Advanced PI coils

[authere]

Hi Mida

I've drawn a X section showing the shielding. As the shields are concentric with the coil x section, there is only minimal eddy current generated in the shield, unlike if you put the tape, or coating, flat on the coil shell with the result that flux links with it. This is the most effective way to shield a coil from the emi point of view, but not cost effective for large scale production. Both shields, in the dual shield coil, are taken to the coax braid, as is the ground end of the coil. For the bifilar centre tapped coil there would be a different arrangement which I will look at in due course. The shield needs a small gap in it so that it doesn't form a shorted ring, otherwise you will get serious eddy currents. I once damped a coil by putting the damping resistor across the shield gap instead of the winding. Seemed to work OK.

Depending on the current you want to pulse, the wire I have used may be a bit too thin, but of course, this method of shielding is for any coil. 1190 copper fabric tape is hard to get now, and lead is even harder as it is classed as a hazardous material. I still have some for my own experiments but there are no doubt materials that are as good. Beware nickel tape though, as being a magnetic material, you do get a small, but constant decay from it.

The preamp gain I used was x500. You would run into saturation on the later dc amplification stages long before the preamp caved in.

Eric.




[ATTACH]22524[/ATTACH]

Hi Eric,

Just for interest sake, the coil i have been mentioning with the tubular coil bundle didn't need any shielding at all to work directly under a power line, i think it worked because the "current" flowed around the coil bundle 3-4 times before exiting providing it's protection from EMI, i never tested this coil on noisy ground as i live on the beach front. This coil was a mongrel to wind as i had to hot melt glue it to a form and it would not work for a large coil as there would not be enought "turns" of wire to cover the form.....anyways as i said, for interest sake

Ron

[Ferric Toes]

As I learned here about the maghemite fields in OZ, they are pretty layered due to the bush fires etc. but otherwise the offending ground is quite homogeneous, and some angle against the ground surface will not spoil the coil's GB ability by much.

Some places in Oz may be like that, but in the Victorian goldfields it is a little different. There is one area where the strata are at a steep angle and in one sweep of the coil you can cross several thin bands of highly viscous material interspersed with bands of low viscosity. Now that's ground noise. Ground noise gets worse on a hot sunny day as rocks, gravel etc on the surface heat up and this alters the decay law slightly. The underlying material which is shaded doesn't change so you have a spread in the exponent which makes GB less accurate. At night or after rain the temperature gradients even out and the ground becomes quieter. Some rocks and areas have a remanent magnetisation (thought to be caused by lightning strikes). Imagine sweeping the coil over a bunch of small ferrite magnets. That's ground noise. Some areas have layers of hard ironstone on top of a red clay. Again the decays are slightly different and in areas of old diggings both can be exposed together at the surface. Bush fires do make a difference. Intense local heating converts magnetite to maghemite which enhances the viscosity. Around trees which are burnt you often get a stronger ground signal which is so localised it sounds like a large deep nugget if the GB is not adjusted correctly, or for auto GB, not keeping up with the changes. Prospectors sometimes dig these signals only to find they fade away as the hole gets bigger and all they have are bits of barbequed tree root.

I have experienced the above at first hand and measured samples of the ironstone on a Bartington MS2B susceptibility meter, and my own viscosity meter. There is nothing like it. The demining literature quotes a lot of tests done in the Hawaiian volcanic islands, Mozambique, Cambodia and other places and they give figures of susceptibilities and viscosity (frequency dependent susceptibility), but it's no exaggeration to say that in the Victorian goldfields you can multiply the figures by ten. I will look out some results and post them in due course. This is important because it shows what a detector coil can be up against in the worst conditions.

Western Australia is reputed to have more uniform ground, but I have never taked a detector there, or measured any ground samples.

Eric.

[Aziz]

Coil Software Upgrade Gonna be Complex

Hi all,

1. pity: no quick & dirty solution possible at the moment. I have to "arrange and adapt"(c)(r)(tm)*LOL* some code parts and this requires more coding & testing time.
2. pity: I can start the coding in the second week of Jan. 2013 or later.
3. pity: tomorrow till second week of Jan. 2013: won't have much time to contribute to the interesting thread.
But keep on posting all the good ideas, so no one in the well-known company can patent it anymore.

Cheers and a Happy New Year to you all.
Aziz

[Ferric Toes]

some kind of coaxial cable?
Eric, what material do you think that will be the best for this "type" of shielding, regardless the cost or availability?

It all comes down to conductivity. The woven copper fabric 25mm wide tape I favour has a resistance of 1 ohm per 300mm. It has an adhesive backing which keeps it in place and it is solderable with care and not too much heat. Another candidate which is often available on ebay is 3M Scotch 24 tinned copper mesh tape. This has a resistance of 5 ohms per 300mm, but has no adhesive backing. Plain adhesive backed 0.05mm copper tape is no good as the eddy current signal generated in it persists for about 30uS even with a gap. It's resistance per 300mm is <0.1 ohms. Basically, look for a solderable adhesive backed tape that has a dc resistance of between 1 and 5 ohms per 300mm.

A search of web sites that deal in EMC products may come up with some useful material which I don't know about. I have enough NOS to keep me going for a couple of years on development projects, so I haven't been looking.

Eric.

[Davor]

CAT5e UTP cable. Just don't forget to leave a slit in a shielding at some place to avoid short circuiting your coil.

Some coil arrangements that have been discussed are described as having the ability to GB. Surely the only thing that can happen is that the amplitude of the ground signal is reduced from that of a mono coil, but that electronic GB will still be necessary. The only situation where a good cancellation takes place is when the coil is perfectly horizontal over a perfectly flat and uniform ferrite mineralised area. As soon as the coil is tilted either way the GB is no more as the amplitudes are no longer equal.

There is no question about replacing the rig GB with a coil, but complementing it and improving the overall situation. As I learned here about the maghemite fields in OZ, they are pretty layered due to the bush fires etc. but otherwise the offending ground is quite homogeneous, and some angle against the ground surface will not spoil the coil's GB ability by much. My situation is a combination of red clay (homogeneous) and shallow sea (mostly homogeneous). A bit of slant will surely spoil GB a bit, but it will still improve things. I already made a wet salt ground balancing add-on project for my VLF so I'm not going there unprepared.

Right now I'm covered in paint and putty, but I'm also FINALLY equipped with a real 2 channel 20 MHz CRO, so my lab is now up to something.

[Bill512]

I've drawn a X section showing the shielding. As the shields are concentric with the coil x section, there is only minimal eddy current generated in the shield, unlike if you put the tape, or coating, flat on the coil shell with the result that flux links with it. This is the most effective way to shield a coil from the emi point of view, but not cost effective for large scale production. Both shields, in the dual shield coil, are taken to the coax braid, as is the ground end of the coil. For the bifilar centre tapped coil there would be a different arrangement which I will look at in due course. The shield needs a small gap in it so that it doesn't form a shorted ring, otherwise you will get serious eddy currents. I once damped a coil by putting the damping resistor across the shield gap instead of the winding. Seemed to work OK.

Depending on the current you want to pulse, the wire I have used may be a bit too thin, but of course, this method of shielding is for any coil. 1190 copper fabric tape is hard to get now, and lead is even harder as it is classed as a hazardous material. I still have some for my own experiments but there are no doubt materials that are as good. Beware nickel tape though, as being a magnetic material, you do get a small, but constant decay from it.
Eric.

[ATTACH]22524[/ATTACH]

some kind of coaxial cable?
Eric, what material do you think that will be the best for this "type" of shielding, regardless the cost or availability?

[Ferric Toes]

Hi Eric,

Thanks for all your very informative posts of late. What exactly do you mean by double shielding ? And what sort of pre-amp gain were you running when you say you were 'nowhere near' saturation? How much do you think you could run before it did become a problem? I seem to remember reading on here somewhere that more than 50 could become a problem in some ground.

Midas

Hi Midas,

I've drawn a X section showing the shielding. As the shields are concentric with the coil x section, there is only minimal eddy current generated in the shield, unlike if you put the tape, or coating, flat on the coil shell with the result that flux links with it. This is the most effective way to shield a coil from the emi point of view, but not cost effective for large scale production. Both shields, in the dual shield coil, are taken to the coax braid, as is the ground end of the coil. For the bifilar centre tapped coil there would be a different arrangement which I will look at in due course. The shield needs a small gap in it so that it doesn't form a shorted ring, otherwise you will get serious eddy currents. I once damped a coil by putting the damping resistor across the shield gap instead of the winding. Seemed to work OK.

Depending on the current you want to pulse, the wire I have used may be a bit too thin, but of course, this method of shielding is for any coil. 1190 copper fabric tape is hard to get now, and lead is even harder as it is classed as a hazardous material. I still have some for my own experiments but there are no doubt materials that are as good. Beware nickel tape though, as being a magnetic material, you do get a small, but constant decay from it.

The preamp gain I used was x500. You would run into saturation on the later dc amplification stages long before the preamp caved in.

Eric.

[Aziz]

The Dual Field Mono PI Coil vs. Mono Coil

This is the performance relation of the dual field mono PI coil vs. mono coil. Both coil parts (outer and inner) have the same loop turns count but the total same inductance as the 10 inch mono coil (our reference coil). The inner coil has the half of the outer diameter TX coil. BTW, TX=RX (=300 µH), it's a mono PI coil.

It is definitely a pin-pointing coil.

Coil Comparison (Calculated Response Ratio to Reference Coil)
© 2012 by Aziz Ögüt. All Rights Reserved. 29-Dec-2012
Reference Coil: 300 µH Round Mono Loop Coil
		10" Dual Field Mono PI Coil (0.5)
Target Depth [inch]	Target Depth [cm]	Relation to Reference
1	2,5	2,421
2	5,1	2,778
3	7,6	2,046
4	10,2	1,622
5	12,7	1,400
6	15,2	1,273
7	17,8	1,195
8	20,3	1,144
9	22,9	1,109
10	25,4	1,084
11	27,9	1,065
12	30,5	1,051
13	33,0	1,040
14	35,6	1,032
15	38,1	1,025
16	40,6	1,019
17	43,2	1,014
18	45,7	1,010
19	48,3	1,007
20	50,8	1,004
21	53,3	1,002
22	55,9	1,000
23	58,4	0,998
24	61,0	0,996
25	63,5	0,995
26	66,0	0,994
27	68,6	0,992
28	71,1	0,991
29	73,7	0,991
30	76,2	0,990

Cheers,
Aziz

[Tinkerer]

Some coil arrangements that have been discussed are described as having the ability to GB. Surely the only thing that can happen is that the amplitude of the ground signal is reduced from that of a mono coil, but that electronic GB will still be necessary. The only situation where a good cancellation takes place is when the coil is perfectly horizontal over a perfectly flat and uniform ferrite mineralised area. As soon as the coil is tilted either way the GB is no more as the amplitudes are no longer equal. Also, most operators swing in an arc to some degree. The GB must be a system that works on the t^-1 decay and be independant of amplitude. Since IMHO you have to do this, you might just as well stick with a mono coil or a coplanar concentric TX plus separate RX. The humble mono can be improved as we have seen by bifilar winding, differential preamp etc, and maybe double shielding. I have been making tests on a double shielded 12in mono and I can sample at 6.5uS with a non differential LME49990 preamp with input diodes. The double shielding, I hope, gives more like a 2nd order filter to rf frequencies. Certainly it is much less noisy and I can pick up more stations if I use it as a radio antenna. A bifilar mono with dual shielding should be even better. Maybe a 2.5 order filter .

One interesting thing I noted when I was doing tests in Australia, was that for mono coils sitting on the ground surface the signal level at the output of the integrator (no GB) was the same for any coil diameter. This makes sense when you think about it. Also, the level was nowhere near saturating the preamplifier as some have thought in the past. This was on one of the worst ironstone areas in Victoria.

Eric.

My idea of using the Tandem coil configuration, is to measure the gradient only between the 2 coils. Unfortunately, this brings up the problem of sensitivity to any deviation from the parallel plane of the ground. However, there are circumstances where this is possible, for example with a underwater towed array or a wheel mounted array on a salt lake etc.
Just ideas for now, not tried yet, but would appreciate any feedback.

[Tinkerer]

Thanks Tinkerer,

I'm thinking of of another coil software upgrade to handle such coils too. Unfortunately, the target response is highly dependent on its orientation and the current software isn't capable to rotate it in it's maximum response orientation. Its currently a fixed orientated target position sweep.

I'll look and see, whether I can make a quick & dirty upgrade. Although, I could calculate & analyse the two-box coils too.

But one trivial coil is simple to do:
The dual-field PI coil

Cheers,
Aziz

Thanks Aziz,

The dual field PI coil will be interesting to see.

Intuitively, the Tandem coil would have to be 2x10" to attain the same depth. However, the depth is not everything. One of the problems with FE discrimination, is the very thin, rusty fragments of steel sheet. In the flat orientation, often the X response is less than the R response and then the fragment is not recognized as FE.

The response of these fragments is highly susceptible to orientation. When presented vertical to the coil, such that the field lines are parallel to the surface of the target, the FE recognition is very easy.

Having a detection field with a field line orientation that has a different angle, at a known distance from the coil, would help discrimination. It would also help with coins and nails, in short, any type of target that is highly susceptible to orientation.

[Aziz]

Here are the Tandem coils, same phase and opposing phase

Note the distance of the coils is not the same for the 2 pictures

Thanks Tinkerer,

I'm thinking of of another coil software upgrade to handle such coils too. Unfortunately, the target response is highly dependent on its orientation and the current software isn't capable to rotate it in it's maximum response orientation. Its currently a fixed orientated target position sweep.

I'll look and see, whether I can make a quick & dirty upgrade. Although, I could calculate & analyse the two-box coils too.

But one trivial coil is simple to do:
The dual-field PI coil

Cheers,
Aziz

[Aziz]

Yes, I try to make the inductance as close to 300uH as I can for any coil size, also I use a constant current source for the TX supply so that any changes in coil/cable resistance is compensated for as well.

Eric.

When looking at SNR, the SNR (the N(RX)*A(RX) product) gets worse with larger mono coils due to picking up more EMI noise. That's the drawback with large mono coils.

But we can split the large mono coil of course:
Large TX, seperate small concentric co-planar RX.
The magnetic flux linkage gets better too (reduced) to reduce the overall ground noise.

Cheers,
Aziz

[Ferric Toes]

Hi Eric,

yes, my coil software is making the same observation too. The total magnetic flux doesn't change (much) when the all sized mono coils are designed to a specific inductivity (300 µH).

Aziz

Yes, I try to make the inductance as close to 300uH as I can for any coil size, also I use a constant current source for the TX supply so that any changes in coil/cable resistance is compensated for as well.

Eric.

[Aziz]

...
One interesting thing I noted when I was doing tests in Australia, was that for mono coils sitting on the ground surface the signal level at the output of the integrator (no GB) was the same for any coil diameter. This makes sense when you think about it. Also, the level was nowhere near saturating the preamplifier as some have thought in the past. This was on one of the worst ironstone areas in Victoria.

Hi Eric,

yes, my coil software is making the same observation too. The total magnetic flux doesn't change (much) when the all sized mono coils are designed to a specific inductivity (300 µH).

Aziz

[Aziz]

DD PI AI Figure-8 Coil Added

Hi all,

the DD PI AI Figure-8 coil (TX=RX, formed as back to back DD figure-8, 300 µH) doesn't go deep even with additional gain of 10x. Forget this coil once and forever if you want to chase deep targets.



Cheers,
Aziz