load to the coil and the EMI was using this to migrate to the coil.

Seems to me that a Faraday shield is specific to each detector and search head design and there are no set values other than rules that work for all. What is good for one design may dumb down another, hence the need to experiment with your own setup, as you cleverly did to arrive at the best compromise.

Cheers
Kev.

[Comp]

I'm currently (!!) playing with this tape,........ Seems Ok, and a doddle to solder to!!

http://www.ebay.co.uk/itm/Copper-Foil-Tape-Tiffany-stained-glass-work-1-roll-/270867018673?_trksid=p4340.m444&_trkparms=algo%3DCRX%26its%3DC%252BS%26itu%3DSI%252BUA%252BLM%252BLA%26otn%3D5%26ps%3D63%26clkid%3D4982322097831220363

Cheers, Fred

[Tinkerer]

I don't really know what shields actually do for metal detectors. EMI blocking doesn't make sense to me because it seems it would block the TX/RX signal as much as the EMI and not improve the signal to noise ratio. Also, it is not really shielding the coil loop as a whole so flux changes through the loop still get detected. The static discharge purpose seems to make some sense, but then why need such low resistance? Then there is something about ground capacitance???? I'd like to see some actual experiments with a variety of shields or some theory on what they do.

Certainly worth experimenting with and not accepting conventional wisdom, unless there is a ton of data out there we can look at.

-SB

You said it: flux changes.

The TX and RX, are of magnetic nature. Changes in the magnetic flux (B field?) of the TX and RX coils.

We want to shield against EMI (E field?) that is generated outside the metal detector, often far away, although I found a good test for the shield to be the EMI that the hand emits.

I looked at the FFT of the EMI that my hand introduced into the un-shielded coil, at the output of the preamp. It was of a very wide spectrum, with the upper limit rolling off above the coil's self resonant frequency (300kHz).

Then I tinkered with the shielding, until I could not see any "hand noise" anymore. Good enough.

An exact definition of coil shielding is related to the sensitivity of the coil. A thumb rule would be, the detector should not pick up the shielding if presented to the coil.

Once I got the detector sensitive enough to detect gold leaf (1/250,000 of an inch thickness), at several inches distance, I found that the shielding with graphite compound was still working fine. Why?

The shield should not sustain eddy currents.

Tinkerer