PI as a variable reluctance sensor

[Anonymous]

Ref: US patent 6,326,790 (Ott et al, "Periscope")
Variable reluctance sensors of various types have been, and are, used in industry for the detection of metal. This class of sensors does not have the sensitivity needed to compete on an equal footing with the PI and VLF methods for ordinary hand-held metal detectors. However, the Otts have discovered an application for the variable reluctance principle which doesn't require high sensitivity, and which has requirements (esp. size) which are not easily met with the other technologies.
I would suppose that there are patents on the basic principle going back at least 40 years, and possibly 100. However, I have not researched this. Although the Periscope uses this principle, the Otts' patent is not on the principle (long since in public domain), but on the specific things they needed to do to get a good Periscope.
Now, back to PI.
The time it takes for the flyback pulse to terminate is proportional to the inductance of the searchcoil. I presume that most engineers who have played around with PI realise from first principles as well as by experiments, that magnetite extends the duration of flyback.
If the receiver timing is pushing up as close to the flyback as possible in order to detect low-conductivity targets, lowering the searchcoil to ground with high magnetic susceptibility pushes the flyback decay into the receiver, causing "ground pickup". The obvious solution is to leave more room between flyback and receive gating, but that solution sacrifices sensitivity on the smallest targets.
There are several other approaches to this.
1. The receiver turn-on can be controlled by a fixed delay timer which is triggered by the collapse of the flyback voltage below a certain threshold.
2. The flyback duration can be demodulated as a separate signal, and this demodulated signal can be summed or subtracted in the proper proportion with the primary demodulated signals in order to balance out the unwanted ground pickup.
3. The flyback duration can be demodulated as a separate signal, and this demodulated signal can be used to adjust either the receiver timing or the up-front receiver complex impedance, in order to provide cancellation of ground effect resulting from the real component of magnetic susceptibility (i.e., magnetite).
4. If the flyback duration is demodulated, the corresponding signal can be balanced against the other demodulated signals in a proportion controlled automatically or manually by the operator, in order to balance the combination of the real and imaginary components of magnetic susceptibility, which is necessary for quiet operation in ground which contains minerals such as maghemite which have a substantial magnetic loss angle.
5. As with Ott's arrangement, demodulation of the flyback duration can provide ferrous-nonferrous discrimination, although the discrimination signal thus obtained will not have high sensitivity, and will be affected by ground minerals.
---------
The Minelab GP Extreme advertises iron discrimination. In view of the fact that this product can operate without induction balance, and the fact they state that the discrimination is for shallow iron only, it would be a reasonable guess that they are demodulating the flyback duration and using that signal for discrimination. However, there are other ways to do iron discrimination, so they may be using some other method.
--Dave J.

 

[Anonymous]

Sorry, poor choice of words!! "INDUCTANCE SENSOR" would have been better. "Variable reluctance" most often refers to a coil which has a permanent magnet core: when iron is introduced into the field, the resulting change in flux density induces a voltage in the coil which is then detected-- there is no active transmitter.
For a quick download of US patents without the graphic figures, text only, go to www.uspto.gov (the US patent office). Carl's website www.thunting.com/geotech (click on "miscellaneous") has a list of metal detector patents including graphics. However, if you want to look at the Ott patent be aware that on Carl's website the download files got listed with the wrong patents, and the Ott file is actually the monster file listed with the patent immediately above it (and vice versa).
--Dave J.

 

[Anonymous]

I've never had much trouble with magnetite and PI's. A few years ago, I designed an industrial PI to work in Sweden's largest underground iron ore mine, of which the ore is 96% magnetite. I had to find small metal parts in powdered ore running on a conveyor into a sampling plant. Magnetite has high susceptibility but negligible viscosity; which is the killer. Many black sands contain magnetite grains and PI is effective on these. However, I once took a straight PI to Lanzarote, in the Canaries, where the beaches have sand derived from volcanic lavas. There it was almost impossible to use the detector because of the viscosity signal. Any magnetically susceptible material will increase the effective inductance of the search coil. How much, I had no idea till today. This is where the exchange of ideas and experiences is useful. It spurs one on to look at more detail, things that have been taken for granted in the past. Using a 10in coil with an

 

[Anonymous]

Thanks, Eric.
In principle, mineralization sufficient to increase inductance by 1% would also increase flyback by 1% (assuming resistance-limited transmitter). That's just a small fraction of a microsecond, but easy enough to demodulate, low pass filter, and then differentiate to remove the quasi-DC component.
The S/N ratio of the resulting "reactive" signal wouldn't be as good as that in a VLF machine, but you only need a little of it to balance out magnetic viscosity (for instance), so it probably wouldn't result in any significant impairment of the S/N ratio of the balanced signal. Of course using a "reactive" signal to balance magnetic viscosity produces a result similar to that of VLF ground balancing-- if the ratio of magnetite to maghemite changes, you're not balanced any more.
Here are my theories about the so-called "imaginary component of magnetic susceptibility" which seems pretty real when you can hear it beep on a metal detector.
Magnetite mined as iron ore is usually pretty pure stuff, with large crystal sizes, and most of the impurities zone-refined out to the crystal boundaries, leaving the actual mineral magnetite content relatively free of crystal lattice defects. So-- low loss angle. A naturally occurring form of ferrite.
Most of the magnetite sand found on beaches had its origin in granite and other coarse-grained igneous and metamorphic rocks, in which the magnetite was a minor constituent. Therefore as the crystals form, they are having to separate themselves from huge amounts of contaminants, and so the crystals are much less pure and have many lattice defects. The loss angle may be as high as several tenths of a degree. Then, these magnetite grains get weathered out of the rock and subjected to oxidation (maghemitization) and hydration (lepidocrocitization, you never heard that word before, eh?). The oxygen and water molecules create additional lattice defects. That's the black sand we find on beaches. As a very general rule of thumb, the finer the sand, the higher the loss angle, because there was less zone refining in the crystal formation process, and greater surface-to-volume ratio creating the opportunity for a higher percentage of the material to be maghemitized. (OK, let's forget "lepidocrocitized".)
And then there's the dreaded basalt. Because the stuff cools so quickly (if it cooled slowly, it wouldn't be basalt), little zone refining takes place, the magnetite crystals are smaller, some of the magnetite may be of superparamagnetic size which is easily bounced around by thermal agitation, and the stuff readily weathers to maghemite and other (magnetic) gamma ferric oxides and hydroxides. The unweathered rock often has a loss angle exceeding one degree, and of course the weathered stuff can be a lot worse.
OK, this is an invite for a professional geologist to step in and take over this thread.
--Dave J.

 

[Anonymous]

I've heard of Limonitisation, where Limonite is a mixture of Goethite and a little Lepidocrocite. Basically, hydrated iron oxide. We get layers of Limonite in the sandstones around Oxford.
Eric.

 

[Anonymous]

Limonitization.. that's a mineralogical process?
I thought it was defined as "the dumping of a citrus beverage on a person's head."

 

[Anonymous]

Hi Dave,
You are right but it is only a common practice in the "ite" family. You know them, magnet ite, hemat ite, later ite, maghem ite, etc.
Hard to tell the family members apart since in many cases they all look alike.
Reg

 

[Anonymous]

The fact that some pretty innocuous rocks look the same as the wicked ones, makes possible certain "dealer demo tricks".
In all fairness, most dealers don't stoop that low.

 

[Anonymous]

Eric:
I have from time to time made measurements of "inductance pulling" by iron minerals but those numbers are not readily at hand. So I just tried a 6 1/2 inch coil 3/4 inch above a crude 40,000 micro-cgs (emu) magnetite sand mix standard 2 inches thick. Inductance change was about 0.9 percent.
In "pretty bad ground" of 500 micro-cgs units, that would be an inductance change of about 1 part in 10,000. I think that could be detected by measuring flyback duration, but my gut level feeling is that the S/N ratio would not be very good. Using it to ground balance might sacrifice some sensitivity.
Your straight magnetite sand test gave about 5% change under conditions which were geometrically in the same ballpark. My "standard" is 16% by volume magnetite, so extrapolating linearly (a risky business when magnetite particles are contacting each other) your stuff would be about 70% by volume magnetite. Sharp (mechanically crushed) sand usually isn't quite that dense, but that's obviously the right order of magnitude.
Our results seem to be roughly in the same ballpark.
As a reality check, I tried burying a 100 microhenry ferrite core RF choke in the supposed 40,000 micro-cgs material. The inductance went up by 10%. If it went up 60% [as it might be expected to do in your (Kiruna?) material], that would suggest magnetic reluctance about 40% that of air in your material. I would expect a bit lower, but I think we're at least looking at the right order of magnitude.
Published literature I've seen indicates the relationship between magnetite volume "density" and magnetic susceptibility to be highly nonlinear, and on theoretical grounds one would not be surprised to see this. It may well hold true in natural conditions under which it was possible for a magnetically anisotropic "fabric" to form, but my own crude experiments with artificial sand mixtures show fairly good linearity except at the very highest concentrations.
--Dave J.