HOW DOES A METAL DETECTOR WORK?

To understand how a metal detector works, we first need to understand a bit about magnetism and electricity. If we pass an electrical current through a wire, a magnetic field is formed around the wire. Conversely, if a magnet is passed over a piece of wire, it induces an electrical current into the wire. This is called an eddy current. In a straight piece of wire, the induced magnetic field is very short-lived as the eddy current has nowhere to go, dies out quickly, and consequently the magnetic field created is quite weak. If the same eddy current is magnetically induced into a piece of wire with both ends electrically joined (like a ring), these eddy currents effectively run round and round, creating a stronger, more concentrated magnetic field which lasts longer.

The transmitter current of a metal detector is applied to the coil (of wire) and creates a large, concentrated magnetic field around the coil. This magnetic field will induce eddy currents into any metal targets in the ground and they will in turn create their own magnetic field. This magnetic field around the target then induces a current back into the detector coil. This is processed in the receiver and results in a sound from the detector. (Or in the case of a detector with a threshold, a change in that threshold.)

Consequently, one of the “best“ targets (or one that can be found at greatest depth) is a high carat gold ring (it is a shape that is most conducive to greater eddy currents, has high conductivity for larger currents and therefore a very strong magnetic field). If this same ring is cut, you may only find it at a quarter of the depth as the eddy currents quickly die out! “Worst“ target may be a thin straight piece of wire (weak, “short“ eddy currents and therefore a weak magnetic field).

GROUND NOISES, WHERE DO THEY COME FROM?

Ground noises are the result of magnetic properties of minerals in the ground. These minerals create their own magnetic field in response to the transmitter and this field induces an electrical signal back into the detector coil, producing audible sounds from the detector electronics, hence the term “ground noise.“ (If this is going a bit fast for you, first read the paragraph “How do metal detectors work.“) Different minerals have differing magnetic properties (eg magnetic permeability, memory and saturation). These differing properties will produce differing sounds in metal detectors. Areas containing these minerals are termed “mineralised,“ “noisy“ or “contaminated.“ Because of these mineralisation effects, as a metal detector is swung over these areas, the metal detector will produce a variety of sounds. These sounds, (ground noises), often sound the same as the sounds normally heard when a target is detected. Differentiating between targets and ground noises can be impossible at times, especially as ground noises can often be louder than target returns, thus drowning them out. Water, and especially seawater, can cause the same effects in many metal detectors.

BBS/FBS/MPS TECHNOLOGY

BBS technology is used with the Sovereign/Excaliburs and FBS with the Explorer series of Minelab metal detectors. FBS is enhanced BBS technology. These technologies have the ability to cancel out ground noises because they have the ability to differentiate between ground noises and metals. Additionally, they have the ability to determine conductivity of metal targets (and the Explorer has the additional capability of determining inductivity).

MPS technology is used in the Minelab series of “Super Detectors,“ SD2000, SD2100 and the SD2200D. These detectors were specifically developed to work in highly mineralised areas in which gold is normally found and can cancel even the most severe ground noises encountered in heavy lateritic areas (dark red ground).

These technologies transmit a train of pulses which vary in width. These pulses are square and are therefore made up of a combination of frequencies. The pulses from a Sovereign for instance, use a fundamental frequency of about 1.5KHz and odd harmonics every 1.5KHz. So there will be energy in the pulse at 1.5KHz, 4.5KHz, 7.5KHz etc. Ground or target response depends, to a certain degree, on the transmit frequency. Lower frequencies tend to give more depth and higher frequencies tend to find smaller objects. Thus, a range of frequencies, rather than one or two gives a much better response across a range of ground conditions and target types. In addition, the receiver circuits will receive a much wider range of information, giving effective ground canceling and more accurate discrimination. Receive frequencies up to 25KHz are utilised in the Sovereign and up to 100KHz in the Explorer, this is one of the reasons that the Explorer has more depth (especially in benign ground) and more accurate discrimination than the Sovereign and many other detectors.

SD2200D GROUND BALANCING

Because the SD series is designed for very noisy ground, correct ground balancing is critical for optimal operation (ie max depth). The SD2200D needs to sample the mineralisation that it is balancing out. To do this, the coil ideally needs to be raised and lowered to 125mm from, and parallel to the ground. If the coil is lifted more than 300mm above the ground, it will be too far away to correctly detect ground noises. If it can’t detect ground noises, it can’t balance them out! Also remember that when the Fixed/Track switch is initially switched to track, the SD2200D ground balances very fast for the first five seconds, then slows down. This means that if you haven’t balanced fully in five seconds, switch to fixed then back to track and raise/lower the coil again until all ground noises disappear. A slower ground balance is used whilst detecting so that the detector will not balance out targets.

WHY DO WE NEED NOISE CANCEL IN THE EXPLORER AND SD2200D?

The BBS/FBS/MPS detectors transmit and receive a wider range of frequencies than single/dual frequency detectors. Consequently, a wider range of atmospheric electrical noise can interfere with the internal processing circuits, resulting in an oscillating (warbling, chirping) threshold and resultant loss of sensitivity. The “received electrical noise,“ referred to consists of electrical noise radiated by power lines, air conditioners, computers, other metal detectors and some weather conditions (storms etc). All electrical appliances radiate some electrical noise! Under these conditions, the microprocessor is not only trying to differentiate between ground noises and target returns, it is also trying to process this “interference,“ so some target and ground information is not correctly processed, hence the loss of sensitivity (Turn one of these detectors on whilst inside a house, turn up the sensitivity and you’ll hear what I mean!). Due to the larger spectrum used by Explorer and SD’s, at higher sensitivity settings, this problem may be more noticeable. The internal noise cancel feature overcomes this problem. This feature also allows a number of detectors to be used in close proximity. The following noise cancel explanation refers to the Explorer but applies equally to the SD series (SDs transmit on more frequencies as the variable frequency oscillator is taken through many more steps during the noise cancel procedure).

The 28 frequencies transmitted by the Explorer are all produced from one variable frequency oscillator, and all 28 are multiples or sub multiples of this oscillator frequency. When noise cancel is initiated, it steps the variable frequency oscillator through its range of 11 frequency steps (thus slightly changing all 28 frequencies by the same ratio). It records the received electrical noise at each step. At the end of this procedure, it selects the quietest frequency (ie the one that received the least noise). It also remembers this frequency when you switch the detector off and sets it to that particular frequency when you turn it back on again. The receiver is designed to properly receive the whole range of possible transmitted frequencies, so neither depth nor discrimination will be affected by which frequency the noise cancel circuit finally selects.

Why it is important to hold the coil very still during the noise cancel procedure:

If the coil is moved during this procedure, the noise being picked up in that instant may change because the coil has moved, therefore whatever frequency was being sampled at that instant will have “noise“ recorded against it, and it means that it won't use that particular frequency (even though that frequency may have been the quietest).

WHAT IS HALO EFFECT?

Halo effect makes targets that have been in the ground a long time appear much larger than they actually are. Imagine an iron nail that has been in the ground for a number of years. As moisture in the ground slowly rusts the nail, forming iron oxides and iron salts, these oxides and salts migrate from the surface and tend to form a “halo“ around the nail. This, of course, makes it a lot “easier“ for a metal detector to detect. This effect is basically the same for most metals. When the target is dug, this halo effect (which takes years to form) is destroyed. So if you find a target due to the halo effect, and, for instance, it was a very weak signal, you will find that if you rebury it at the same depth, you will probably not now detect it (unless you leave it for five years!). For you beach-hunters, halo effect rarely has time to build up and will be destroyed if the ground moves. Relics are often found at greater depths due to these effects.

Gold can be very hard to find due to the odd shapes that are often formed. Halo effect really helps us liberate a lot of gold, as most gold hasn’t moved in a long time! But gold doesn’t rust or corrode I hear you say. Correct. Often where you find gold there are also salts and chemicals (like cyanide) and these will leach out the gold into the surrounding minerals, causing a halo effect.

WHY CAN CHAINS BE VERY HARD TO FIND?

Chains consist of many individual links. Unless chains are tangled, or have built up a halo effect, metal detectors can only pick up the individual links, not the complete mass. So if the individual links can be detected (or the largest bit, eg the clasp), the chain will be detected, fine chains with small catches are often impossible for metal detectors to detect.

Ralph @ Sun Ray Detector

PULSE INDUCTION

S.A.T. - Self Adjusting Threshold. This is a circuit that maintains the threshold settings as you sweep the coil over ground. Changes in ground makeup alter the threshold as you sweep. With S.A.T. it does the circuit re-setting for you. The Deepstar, Aquastar, and the Goldquest SS, have a control the user can adjust, to change the rate of when the circuit re-sets automatically.

Most detectors today employ S.A.T., VLF and PI. It is the motion needed, in the motion mode. Without motion, the S.A.T. tunes out the target, trying to maintain the threshold. In no motion pin point, as in some VLF detectors, that turns off the S.A.T. so you can pin point easier.

With the adjustable S.A.T. control, one can set up the threshold re-set speed for the different hunting conditions encountered while detecting. With slower S.A.T. speeds, one can detect targets deeper than they can in the same instance with a fast S.A.T., because your giving the circuit more time to analyze the signal. The trade off is that the ground will also have an affect. Tougher ground generally requires a faster S.A.T.

As far as what adjustments to use, changing from beach jewelry hunting goes, I’ll start with the Headhunter PI because it’s the most basic. A user would have to determine these setting on his own, but one “may“ have to re-set the frequency/pulse delay control. The same on the Goldquest SS plus the addition of the S.A.T. adjustment. With the Deepstar, that starts to get a little more complicated, because it has a whole lot of adjustments for the user to set, “if needed.“ With the Deepstar we can get into transmitter pulse width settings, receiver sensitivity settings, frequency settings, pulse delay settings, audio adjustments, etc.

Eric Foster

There are usually three things in a PI receiver that affect the detection range and/or sensitivity. Firstly, the sample pulse delay. The shorter this is, the more the detector will respond to lower conductivity items. There are certain items in this category that we want to detect, such as thin low carat gold rings, platinum rings, gold chains, ear rings etc., However, a salt wet beach constitutes a very large low conductivity object and will give a signal which increases rapidly as the sample pulse delay is shortened. Near, and just in, the water, 10uS is about as short as you can go with standard coil and electronic arrangements, but even at 15 - 20uS a signal is still apparent. This makes itself known by a rise in the amplitude of the audio tone as you lower the coil onto the wet beach. Those detectors that have SAT will adjust this signal back to the threshold if the coil is left stationary. A variable SAT control alters the rate at which this adjustment takes place. As beach conductivity is reasonably constant over a large area, and we keep the coil at a fixed height when scanning side to side, then the tone variations will be kept to a minimum. Some detectorists swing the coil in an arc over the ground, which is not the way to do it in any circumstances, as you lose range at the ends of each swing as well as causing large variations in beach signal.

Salt water and wet salt sand are the only situations which cause a ground conductivity signal in a normal PI. Fresh water and freshwater beaches give no conductivity signal and even brackish water gives little or no response. Some times there are areas of an ocean beach where a fresh water stream or spring flows into the sea. This dilutes the sea water and although the beach is fully saturated, it gives no signal.

Beach conductivity does vary with salinity from place to place and if the audio background variations are a nuisance, they can be reduced by increasing the sample pulse delay, remembering that the sensitivity to small objects will also be reduced. This leads on nicely to the second feature of a PI receiver that can be made variable. This is the receiver gain, or amplification. Not all PI detectors have this, but it is a useful feature. I usually vary the gain of the dc amplifier at the end of the receiver chain. There are generally less problems with this, than trying to vary the gain of the front end amplifier. In situations of ground noise or electromagnetic noise from radio transmitters and power lines, reducing the gain just reduces everything in proportion, but you can get back to an acceptably smooth threshold. You can also still run at a short delay and detect small objects, but at less range of course.

The third method is to turn the audio threshold control back until the background variations disappear. The result of this is similar to reducing the gain; the range on all objects is reduced in proportion to the backoff of the threshold. Usually a little bit of each, plus adjusting the TX frequency (if available) is all that is necessary to achieve sweet operation. Here's an analogy to a car. Maybe you have a 300HP engine, but the power that you can actually use depends on the road conditions.

Eric Foster

There is quite a variation in PI timing; it not only depends on what you want to achieve in detection sensitivity, but how it is generated in the timing circuits. A typical PI for beach use, where one wants good sensitivity to rings, coins etc, would have a TX pulse width of 50 - 100uS. The switch off of the TX current would occur in 5uS or so with say another 5uS for receiver settling and damping of residual ringing. This would allow receiver signal sampling to start at 10uS for the highest sensitivity, or 15uS for more general use, as in the majority of beach PI’s.

10uS is about the shortest sample delay you can use around sea water, which is moderately conductive, otherwise you get too much response from the wet sand. With fresh water you could use even higher sensitivity but coil and circuit design becomes ever more critical for stable performance.

A TX pulse width of 50 - 100uS is OK for rings and small gold objects, but if you want to search for larger objects i.e. silver ducats, gold bars etc, you will get a better signal and more range from a wider pulse. I have used pulses up to 1mS for very large non-ferrous objects. Also, the sample pulse delay time can be made longer to reduce sensitivity to small items when searching for the big ones.

TX pulse frequency can vary a lot between different detectors. Current units run from a few hundred pps to several thousand. The actual frequency used is generally a compromise resulting from interference reduction, response speed and battery consumption.

Some timing circuits can alter the different pulse widths and timings independently, Deepstar, Aquastar, while others, such as the Goldquest, CS6PI and CS7 alter all of the pulse widths and timings proportionally together, by means of a clock frequency control.

Eric Foster

Electromagnetic waves propagate at the speed of light, and this exactly what happens with any metal detector. A coin or ring at 15" is going to see the transmitter field virtually instantaneously, and likewise the return signal is at the same speed, so there is no problem with sweep speed here.

For a small object at that range, the return signal is very minute and will probably be at the microvolt level. On the other hand, electromagnetic noise picked up by the coil from power lines, radio transmissions and general atmospheric noise can be many times higher than the object signal. Some sort of signal processing then has to be done to extract the object signal from the noise and the simplest way of doing this is by averaging over a small period of time. In the process of averaging the noise, being to some degree random or non-coherent with the transmitter frequency, will average to zero, whilst the object signal will be enhanced to a useable level. You may have to average over a thousand pulses to achieve this, and if you are transmitting 1000 pulses per second, then there is going to be an apparent lag in the detector responding as you pass the coil over an object. If you sweep too fast, then you could miss a deep target. Sweeping more slowly allows the electronics to catch up with the signals being processed.

One of the reasons why PI detectors (mine in particular) use high pulse rates, is so that the object signals are coming in faster and averaging can be done over a shorter time, thereby minimising the lag.

There is another reason for not sweeping too fast, and that involves the SAT, or Self Adjusting Tuning. It’s not tuning in the accepted sense, and Tracking or Threshold would be a better term. Motion type detectors have this feature and it is an electronic way for the detector to adjust itself to slowly changing background signals such as those generated by a wet beach or mineralised ground. This automatic adjustment has to be slow, otherwise the object signal would be tracked out too. It you hold a coin in front of the Goldquest coil, you hear the signal fade away after a short time. The rate of fade out is governed by the SAT control.

If you are searching a beach where you are sweeping from dry sand to wet, or wet sand to the water, the SAT is trying to keep up with, and compensate for, the varying signal from the beach. If you sweep too fast, then you can beat the SAT and get annoying threshold variations.

Detectors are one big compromise, and striking the balance between response speed, sweep speed and SAT speed for any given situation of ground mineralisation, object size and object depth, is not an easy task. Practice on the part of the user is important too, as they learn the best settings for particular situations.

Eric Foster

Goldquest SS Coil Comparisons.

I recently had some requests to make some 8in coils for the SS, in addition to the standard 11in that is normally supplied. A smaller the coil is useful in areas where there are rocks or vegetation, which make the use of larger coils difficult or even impossible. Also, it gives more precise separation of close targets, sharper pinpointing, and less ground signal in areas with iron mineralisation.

The interesting question, though, is what do you sacrifice in terms of detection range? Generally, the requests for a smaller coil have come from persons wishing to use the SS for nugget hunting, where conditions are rather different from those encountered in beach searching.

The results are interesting and fall broadly in line with the range curves for different coil sizes that I posted recently. All of the tests, except one, were conducted with gold nuggets that originated in Victoria, Australia. The exception was a gold ring, which was used to give a larger area target.

Initially a 0.1gm flat nugget was tested but was not detectable at any range. This is because the decay time of the signal was faster than the 10uS sample delay used in the SS.

0.4gm flat-------------3.5in (8in coil)----2.5in (11in coil)
0.5gm irregular-----2.5in-----------------2.5in
0.9gm flat-------------6.0in-----------------6.0in
2.0gm flat-------------7.0in-----------------7.3in
3.6gm irregular-----7.5in-----------------8.0in
3.4gm 22k ring------13.in-----------------15.in

All the tests, except the last one, were done by stacking plastic blocks on the coil and sliding the nuggets horizontally on the surface of the top block. This gives a more accurate figure than just using a ruler and holding the target manually. The audio was a just discernible, but clearly identifiable, change. The ring was done manually with a ruler for lack of enough blocks. Another inch could possibly be added to the ring for a just discernible signal.

The results show that, for small nuggets the 8in coil in not much different in range to the 11in coil, and at the lower end starts to win out. The 0.5gm irregular nugget shows less range because of its convoluted shape and maybe different composition. In fact it should be born in mind that the variation in composition of nuggets from different locations can result in widely differing detection ranges. The 3.4gm ring demonstrates that the radiating area of an object has a great effect on range and that the large coil then shows a distinct advantage. Both size coils are wound for the same inductance, which means that the 8in has a few more turns. This helps to some degree in making up the loss in range due to the difference in diameter, which likely explains the identical ranges on two of the targets.

Eric Foster

Goldquest SS Tuning

Don't be afraid of turning the SAT up to half way, or even more. This with make the GQ tune quicker to the changing beach signal without noticeable loss of depth. It is one of the facts of life that having a detector running at 10uS, you are going to get more response from the wet sand. You could turn the REJECT control a notch or two back to reduce the effect, at the cost of losing a bit of small object sensitivity. The response time of the GQ is considerably faster than the Surfmaster due to its 10,000 pulse per second rate compared to 700 or so. This will make the whole thing seem livelier, plus there is no series diode in the signal chain which smooths out low amplitude fluctuations to give an artificially smooth threshold.

Eric Foster

PI Terminology

PULSE WIDTH. The width of each individual transmitter pulse. Depending on the detector and its intended application, the pulse width can vary between 50 microseconds (uS) and 1000uS. 50 to 150uS is usually the width for ring and coin size objects while wider pulses suit detectors designed for locating larger objects. A boat towed PI designed to look for cannons would likely use 1000uS. Examples are:- Whites Surfmaster 50uS, Deepstar 100uS, Aquapulse 300uS, Superscan and Aquapulse towed 1000uS.

PULSE FREQUENCY or PULSE REPETITION RATE. The number of times the transmitter pulse is repeated every second. The frequency affects the response speed, power consumption and interference rejection, although all of these are influenced by many other factors in the circuit design. Examples are:- Surfmaster 800 pulses per second (p.p.s.) Deepstar 3000 p.p.s. Aquapulse 170 p.p.s. Superscan 66 p.p.s.

PULSE DELAY. The time that is allowed to elapse between the transmitter switching off at the end of each pulse and the start of the receiver function or receiver sample pulse. The shorter this time is, the higher the sensitivity to smaller and/or poorer conductors. On some PI detectors (Deepstar) this is called a REJECT control. The shortest pulse delay is with the control on MIN and if it is turned clockwise the pulse delay time is lengthened which enables bits of foil and pulltabs to be rejected. However, thin and lower carat rings may also be lost. For beach work on salt wet sand a minimum pulse delay of 15uS is about right so as not to get too much signal from the conductive salt water. Detectors for diving have to use longer delays as the water signal is much greater. Detectors for large objects, as well as a long TX pulse, use longer delays to minimize signals from small bits of metal.

DECAY TIME or OBJECT DECAY. The time that the small electric current induced in a metal object flows. The energy is dissipated in the electrical resistance of the object so that poorer conductors (higher resistance) have a shorter decay time. A thin ring may have a decay time of 50uS while a silver cob could be 500uS. A detector set with a pulse delay of 100uS would pick up the cob but not the ring.

SAMPLE PULSE or SAMPLE WINDOW. The period of time from the start of receiver sampling to the end of the sampling period. Usually between 15 and 50uS. In the Deepstar it is set at 20uS irrespective of the pulse delay. In the CS7 the sample pulse tracks the pulse delay i.e. it is always the same value as the delay.

DUAL SAMPLING. This is a system used by most PI’s where there is a second sample pulse that occurs much later (a few hundred microseconds) but before the next transmitter pulse. The signal from the second sample is subtracted from the first to give interference cancellation (powerline noise) and also to eliminate the signal from moving the search coil in the earth’s magnetic field. Most of the object signal has gone by then so that object signals are not cancelled.

INTEGRATION TIME. As the received signals are in the form of pulses and these are then sampled in short windows, an averaging system is needed to smooth these out and end up with a d.c. voltage that can operate a meter and/or and audio generator. The integration or averaging time relates to the number of pulses over which the average is taken. With the Deepstar, around 1500 samples are averaged to give the d.c. output. Hence the response time is about ½ second (3000 p.p.s.)

Eric Foster

Ground balance is required when certain ground conditions create signals that sound much like one large continuous target deep target beneath the coil. The closer the coil gets to the ground, the louder the response. Under such conditions, one has to keep the coil very level above the ground to reduce the audio variations caused by the ground if no ground balance is present.

Now, the objective of ground balance is to eliminate the signals caused by the ground and to do this, about the only way I know of, is to perform some form of subtraction process such that the signal being subtracted is equal to the offending ground signal. If this is done, then the ground response is minimized.

I say minimized because the ground response is not linear, so any form of simple subtraction will not be equal over the range of signal changes that occur as the coil is lowered. The shorter the delay, the more likely will be the degree of error. Now, to compound matters, certain targets will have responses such that they are very similar to the ground signals, so they will be eliminated also.

If the pulse is lengthened, then the ground signal characteristics change. Also, if the primary sample is taken a little later, a subtract process will require a different subtraction ratio which will then cause different objects to be ignored. Finally, if the target signals from the two different pulse lengths are combined, a full range of targets can be detected.

Discrimination is another can of worms even if just iron objects are to be rejected. The reason lies in the fact that iron objects span a wide range of target responses. Large or thick iron object decay signals look nothing like signals from something like a piece of an old tin can.

Simple delay techniques may minimize the digging of only certain iron items (i.e. large or thick iron) while other pieces of ferrous metal (pieces of tin cans) will respond more like a good target. Now, flip things around and we will see a large non-ferrous target such as a very large gold nugget may respond much like a iron object by having a long decay time. So, the design of a good discriminating PI is much more complex than most people realize.

Reg