Advanced PI coils

[Aziz]

Hi Aziz,

I played with coils a while back and came up with an anti interference coil of sorts,but i never really tested it fully and never had a dc ohm and inductance meter,but directly under power lines it was as quiet as a mouse and could detect nuggets

The coil bundle( cross section )was wound in clockwork fashion 12-2-4-6-8-10 0'c then alongside each one of those until all the gaps were filled so that the coil bundle was hollow and in a full circle

The current would flow around the outside of the coil bundle 3-4 times before exiting and the inductance comes from the induced inductance between each wire(damn thats hard to explain)

I don't know if your programme can handle that or if in fact you understand what i was trying to explain

I'll leave you's a while to try and take it in!!!...or not!!!

Ron

Hi Ron,

you surely mean the low capacitance coil weave technique (see examples below):

http://www3.telus.net/chemelec/Projects//Metal/New-Coil.jpg

http://www.vk2zay.net/article/file/865

http://img172.imageshack.us/img172/2733/silk1ir1.jpg

Yes, I can do such coil simulations as well. But they won't perform better than the mono coil. The inductance law remain the same of course. It's just a means of reducing the coil capacitance.

The benefit I see is, that the magnetic field strength is getting diluted over a wider region range due to a wider coil bundle, which is causing less disturbing ground magnetising effects.

Cheers,
Aziz

[authere]

Hi Aziz,

I played with coils a while back and came up with an anti interference coil of sorts,but i never really tested it fully and never had a dc ohm and inductance meter,but directly under power lines it was as quiet as a mouse and could detect nuggets

The coil bundle( cross section )was wound in clockwork fashion 12-2-4-6-8-10 0'c then alongside each one of those until all the gaps were filled so that the coil bundle was hollow and in a full circle

The current would flow around the outside of the coil bundle 3-4 times before exiting and the inductance comes from the induced inductance between each wire(damn thats hard to explain)

I don't know if your programme can handle that or if in fact you understand what i was trying to explain

I'll leave you's a while to try and take it in!!!...or not!!!

Ron

[Aziz]

Hi all,

I'm open to other ideas and wishes now as I have covered most interesting coil architectures. Maybe I did forget something. Please, just remind me.

- I have left over the two-box principle (trivial and causing a lot of work, response is dependent on target orientation, would require a coil software upgrade).

- The tandem coil - what was meant by it? I need a description or a sketch so we know, we are talking about the same thing.

- Spiral coils: trivial but causing a lot of processing time (more wire elements required).

Anything else you would like to know and how a new coil performs?

Or should we invent a new coil together? Yes, we can do it.

Cheers,
Aziz

[Aziz]

Zipped Excel Table of Coil Comparison

I'll add from time to time an updated version.
This analysis hasn't been finished at all. But you can enjoy the latest version of course.

CoilComp5(CoilCMP2)-03.zip

Cheers,
Aziz

PS: Ignore all Copyright (C) infos & rights. It's free for all of course!!!
"MadLabs Inc."(c)(tm)(r) first mad engineer is metal detectorists best friend.

[Aziz]

Detection Depth Improvement: Gain Compensated AI Coil Comparison vs. Mono Coil (Reference)

Hi all,

provided that, we can eliminate some EMI noise contribution due to anti-interference (AI) coil configuration and compensate for the target response losses with additional gain in the amplifier (due to less EMI noise), we can directly see the detection depth improvement for the specific coil configuration compared against our reference mono loop coil.

You can see, that the detection depth improvement increases at low noise floor levels much more. That's the reason, why you should reduce the noise floor level using an ultra low noise amplifier. There is the real magic buried.

I'll show you the comparison of selected AI coils and additional gain compensation vs. the mono coil.







And this is how to read the detection depth improvement between the mono coil (reference) and the other coil:


Cheers,
Aziz

[Ferric Toes]

That's a good EMI noise pickup configuration Eric. I love it.
Aziz

What is noise to a metal detector is music to a radio.

Most manmade noise is vertically polarised, so with the shielded coil oriented as shown in the photo, the e field noise pickup from sources of random noise is minimised, while the magnetic component of the wanted radio signals are maximised. Hence my radio is quieter. With a PI detector and a horizontal coil, luckily the radio pickup is minimised, but not eliminated.
Using a stacked noise cancelling arrangement will reduce r.f. pickup but not eliminate it, even if the coils were perfectly balanced. This is because there is a steep gradient (depending on the frequency) as one approaches the ground surface. Hence the lower coil in the stack will have reduced signal compared to the top one. This will get worse for greater coil spacing. (the trade off effect). Back to back D coils would be better, but there is a trade off there too.

One area I am going to look at is to make various shielded and unshielded AI coils (bifilar centre tapped ones also) and use them as radio antennas to find which gives the least pickup and best HF cancellation. I reckon that the damped and shielded RX coil response ought to roll off sharply somewhere in the region of 200 - 500kHz. The best arrangement would be coil/s plus preamp as used in a PI detector, with the preamp output into my 1941 radio which goes down to 75kHz. Let's see what we can hear!

With regard to conventional monos and DDs I suspect that the shielding used is often woefully inadequate and may even be making the r.f. pickup worse.

Eric.

[Aziz]

My point was that ,in general, you can not name an "ultra low noise op-amp" to an "ultra low noise amplifier".


yes you need, even when you are forced by some reason, to use a relative big and "noisy" resistor.
The 1k resistor with the 1.3uV noise across the 100kHz ,it will give an output of 130uV after the opamp(gain of 100).
Consider an opamp with noise of 5nV/sqrt(Hz), then the total noise output from the opamp only, it will be at 160uV.
You see, is not only comparable to the resistor noise but for a common op amp, simply bigger.
For a opamp of 1nV/sqrt(Hz) the total noise output is 32uV. Between these two op amps there is a noise difference of about 130 uV.
Personally, i will be prefer to have 130uV less noise, for a few bucks more.

My definition of an ultra-low-noise op-amp/amplifier is:
Input referred noise density <= 1 nV/sqrt(Hz)

But if want to use the 1k resistor at your op-amp input, you're screwing up the op-amp's low noise feature. In this case, you can use the NE5534 further and save your money.

When I mean, you need an ultra low noise amp for some configurations, then it is so.
Aziz

[Bill512]

Yep, confused.

My point was that ,in general, you can not name an "ultra low noise op-amp" to an "ultra low noise amplifier".

You don't need an ultra low noise op-amp, when you have 1 k resistor at the inputs of it.
Aziz

yes you need, even when you are forced by some reason, to use a relative big and "noisy" resistor.
The 1k resistor with the 1.3uV noise across the 100kHz ,it will give an output of 130uV after the opamp(gain of 100).
Consider an opamp with noise of 5nV/sqrt(Hz), then the total noise output from the opamp only, it will be at 160uV.
You see, is not only comparable to the resistor noise but for a common op amp, simply bigger.
For a opamp of 1nV/sqrt(Hz) the total noise output is 32uV. Between these two op amps there is a noise difference of about 130 uV.
Personally, i will be prefer to have 130uV less noise, for a few bucks more.

[Aziz]

Here is an interesting paper regards to AI configurations:
UXO TARGET DETECTION AND DISCRIMINATION WITH ELECTROMAGNETIC DIFFERENTIAL ILLUMINATION
www.serdp.org/content/download/3749/59543/file/UX-1355-FR-01.pdf

Cheers,
Aziz

[Aziz]

confused???

Yep, confused.

You don't need an ultra low noise op-amp, when you have 1 k resistor at the inputs of it.
Aziz

[Bill512]

You're a bit confused, don't you?
Aziz

confused???

[Aziz]

I was idlely thumbing through a 1944 volume of Wireless World when I saw something that looked a bit familiar. It was a schematic for a frame aerial that was balanced, i.e. it had a centre tap and could be connected differentially. One benefit, according to the article was that it did not need to be shielded, unlike a single ended winding. The article pointed out the benefits of a frame aerial, either shielded or balanced, as reducing capacitive noise, or the electric field component. Having a restored and working 1941 RAF Bomber Command radio, I thought I would give a wrap around lead tape shielded single ended 300uH mono a try as a radio aerial. Wow, it works much better than the internal 20ft wire aerial I have in my workshop. It is easily turned for optimum alignment to the transmitting station too. The electric noise, which seems to come from wall plug switched mode power supplies, solder station, fluorescent lighting, central heating controller etc is way down. Clear signals are heard right up to 18MHz, which is surprising. When I get a bifilar wound differential coil of similar diameter it will be interesting to compare - with and without shielding.

Eric.

That's a good EMI noise pickup configuration Eric. I love it.
Aziz

[Aziz]

maybe is trivial to mention it, but using an op-amp for the first stage,actually it's not a "ultra low noise amplifier".
For example, using an opamp with a noise spectral density of 1nV/sqrt(Hz) ,gain 100 and for 100kHz bandwidth, the total "internal" noise (only from the opamp) will be around 32uV.
Not to mention the noise contribution from the resistors on the opamp configuration.
For a quick comparison ,the 1k resistor on the same bandwidth it will give a noise about 1.3uV.

You're a bit confused, don't you?
Aziz

[Aziz]

Not exactly. The tandem or top hat configurations use doubled the area with the same radius, but doubled nevertheless. IMHO all the AI coils should be given the same treatment as they all are using double the geometry this way or another.

No the comparison is fair as the RX- coil becames a reference EMI pickup coil.
In regards to EMI noise induction:
N(RX)*A(RX) = N(10" TX)*A(10" TX) for normal configurations (not AI),
N(RX-)*A(RX-) = N(RX+)*A(RX+) = N(10" TX)*A(10"TX) for AI configurations.

It is important to show the target response loss in the AI configurations to see the additional gain you would require to compensate for it. There won't be an ideal (full) EMI noise rejection either.

Aziz

[Bill512]

Davor,

Ultra low noise amplifier noise density := 1nV/sqrt(Hz)
(This is enough for us and we need some noise for the Gaussian distribution.)

The ultra low noise amplifier is important for AI configurations, where you have to compensate for the response losses. You can't beat the best without considering the amplifier noise. But we are going to beat the best.

Aziz

maybe is trivial to mention it, but using an op-amp for the first stage,actually it's not a "ultra low noise amplifier".
For example, using an opamp with a noise spectral density of 1nV/sqrt(Hz) ,gain 100 and for 100kHz bandwidth, the total "internal" noise (only from the opamp) will be around 32uV.
Not to mention the noise contribution from the resistors on the opamp configuration.
For a quick comparison ,the 1k resistor on the same bandwidth it will give a noise about 1.3uV.