Shielding, and more...

[Anonymous]

I am building my own PI detector. I have a few questions if anyone knows the answers? much appreciated!
1) Shielding of the coil! why is it so critical? I get drift and interference if its not very even. I am using nickel shielding spray, any better suggestions? Can I make it less critical?
2) Coil windings. what difference does the winding neatness make? what about the shape or footprint? What limits the shape, turns, wire grade, etc?
3) I have tried Litz wire and got even more drift and noise then the usual wire?! What am I doing wrong with it?
4) Interference! It comes and goes but how do I know where its coming from? How do I detect its started? I usually have a couple of identical circuits running and when the same disturbances can be seen on both I make a huge assumption that its external interference! How do I make this more scientific?
5) I need to detect all metals, but generally alluminium is the main target, can anyone tell me a suitable coin or whatever, and a suitable detection distance this coin can be detected at? If I detect a martian shackle at 4" is this good detection distance? Ive got no way of knowing!
Thanks alot folks, sorry theres so much, but I am busting to understand these things.
Roy

 

[Anonymous]

Hi,
Great questions that will take a lot to answer, and I am interested in the answers also. But first, being a city boy who lives out west, I got to ask, what is a martain shackle?
Things that come to mind are, is this one that is used on the little green men or one for the large robot. All, kidding aside, what is the martian shackle, what is it made of, how large is it, and finally, what is the general shape?
Reg

 

[Anonymous]

Point number one. Have you grounded the shield to the ground side of the coil? An ungrounded shield is worse that no shield at all. I'm not sure why you get drift. If this is the case then it is likely that something else in the electronics need attention.
Point 2. PI coils are pretty tolerant in the way they are wound. Unless you are going for super small object performance where you need to keep the winding capacitance down, being overly neat is not essential. Make the shape and size whatever you want. Usually one tries to keep the inductance and resistance about the same.
Point 3. I've never used Litz, but there is no reason I can think of as to why it should be noisier.
Point 4. External interference can be nulled out by rotating the coil into a different plane. A scope on the first amplifier will show this, and also the frequency of the noise, which may be 50/60Hz or low frequency radio.
Point 5. You need to specify what size and thickness a typical target would be. Aluminium varies dramatically in conductivity according to its alloy.
Like Reg, would like to know more about Martian shackles.
Eric.

 

[Anonymous]

Hi Again Roy,
First of all, I am just an experimenter like you so there are more qualified people than me who frequent this forum who can give better answers. It is from them that I have learned the most. However, experimenting has really helped, so I will try to provide some basic answers are reasonably accurate.
1. Shielding, or Faraday shielding as it is called, is used to reduce unwanted noise, whether it be external or self generated. Ideally, the shielding should be sufficient to eliminate or reduce the electrostatic field or electrostatic buildup. The material used should be able to do this without introducing more problems, like reduced sensitivity or lengthening the minimum delay capabilities. The amount of shielding, the type of material used, and how it is applied is also critical. Too much shielding is just as bad if not worse than not enough. Also, the shielding normally wraps around the windings or covers the top and bottom, but does not totally wrap the entire coil circumference. Instead it has a very small air gap between ends. Also, the shielding should not be detectable, or barely detectable by the PI.
As for drift, I fully agree with Eric, some part of the electronic circuitry is drifting, or if you are building one of the more common kit type PI's, the powersupply is probably not stable.
I have used Lead foil, a super thin copper mesh, semiconducting tape, and even the very thin foil backing off sheets of foam insulation, all with good success as shielding material. I try to have a space between the shielding and the windings, and make sure the distance between the windings and the shielding cannot change.
As Eric noted, the shielding has to be connected to the coil lead that is not connected to the FET or whatever transistor is used.
2. As for coil windings, the circuitry, the desired minimum delay, and the maximum current drawn will mainly determine the size of wire used and the number of turns.
As for footprint or size of the coil, they will be basically determined by what you want to detect. The bigger the coil, the bigger the object has to be for detection, and the area you want to cover.
For example, I am using one of Eric designed detectors I have modified. I use a 9" by 13" coil form, Litz wire equivalent to about a 28 awg and has 26 turns. This gives me a measured inductance of about 530 uh with shielding applied. On this coil, I am using semiconducing tape with some mods.
The detector operates fine at 10 usec minimum delay.
3. I have found Litz wire seems to reduce the capacitance which allows for a shorter delay. Just changing to the Litz wire I was able to shorten the delay by 2 1/2 usec. Again, the Litz wire size makes a difference. In my case I am using a 7/36 size, which is 7 strands of #36 wire.
4. Most interference is externally generated and I have found that it varies throughout the day. However, on the PI's that I built from plans, the powersupply limitations also introduced noise. As for the external noise, normally changing the pulse rate will work to reduce this noise. By making the pulse rate variable, then one can normally find a rate where the noise is minimal. Generally, I have found that it doesn't take much of a change in the rate to make a big difference in the noise level.
5. As for detection range and target distance, again this is a function of coil size, the minimum delay, and the material being detected.
In my case, I can detect a 5 grain gold nugget with a decent signal at a distance of 3 to 5 inches, depending upon the orientation of the nugget. I can detect a nickel with a nice strong signal at 13" and a very weak signal somewhere between 15" and 17" depending upon the external noise level. However, I have now idea how effective it would be on your martian shackle. Heck, I am not even sure if Martians are detectable with a PI.
Hope this helps a little,
Reg

 

[Anonymous]

Thanks to Reg and Eric! Can I just ask where the info on coils and stuff comes from? I have been building my own detectors for a few years and I have searched hi and lo for practical info like this and so far my sum total is zero! (I only discovered Findmall the other day?!?)Are there any books on this kind of stuff?(a book? oh joy!)
I have been playing around with different coils. I have had pretty good success with a kind of figure 8 shape recieve coil, signal goes + or - depending on which side target is. Transmit coil is plain round shape. Excellent pin pointing of targets. However the shielding is very critical. I use low capacitance wire as well. With Litz I seem to get erratic signals(clumsy to work with also!). My shielding forms a U around the sides and top of the coil, and around the circumference I have found little difference when I break the shielding loop. The recieve coil outputs feed into sort of instrument amp, 10usec delay then sampled. This means that I do not effectively have an earth side of the recieve coil, is this my problem? The recieve coil and circuitry is clamped at 0v until 5usec after transmit pulse. How would one keep the winding capacitance down?
Truly thankful to you all.
Roy
Regarding the Martian Shackle. Haven't you found one yet? We can't move for the blasted things round here! The thing to remember is they are ONLY detectable at 4 inches! I once made a machine that would cancel out all other targets! When it went quiet I would dig. Got a false alarm nearly every time!

 

[Anonymous]

Hi Roy,
I got most of my information from this website and experimentation. However the Geotechnology website, see below, has some good basic information and plans on kit type PI's.
Eric has been a tremendous help in supplying information on this site, as have other very knowledgeable people.
I attribute much of my knowledge on VLF's and some on coil design to George Payne, who is, in my opinion the "father" of the VLF motion detector. Now, we have Dave Johnson, another extremely knowldegeable VLF designer monitoring this site who also has been an active contributor to this forum.
I have never really found a good book on the subject of PI's. Maybe we can talk Eric into writing one. I know I would buy a copy. In the meantime, I just copy inportant information as it is posted, and keep experimenting to see what works or seems to work best. However, I still have a long way to go on the learning curve.
Now, as to your coil, having a separate receive winding is a little different than a simple mono coil. However, one side of the receive winding feeds the preamp and the other lead has to connect to something. I would suspect it is the common or ground connection. This is normally the center point between the + and - power supplies for the opamps.
Since you are using an instrumentation type opamp, are you differentially connecting the coil to both the + and - inputs? If so, how do you reference this to a common point. Do you centertap the receive coil, and tie the centertap to the common? If so, then I would think this would minimize both the noise and earth field problem. At least that is my first thoughts.
I am curious about what happens when you use the figure 8 coil? Is the noise level significantly lower? My experimentation with this type of coil showed that was the case.
Do you have a lot of noise with the coil disconnected? If you still have noise with the coil disconnected or when using a figure 8 coil, then most likely the noise is internally generated. I haven't done nearly as much experimenting with IC's for preamps as Eric has, but I have found there are signifiant differences in noise levels between them. So far, the basic NE 5534A seems to work the best for me. Again, I have to thank Eric and other people on this forum for that advice.
As for a faraday shielding, I would recommend you reverse your shielding so the bottom is definitely shielded. This is the critical area to shield. I prefer to make sure the shielding is wrapped completely around the windings. I have found that shorting the Faraday shield doesn't seem to have that dramatic of an effect, but I was just told that would indicate the shielding is insufficient. However, the level of shielding I have, completely eliminates any hand capacitance and I don't see any difference between what I am using now and the lead foil shielding. However, I haven't done any real serious experiments either. I just know I have minimum ground pickup and minimum noise.
As for keeping the winding capacitance down, I use Litz wire. this seems to work the best for me. Dave Emery just recommended to me to space the windings a little so they are stacked and not layered as much. This will change the inductance also.
Now, as for the Martian Shackle situation. I suspect they are not found in extremely dry climates is why I have never encountered one. Probably don't like forest fires either, and of course, we have our share of them here in Colorado. I heard that Smokey the Bear wasn't happy when he found out who started the big fire we have here now.
If I am wrong about Martian Shackles being sparce in hot dry climates, then maybe we can find charred Martian Shackles once the smoke clears. That is if they are detectable then. Who knows, they may mutate and become non-detectable because of the heat.
Reg

 

[Anonymous]

Hi Roy,
Just keep tuned in to this forum for info on PI coils and electronics. I hope to get down to writing a comprehensive book before too long, but at the moment, there is not much "in depth" info that is published.
Your round TX with a figure of 8 RX is something I have started to look at. It has good noise cancelling ability, and ground cancelling too. Only thing is that the signal goes to a null when crossing between the two RX halves. There is a PI mine detector made by Guartel in the UK, that successfully uses this technique.
You can use a differential input arrangement for the RX input, but the faraday shield still needs to be grounded. I would use a shielded twin cable rather than a coax, with the cable shield connected to the faraday shield at the coil end. Is this what you have? I have tried centre tapping the coil to ground but did not notice any difference.
Eric.

 

[Anonymous]

The best explanation I can find is for a shackle to hold a 22in torpedo underneath the Martin B26 Marauder. The original name for this plane was Martian, but it was decided that Marauder sounded better. Only 4in though? Must have been a funny metal.
Eric.

 

[Anonymous]

First I apologise for not scouring the message board before asking a question, but there is so much of it, its a bit daunting (free time? its not cheap!). Is there any way of searching for keywords among the messages? You may think this is lazy but I like to call it efficiency...And its also too distracting, I look for 1 solution and end up with another 10 problems!
Anyway thanks for the advice, I shall now go and do an experiment frenzy! I will go and try shielding underneath as well. It makes sense to try this as this would explain the critical shielding up the sides. What about the Tx coil, does this need shielding? I will go and try a few combinations.
The Rx coil ends are connected to opamps and then converted to single +- signal. If the Fig 8 'lobes' are symetrical, it doesn't even see the transmit pulse. Its the imbalance that is detected, ie target.
I have found the drift to be caused by the Tx transistor getting warm. Any ideas for electronically compensating? No room for heatsink either.
I have found that by increasing the pulse power considerably, does not result in a substantial increase in depth! Am I reaching a limit this system? Is higher voltage better? Or longer pulse length(if limited to a max voltage)?
Many thanks, and more!
I thought that too much heat makes a Martian Shackle transparent. Wack 12 volts across one and see the little feller dance! Ends up practically invisible apparently!

 

[Anonymous]

Hi Roy,
It is best to shield both coils. The primary purpose of the shield is to prevent capacitance change when the coil is brought near to the ground. The coil sees only the fixed capacitance to the shield and not the varying capacitance to the ground surface. The most important shielding, therefore, is that underneath the coil. As a secondary benefit, overall shielding also attenuates high rf frequencies. It is important that the coil does not move relative to the shield or you will get a false signal each time the coil bangs against a stone.
You are right about the TX transistor causing drift. If you are using a Mosfet, they have a positive temperature coefficient. This means that the TX current will drop slightly as the device warms up. There are always some spurious eddy current signals that are generated in the connectors, coil components etc., and these will reduce slightly on warm up, giving the resultant drift. Luckily the drift is negative, which is the opposite way to a wanted signal, and can be compensated for. I use a system where, if the output of the final dc amplifier goes negative by more than 15mV, the amplifier re-zeros.
A heat sink will help if you can squeeze one in, but another way to reduce drift at source is to use a constant current transmitter arrangement. There was a good discussion on this a few months back on this forum.
Doubling the pulse current will double the signal amplitude from an object, but the detection range will only increase by about 10%. To double the range you would need 64x the current. Then you would have some warm up drift to worry about. What pulse length and pulse rate are you using, and what is your peak coil current?
Eric.

 

[Anonymous]

Hi Eric
Have you ever noticed that when you require a tiny metal-free bit of ground to test on, they are rarer than Venetian shillings! I have been trying the search head upside down, meaning the top is now unshielded, and I didn't notice any major difference, although I must admit that this test lasted about 10 secs before rain stopped play! and my garden is full of metal, so I'm not sure I would have noticed anyway.
I think(my notes are elsewhere!)I can vary the pulse width in 5usec steps to 50usec, this defines the current. At 50usec pulse width V=12V I=0.1A and at 45usec V=12V I=0.086A etc. The transistor is a mosfet rated at 400v (memory tells me its a IR74XXX or something similar!). It is limited by a couple of 180v Zeners giving 360v.
I have noticed something peculiar, the fig8 coil detects stainless steel @10usec delay, that an ordinary round coil could not with exactly same everything. How can this be? I need to increase SS up to 100x before the round coil detects it at all!?! However the round coil gives an extra 40% depth on some other metals.
Last question, is there a standard target(official or not) that people use to compare one machine against another?
Thanks alot Eric & co! Ive learnt more in a week than I did the year before!

 

[Anonymous]

Hi Roy,
One of the tests I use to check my shielding is a simple hand technique. With the detector at minimum delay, just take your hand, less rings and watches of course, and move it towards the coil until it is up against the housing. Insufficient shielding will result in the detection of your hand. The worse the shielding, the greater the signal increase as your hand approaches the coil.
This works quite well. I started using this highly technical procedure many years ago when I was having fits with a couple of coils on a VLF.
Now, Venetian Shillings? I gotta ask, how many miles from the sun do you live? Rounded off in millions will be fine. hehehe
Reg

 

[Anonymous]

Hi Again,
I forgot to mention two quick things. First, one of the more popular targets is the good old US nickel. Isn't much good for anything else nowdays except as a reference target to see what kind of distance one can obtain in an air test.
The second item is, if you have planted aluminum targets (like you mentioned in your earlier post as the desired object you want to detect) in your garden to check the depth capabilities of your detector, you might be sprouting tinfoil. It grows in abundance everywhere and hard to get rid of.
Finally, that must be an interesting coil you have that your can mount it upside down also. How about inside out? hehehe
Reg

 

[Anonymous]

Hi Reg
I perform the very same test with mine. I have noticed that it is even worse if you spread the fingers out to touch both sides of the coil. Also someone else may perform the same test and it will be fine, then I do it and it goes bananas!?
Today I went to a golf course and tried out a new coil housing with all round shielding, against an old one with shielding on top and sides, none on bottom although the sides extend down maybe 8-10mm beyond the coil. To tell the truth, if there was a difference, it wasn't a huge amount. The nearer the coil to Earth, the more it suffers from capacitance variations(obvious, but bear with me), regardless of the shielding. I tried keeping the coils 1 inch off the ground and then tilting the coil left and right, and I think the all round shielded one was marginally better and overall it sounded off less often. Although I have seen similar differences just changing one connector for another! Verdict: erm...I'll have to try it a tad more scientifically, or at least with a ruler!
Lastly I am using a pre-amp to boost signal up wires to control board. Is this worth it? It seems to improve the earth connection to the shielding.
Many thanks and any idea's on the stainless steel response mentions earlier, it is really bugging me.
Regards
Roy

 

[Anonymous]

Hi Reg
I perform the very same test with mine. I have noticed that it is even worse if you spread the fingers out to touch both sides of the coil. Also someone else may perform the same test and it will be fine, then I do it and it goes bananas!?
Today I went to a golf course and tried out a new coil housing with all round shielding, against an old one with shielding on top and sides, none on bottom although the sides extend down maybe 8-10mm beyond the coil. To tell the truth, if there was a difference, it wasn't a huge amount. The nearer the coil to Earth, the more it suffers from capacitance variations(obvious, but bear with me), regardless of the shielding. I tried keeping the coils 1 inch off the ground and then tilting the coil left and right, and I think the all round shielded one was marginally better and overall it sounded off less often. Although I have seen similar differences just changing one connector for another! Verdict: erm...I'll have to try it a tad more scientifically, or at least with a ruler!
Lastly I am using a pre-amp to boost signal up wires to control board. Is this worth it? It seems to improve the earth connection to the shielding.
Many thanks and any idea's on the stainless steel response mentions earlier, it is really bugging me.
Regards
Roy
PS What depth would a nickel be found at or air distance? What is a nickel anyway? Because for the life of me I can't remember its centage, if thats the right word!

 

[Anonymous]

Hi Roy,
One thing I forgot to mention when I check my coils is I do wrap my hand around the coil. I also check completely around the circumference to make sure there the shielding is even and complete.
What I have found is I have gotten strange "boings" on a coil that I know is not fully shielded. Strangely, the "boing" response occurs when I bump the coil over certain weeds. This is something that I have experienced with a VLF also.
Now, I would like to duplicate your stainless steel response, but need a little more information. What was the stainless object, how big was it, and what was the general shape?
Also, are you taking two samples of your receive signal to eliminate the earth field effects?
What is the general size and shape of your two coils you used when you tested the stainless? Was the figure 8 about the same size? Finally, I am not sure of the design of your normal coil. Is this a concentric type coil, or more of a DD type.
Finally, was the delay the same on both tests?
Reg

 

[Anonymous]

Hi Roy,
I would guess that the signal you are getting from the ground is not due to capacitance, but due to iron minerals. Very few soils are completely free from these, and they give a decaying signal much like metal, but it is a magnetic decay rather than due to eddy currents.
Another guess, on the stainless steel response. As the figure of 8 RX and the circular TX are in an induction balance arrangement, the presence of the stainless steel is upsetting the balance. Even though you are sampling in the TX off time and the stainless steel eddy currents are too fast to detect, the resulting unbalance is causing spurious TX transients or ringing to appear in the sampling period. When the coils are well balanced, these would cancel and not be seen. When I was developing the PPD1 (A discriminating PI with a stacked concentric balanced coil), I noticed that the balance was quite sensitive to the poorer conductors.
Another 5 cents worth. As you are not familiar with the famous nickel, and have garden rather than a yard, are you in the UK? Nickels are better than Martian shackles as a test object. If your PI is performing well, you should get 12in plus in air with a 10in coil.
Eric.

 

[Anonymous]

Hi,
Having never ventured out of the US any distance, I never gave it a thought one might not be familiar with the American nickel, our 5 cent coin.
Eric, I think you are on to something about Roy. PI's, Shillings, rain, garden, yep, those are words are seldom used together where I live. As for rain, it is now becoming a rare commodity. Heck, it is so dry here, trees are leaning towards dog houses.
How about it Roy, do you live on the European side of the pond? If so, I will try to adjust my answers to accomodate the European accent.
Reg

 

[Anonymous]

Cor Blimey guv'nor, you've twigged me!
Over here it rains all the bloody time! Nothing too dramatic, just plain boring overcast nonsense rainy weather! Now you've started me off...
Just for the record, a decent minelab or what ever is probably

 

[Anonymous]

Hi Roy,
It is nice to know how to duplicate a strange condition like the one you experienced with the stainless steel. For me, knowing the conditions, combined with experimentation, and the little knowledge I have on the subject, helps me determine the why's of such a condition.
Once knowing the why's, can readily expand the concept even farther, thus maybe finding a totally different coil arrangement that has serious advantages for certain objects.
Eric's comments about the stacked coils caught my interest since he mentioned such a coil design was quite sensitive to "poorer conductors". Now, I have to find time to try this type of coil.
As for your questions, a nickel is basically made of 25% nickel and 75% copper and has a diameter of 21.21 mm. A penny now is good for only two things, one to weight down you pants pockets, and two, so companies can charge $27.99 for a product instead of $28.00. It makes the cost seem so much cheaper, doesn't it?
Actually, there is a third use for the penny. Sometimes you can use 10 of them to buy a penny's worth of candy.
If you don't mind, send a little of your rain our way. No, make that a lot of rain. Unfortunately, it is so hot over here now, it would probably just evaporate before it hit the ground.
Now, getting back to your coil. If you are using a figure 8 design to ground balance, I can see how you could get a ground response by tilting the coil. Have you tried a 3 coil arrangement?
Reg