Understanding The PI Detector


One of the more popular metal detectors used for nugget hunting today is a type of detector commonly called the Pulse Induction or PI for short. A lot has been written on the general principles of operation but many questions are still unanswered or not answered completely about this strange machine. Also, there is a lot of misinterpretations of information that has been written about PI's and how they work.

As an example, in some books there is a statement that a PI does not "see mineralization" so it is therefore a great detector to use in mineralized areas. Is this really a true statement? The answer is both yes and no.

PI's basically do not respond to the typical iron mineralization such as magnetite or black sand. However, other minerals of the same family can and many times do cause a response. Iron oxides such as maghemite, clays, and other things such as salts commonly found in the ground can cause a PI to produce a rather strong signal. So, generally a very sensitive PI, normally used for gold hunting will respond to ground signals, especially if it does not have some form of ground balancing circuitry built in.

One question that is often asked is what is the operating frequency of a PI. This question is often asked by someone who is trying to relate their knowledge of VLF's to the PI. Unfortunately, because of the nature or differences between types of detectors, comparing a PI to a VLF is sort of like comparing an apple to a potato, so trying to relate the operating frequency of a PI to a VLF or sensitivity to small gold is of little value. The differences between the two types of detectors or the affects of their operating frequencies are quite dramatic so it is best to not try to use the same standards when trying to determine certain things about a PI.

As for a PI, the pulse rate or pulses per second (pps) refers to the number of high current pulses that occur over the time specified. Rates vary from a few hundred to several thousand per second: Generally, more pulses allow for a little better averaging and thus a little better signal to noise ratio. However, a detector will have a tendency to consume more current with a higher pulse rate. A faster pulse rate doesn't mean a detector will detect small gold better. In fact, it is quite easy to build a PI that has a very low pulse repetition rate (PPS) that is very sensitive to very small gold while designing a PI with a high PPS that is not sensitive to small nuggets.

Now, both PI's and VLF's will detect metals, respond to different ground conditions, and even respond to salt water. Both use a coil, specialized circuitry and usually generate a similar output to indicate some object has been detected However, the techniques, circuitry and in many cases, the coils are dramatically different.

VLF's generally produce a relatively low power continuous sinewave into the transmit coil and, analyze a signal received with a separate receive coil winding. A signal from an object will increase the amplitude of the receive signal level but will also shift the receive signal with respect to the transmit signal. Thus, an object can be analyzed by not only the intensity or amplitude increase of the signal but by just how much the signal has shifted.

VLF's generally operate at a single frequency but can be produced to operate at different frequencies. However, each frequency has to be analyzed as if it is the primary frequency and as such, both the signal strength and the shift are used to determine the presence of an object as well as type of metal.

PI's are a different beast all together. Instead of transmitting a low power continuous signal, the PI generates a brief high current pulse to energize the coil and this pulse is repeated at some nominal repetition rate, which can vary from a few hundred pulses per second to thousands per second.

The technique to determine whether an object is present is to analyze the signal coming from the receive coil shortly after the high current pulse is turned off. This is done by sampling the signal coming from the coil some time after each high current pulse. This time after the pulse is often referred to as the delay time Remember, on a PI, the transmit coil may become the receive coil once the transmit signal is turned off so there is no need for a separate receive coil winding. This type of coil is often referred to as a Mono coil.

PI History

There has been considerable work on PI type detectors since the early 1960's. One of the main reasons for their design was so they could be used for archaeological purposes. Most of the work in the evolution of the PI occurred in Europe during those early years, and much of this work was done in England by a young engineer by the name of Eric Foster.

As a result of his involvement with PI's during their early years, Eric Foster began his own business building PI's for industry as well as the consumer market. Many of his initial designs are the cornerstones of some of the PI's used today. Sometime in the early 1980'5, Eric Foster built a PI with ground balancing capability and a rudimentary form of discrimination. He also built a much better discriminating PI around the same time frame.

Minelab was the first to introduce a PI specifically designed for gold hunting in the US some time the 1990'5. The introduction of the SO 2000 really started the serious use of PI's to search for gold even though people began using Eric Foster's detectors for nugget hunting sooner in Australia. What made this ML PI detector excel was the introduction of the use of a DD coil on a PI. The DD coil had the ability to eliminate much of the ground problems making it a quieter choice. One other major advantage of the ML was it operated and sounded much more like a VLF. A PI equipped with either a mono or a DD coil will ignore many hotrocks and have additional depth of detection. However, this depth advantage is greatly reduced in very quiet ground.

Eric Foster's ground canceling detector had a putt-putt type audio, required the operator to retune the detector frequently, and had several different modes, some of which made the ground balance mode seem much less sensitive. Also, since only a mono type coil was available, some of the more severe areas still caused problems even when the ground balance was used. As a result, Eric Foster's ground canceling PI, the Goldscan, never really caught on.

Strange as it may seem, one of the first US patented designs using a high current pulse to detect metals that also had ferrous/non-ferrus discriminating capabilities was designed by George Payne in about 1978 or so. This design not only would distinguish iron objects but also had a basic form of ground balance. This strange design used a bi-polar form of pulsing, which was also unique. Unfortunately, because of the high current necessary for operation and thus, the need for a very large battery, the design was never produced and sold. Instead, American manufacturers focused on developing VLF's for both coin and gold hunting.

How do PI's really work and what makes them sensitive to small gold?

As stated earlier, PI's operate on the principle of generating a large current pulse in the coil and then analyzing the signal in the coil a short time after the pulse is turned off. This cycle is repeated on a continual basis.

As simple as this sounds, the design is quite critical. The key to increasing the sensitivity of a PI is to turn off the current pulse as soon as possible, and then stopping the resulting high voltage spike as soon as possible.

By nature, a PI coil is an inductor and as such, any immediate disruption in current will cause the inductor to produce a very large voltage spike in its attempt to keep the current flowing. This high voltage spike is a side affect of the current disruption that has to be dealt with as quickly as possible so any signal from a metallic object can be distinguished.

When the current is flowing in the coil, a magnetic field is generated that expands from the coil. When this field encounters a metallic object such as a gold nugget, current begins to flow in the nugget as the result of this magnetic field. When the current suddenly stops in the coil, the coil field collapses which in turn causes the current in the object to collapse. This secondary collapse of current in the nugget causes it to produce its own field that now generates back to the coil. This target signal ultimately adds to the collapsing coil signal, thus making the coil signal change very slightly.

The signal strength, and just as important the duration or time of the signal produced by a detected object is a function of the size, shape, and actual composition, among other things. Gold and other low conductive materials may produce a strong signal but the duration of the signal is much shorter than a signal from something like a piece of iron, copper or silver. Very small nuggets, in the few grain range, not only generate a very small signal, but also a very short signal.

Small iron objects, on the other hand, will produce a much larger signal as well as a much longer signal than a piece of gold of similar size. Thus, it is much easier to detect a very small piece of iron than it is to detect a very small piece of gold.

Finding the small stuff in Greater Detail

The key to the success or sensitivity of a PI to small conductive objects such a small gold nuggets is the ability of the PI circuitry to turn the coil pulse current off very rapidly, and then be able to analyze a signal very shortly after the pulse of current has ended. This sudden stop of current in the coil will cause a very large voltage spike that rises almost instantaneously to some voltage generally between 50 to 400 Volts (V). Generally, the voltage level is a function of the FET (field effect transistor) used to deliver the high current. Once this voltage peaks, it will then quickly decay to very near 0 Volts ( 0V) in just a few microseconds (usecs). The rate of the decay is extremely important, just as is the characteristics or shape of the decay of this large voltage spike.

One important factor to remember is a large current normally requires more time for the spike to decay. This becomes important when determining the best design for small gold. Another critical factor is the inductance of the PI coil itself. The larger the inductance, the longer the decay time to OV.

It is also critical that this high voltage spike doesn't result in oscillations, which can easily happen. Generally, the coil, for a PI, is made by first determining the desired inductance. Then the coil size or diameter selected. Once the two characteristics are determined, calculations are made to determine the required number of turns of wire to produce the calculated value of inductance.

Once built, the coil of wire is basically an inductor that has some internal resistance. However, because the coil consists of multiple windings, generally a number between 10 and 35, the windings produce a certain amount of capacitance between windings. This capacitance when combined with the inductance of the coil will create a "tuned circuit that will oscillate if additional circuitry isn't added to dampen or stop the oscillation. The basic damping device normally used is a resistor, generally called the damping resistor.

So, by carefully selecting the right resistor, a coil will produce a rapidly decaying voltage spike that doesn't ring or oscillate. If the resistor has too high a value, there will be some very minor oscillation, and if the resistor is too low in value, the spike voltage will take too much time dropping to the OV range.

One other critical part of a search coil that is seldom talked about is the shielding of the coil. Generally, coils have some form of a shield called a Faraday shield. The purpose of this shield is to minimize the capacitive effect between the coil and the ground, reduce static, and to absorb or reduce external noise. Like other factors, the shielding and the technique used, is quite critical. Too much or the wrong type of shielding can reduce sensitivity, especially to small objects such as gold nuggets. Too little shielding will allow other factors such as noise, signal variations due to the ground capacitance, etc to affect the signal. The shielding can also affect the decay time so it can affect the ability to detect small nuggets.

It should be noted that some manufacturers do not use any shielding at all. However, these detectors normally are designed for the detection of very large iron objects so any minor variations in noise or ground capacitance that normally affect very small non-ferrous objects such as small gold nuggets is not a problem. Such detectors normally operate with a very long delay before sampling. This long delay will cause most of the ground signal to be eliminated since it will decay much faster than a signal from a large iron object.

The technical information mentioned above is of little value to the average user of a PI. However, it can be important to anybody who wants to try to build a coil for their detector. The first rule of thumb when trying to build a different coil is to try to duplicate the electrical characteristics of the factory coil. By this I mean, one should try to keep the resistance the same as well as the inductance the same.

How is detection really done?

Both PI's and VLF's take a sample of the receive signal for analysis. In the case of the VLF, the receive signal sample is analyzed with respect to the transmit signal. By doing this, any signal "shift", commonly called phase shift, can be "seen". In other words, the sample is taken by syncing the sample to the transmit signal so the sample is always synchronized to the transmitter. The circuitry used to sample the received signal is normally called the synchronized demodulator.

On a PI, the signal from the coil is initially amplified and some time after the large current pulse is stopped, a sample of the amplified coil signal is taken. Since there is no transmitting going on at the time of the sample on a PI, timing is generally done by waiting a finite time after the termination of the large current pulse and then taking a sample. In this way, there is a form of synchronization also. The time between when the pulse quits and the sample is taken is often referred to the delay time. The delay time on most Gold Hunting PI detectors is 15 usec or less. A delay of 10 usec will show a distinct improvement, especially to very small gold in the few grain range over a detector having a delay of 15 usec.

This delay time is quite critical and is sometimes changed to create a crude form of discrimination, or rather reverse discrimination in the case of gold. As I mentioned before, the signal from gold can decay very quickly. In fact, the signal from most gold nuggets smaller than a 1/4 oz can decay in less than 50 usecs. If the delay is adjusted to 50 usec, then most small nuggets will be ignored, or phrased another way, will not produce any audio response. However, signals from objects made of iron, copper, silver or other highly conductive metal will normally still produce a strong signal. So, if a detector samples the signal at a time later than 50 usec or so, and this sample does not "see" a target, there is a good possibility the object is gold or some other type of low conductive material.

Since the analysis or sampling of this decaying signal is normally only done when the signal gets very near OV, any additional time to drop to the OV level will cause very small gold nuggets to be missed. The reason is because the reflected signal caused by the nugget is very brief and it combines with the normal signal from the coil.

If the nugget signal dissipates before the main signal decays to OV, then, when the sample is taken to determine whether an object is present, the signal from the small nugget will have already subsided and the nugget will be ignored.

Once a sample is taken, this sample voltage is held in suspension, for a better choice of words, until the next sample occurs, which adds to or subtracts from the previous sample. Because of the suspension, normally called sample and hold, and the filtering process built in to reduce noise, multiple samples are required before a true average signal is developed.

Once this average has leveled out, which normally takes a very brief time (in the thousandths or hundredth's of a second), any object that produces a change sufficient to be seen, will cause an additional signal that alters the receive sample average, which then causes the output to change or increase. This subsequent change is further amplified and ultimately is heard as an audio response, normally in a set of headphones.

The Up Side of PI's

Probably the biggest claim to fame of a PI is the additional depth that can be obtained. The key to this is the increased amount of power into the coil that can cause a stronger return signal from a buried object. However, even though there is a significant increase in current, the depth difference between a PI and a VLF isn't as dramatic as one might expect.

Many people question whether this depth advantage between a VLF and a PI is really that great in areas having almost no mineralization but overall, the PI appears to be superior simply because such places are few and far between. Where the PI really excels is in places having a much higher ground mineralization as will as locations where magnetite type hotrocks are common.

Next comes the big debate of just how a PI only using AA batteries can even come close to obtaining the depths of a different PI using a very large heavy duty battery .Obviously, the PI using the AA batteries cannot be pulsing with the same amount of current as a PI using a much bigger battery .

The fact is, the PI that uses AA batteries can approach the depth of other more powerful PI's, especially on gold less than an ounce in weight. There are multiple reasons this can be true. One reason is the fact that there because of a law of diminishing returns, which simply means it takes a whole lot of current to produce a very small depth increase because of shear power alone. As an example, it may take something like 4 amps of current to increase the depth 1 inch over a PI only pulsing with 1 amp, and this is only true if all other factors are equal.

One important factor that determines the sensitivity of the detector is sampling delay time. The sooner a sample can be taken, the stronger the signal that will be seen. In other words, it is quite possible to take a sample sooner and produce a stronger signal on a PI operating with less current than might be seen on a more powerful PI using much more current and having a longer delay. One simple way to allow earlier sampling is to reduce the coil current.

In other words, there are a whole lot of other factors that need to be taken into account to determine what is the best combination. Ah, but someone who just read the previous information might simply say, pulse with a strong signal and then simply sample sooner to make the best detector. Well, unfortunately, the stronger the pulse, the more difficult it is to sample sooner because of the reasons mentioned before. A longer pulse or lower coil resistance will result in more coil current, which will affect how long it takes for the spike to decay. A larger inductance will also result in a longer decay time. In fact, it becomes almost impossible to obtain the very short delay times when using a very strong pulse of long duration.

One way to help shorten the delay time of the decaying pulse is to reduce the number of turns of wire in the search coil. However, the field strength of the coil produced when current flows in the coil is a function of both the current and the number of turns, so reducing the number of turns also reduces the field strength produced. So any reduction in number of turns directly relates to potential depth loss.

If all this seems confusing, it is. Not only does the actual current have an effect, but the actual pulse length or time the current flows has an effect as mentioned before. Therefore, it is possible to pulse fewer times and use a shorter pulse and obtain very satisfactory results.

Pulse lengths of 50 usec or less will still produce a very decent signal from most of the gold nuggets that are found with a detector. Increasing the pulse length to 200 usec will really only have an impact on the signal coming from very large gold objects. The reason why the large increase in pulse time doesn't help on most smaller gold is simply because most of the smaller gold is fully saturated by the shorter pulse. Any additional pulse really does nothing to the potential signal that will come back form that gold object.

As a general rule though, a more powerful PI having a very long pulse will generally go deeper on very large gold, meaning nuggets weighing several ounces or more will be more readily detected to greater depths.

A detector using a high current short pulse will have a tendency to be more sensitive than a detector using less current. However, this difference normally is not dramatic, if at all. The key lies in early sampling and noise reduction.

One other major advantage of a Pl over a VLF is the fact that many of the hotrocks or black sand that make a VLF Scream will cause little or no signal on a PI. Normally, these intense hotrocks create a response on a VLF because of the magnetite within the rock.

A PI will seldom create a response to a magnetite rock or black sand due to the fact a magnetite hotrock or black sand signal will normally dissipate well before a sample is taken. However, an unbalanced earth field effect elimination can cause a hotrock to create some response, as can a very quick sample. In the case of some of the PI kits people build, where there is no earth field effect elimination, a long transmit pulse will cause a magnetite hotrock to produce a very strong signal, much like the signal from a metal object.

When a PI is pushed to the limits, even a PI will begin to produce a much louder signal on more and more magnetite type rocks just due to all the factors involved.

Where a PI really excels is in situations where a nugget is buried under or along side a magnetite hotrock. This combination spells disaster for most VLF's simply because the rock will generate a much stronger negative signal than the slight positive response from a piece of gold. So, it is very possible a VLF will miss a piece of gold in such a combination.

A PI, on the other hand, will look though the magnetite as if it isn't there, so a magnetite hotrock and gold combination will produce a desired signal. In some cases, the rock may just add a little positive signal thus causing the evasive gold target to be detected.

I do want to mention again that if the delay is extremely short, the earth field effect cancellation circuitry isn't perfect, or the rock or black sand contains other types of certain materials, a typical magnetite rock or black sand can generate a small but noticeable signal. One should also remember that it is quite likely that there are other members of the iron oxide family or even other metals in very small quantities that will cause a response and these other oxides may be present in the black sand or in a rock that appears to have large quantities of magnetite. So, any response a person gets from a rock my be caused by a combination of things.

Other hotrocks such as basalt or other similar hotrocks may cause a weak but noticeable signal much like a deep target. Fortunately, the signal from such rocks quickly subsides as the coil is raised a little. Thus, deeper hotrocks will seldom produce a signal.

Page 2


The down side of PI's

All detectors are subject to external noise, but PI's are extremely temperamental in this regard. The reason PI's are more sensitive to noise is due to the design of the preamp or first amplifier stage. For a PI to work at optimum this amplifier has to be built to amplify a very wide range of frequencies to assure the decay signal isn't altered. This type of amplifier is commonly called a broadband amplifier.

On a VLF there normally is one specific operating frequency and the preamp or first amplifier is somewhat tuned to that frequency. Thus, other signals such as noise will not be amplified as much.

On a PI, all signals are basically amplified the same, thus all noise, especially electrical noise is amplified just like the signal from a nugget. This problem is compounded by the fact that the coil itself will generate a very small voltage as it is moved or swept above the ground.

This very small voltage is the result of the coil moving through the earth's magnetic field and this extremely small voltage is often referred to as the earth field effect (EFE) signal.

For a PI to be very sensitive, there has to be a lot of amplification of the sampled signal. As such, even the small voltage produced by the coil movement will create a signal large enough to be heard. To eliminate this EFE problem, a second sample is normally taken much later in time and this second sample is effectively subtracted from the first or main sample. Since the earth field effect is a very slow signal, taking the later sample will eliminate most of the earth field effect signal, thus it is normally not heard.

However, this subtraction process is seldom perfect and because the sample is taken at a different point in time, there will always be a very slight response, that just might cause a very slight increase or decrease in a target response. Most likely this will occur or be noticed on the targets just on the threshold of detection. However, if for some reason the subtraction process is not near perfect, then it is quite possible that there will be a noticeable increase in signal strength when passing over a target from one direction than when passed over on the opposite direction of the coil swing.

The earth field effect is one of the reasons there may be a slight response at the end of a coil swing. The sudden stop of the coil before moving the opposite direction produces a much stronger EFE signal. One should also remember that it is almost impossible to totally eliminate the EFE so some of the minor responses mentioned should be considered normal.

One final noise problem is the noise generated within the detector itself. Just due to the nature of the circuitry used, a lot of noise is generated within the electronic circuitry .Much of this noise is "filtered" by the battery and aided by large capacitors and other filtering devices such as ferrite cores. However, no battery or capacitor is perfect so some noise always gets through.

The result of all the combined noises commonly creates a form of chatter or warble that can significantly reduce the sensitivity, especially to very small or deep objects producing very weak signals. In many cases, the noise may not even be really noticeable but be of sufficient amplitude to cause a reasonable depth loss.

Ground signals and Ground Balance

PI's are susceptible to many types of ground conditions, and, depending upon the type of ground, the sensitivity, and the delay, may generate a very strong signal due to the ground.

Terms like magnetic viscosity are used to explain just why certain types of ground can cause a strong response. Ground conditions having concentrations of maghemite will create very strong ground signals.

Areas having a clay base seem to produce strong ground responses also indicating that clay itself is part of the problem. Articles such as those written about geophysical research indicate that the clay problem can vary dramatically because of the type of clay as well as the moisture within the clay.

An article written by the Army Corps of Engineers indicates that clay will actually create a field that opposes the transmitted field of the PI and moisture enhances the ability for the clay to oppose the field due to the ionic behavior within the clay.

Coils, Coils, and more Coils

VLF's always have a transmit coil and a separate receive coil of wire in the search head. PI coils, however, can be produced in several variations. If the same coil is used for both the transmit and the receive signal, the coil is normally called a "MONO" coil.

If two sets of coil windings are used, and those coils are basically the same size and shape with one coil used as a transmit and the other the receive, and they overlap a small amount on one side, the coil is generally called a DD coil. The name DD generally refers to the design of the coils where they are sort of like D's with one D reversed and the backs of the D's overlapping slightly. This overlap area is the main detection zone and is the area where an object is under at least part of both coils at the same time. This detection zone is most noticeable on deeper objects.

By nature, DD coils are somewhat less sensitive when compared to a mono coil of the same size. One reason for the reduction in sensitivity is the fact that the DD electrical coil windings are smaller in size than the coil windings of a mono coil even they may have the same size coil housing.

One other key factor that is important is the fact to remember about a DD coil is the main detection zone is quite narrow. This narrow detection zone, normally at or near the overlap will create a very brief or narrow signal when compared to the signal on a mono coil. This situation makes the sweep speed of the search coil much more critical. Swinging the coil too fast can easily cause a very weak object to be missed simply because the signal is so short and the circuitry filtering used to eliminate the noise will also almost eliminate such a signal.

The fact that DD coils have smaller diameter windings for the transmit and receive coil, when compared to a mono coil using the same size housing, has some advantages. Generally, the smaller receive coil is not as good of an antenna as a larger mono coil, thus less noise is detected and amplified. As a result, the detector can be much quieter when using a DD coil. In many cases the reduction in noise can outweigh the depth loss due to the size difference.

One other major asset of a DD coil is the fact the receive coil is isolated from the transmit coil. This helps in the fact that any low level noise that is generated by the transmit circuitry during the sampling time is isolated from the receive circuitry. This isolation therefore reduces the combined noise that can negatively affect a target response.

One final advantage of a DD coil is, by nature, a DD coil partially cancels the ground signal. If the coils are properly aligned or positioned, most ground signal in the receive coil is eliminated. This results in a detector that has very little ground response, yet still responds with a strong signal from a buried object.

Another type of coil that is made for PI's is called the figure 8 or "Salt" coil. In this design, there is a large transmitting coil and two receive coils that are wired such that the receive coils are opposite of each other, meaning that one coil will produce a positive signal and the other a negative signal. On this type of coil, it is quite common to build a larger receive coil, elongate it, pinch the center of the elongation and then twist one half of the receive coil one half turn to create two coils, much like a figure 8. As mentioned before, this type of receive coil is also called a "figure 8 coil" just due to how it is constructed.

One advantage of a "salt" coil having a large transmit coil and two smaller receive coils is the design is both ground canceling and noise canceling. The ground canceling relies on the principle that both receive coils are equally spaced from the ground for maximum ground signal elimination.

The disadvantage of the Salt or figure 8 coil is there is a depth loss that occurs when compared to a similar sized mono coil or even a similar sized DD coil. Part of the reason for the depth loss is the fact the two receive coil signals basically oppose each other since one will produce a positive receive signal and the other will produce a negative signal. This opposition will cause some receive signal to be eliminated.

Because the signals from two receive coils have a tendency to cancel each other, any noise detected by the two coils basically is also canceled. This cancellation process has one other advantage and that is, it will eliminate the earth field effect.

A little different figure 8 coil can be built where there is only one coil used as both the transmit coil and the receive coil. This coil is again, elongated, pinched, and one half of the coil is twisted over so half of the coil is transmitting up when the other coil is transmitting down. This type of coil eliminates or cancels external noise extremely well but does not ground cancel. Since the two xmit coils are much smaller, there is also some depth loss on this type of coil because of the size of the coils as well as the signal from the two halves of the coil have a tendency to cancel each other. As such, the signal from a buried object will be the greatest when it is centered or near centered under either of the two coils and the weakest when the object is right at the crossover point of the two coils.

One of the most common coils found on a VLF is something called a concentric coil. In this case there generally is a large transmit coil and a smaller receive coil basically centered in the large coil. For this type of coil to really work correctly on a VLF, there will be an additional transmit coil wound directly on the smaller receive coil, but will wound opposite to the main transmit coil. The purpose of the smaller transmit coil is to cancel any signal in the receive coil caused by the larger transmit coil A concentric coil design can be used for a PI, but it is rare to find one..

Of course, there can be variations of the above coils, meaning they can be rectangular, round, oval, or any other shape a person should desire. Also, the windings can be changed or possibly additional windings can be incorporated to produce the desired results. So, the ultimate design of a search coil is left to the imagination of the designer.

Finally, some mention has to be made regarding coil size. The coil size of most PI's normally ranges from an 8" diameter coil to greater than 3 feet in diameter. It is quite common to hear of a person using an 18" diameter coil, but the most popular sizes range from 11" to about 14".

Recently, Eric Foster posted some interesting findings regarding the general detection ranges of different sized coils versus target or object size. This information can be viewed on the PI forum and displayed on September 16, 2002. Several discussions occurred during that time pertaining to the depth verses coil diameter, versus target size.

As one might expect, the larger the coil, the deeper one may find objects. However, it is quite possible a smaller coil will find an object deeper than a large coil, especially if the object is small. Contrary to the some of the discussion that resulted on the above mentioned forum, there is a more direct relationship between the size of the coil, size of the object and the ideal maximum depth such an object can be detected. An error in calculations led to some information being incorrectly noted.

One should realize that information such as what Eric Foster posted is generally theoretical and as such is subject to some distortion in the real world. However, as a rule, the general principle is quite accurate.

When searching for information about depth or size of objects that can be found with different size coils, extreme cases always show up. For example, many people have found extremely small nuggets ranging in the few grain range with an 18" coil. Normally such a large coil will not be able to see such a small target at any depth, or even in the middle of the coil if the nugget is small enough. However, this small nugget can produce a signal if it is very near the coil windings themselves.

One final point that I am sure will cause controversy and that is a smaller coil will not show as dramatic increase to sensitivity to small gold on a PI like it does on a VLF. The reason, again, lies in the fact that the sensitivity to small gold on a PI, is much more dependent upon the delay before sampling than it is on the coil itself.

Ground Balance Differences There is a world of difference in ground balancing techniques between a VLF and a PI. On a VLF, a sample can be taken such that the signal from the ground appears to be eliminated. Actually, it is still there, but by the sampling at the right time half of the signal is positive and the other half negative, so the net effect is 0.

On a PI, no such condition can exist because of the fact the transmit time is separate from the receive time. So, another method has to be used. One common method is to take advantage of the fact the ground signal lasts for a long time. So, if the initial sample is taken to look for a target and then a later sample is taken that still contains ground signals and this later sample is amplified, and then subtracted from the first or main sample, the ground signal can be minimized, thus leaving the target signal.

Unfortunately, any subtraction process also reduces the signals of targets also having a long decay. As it turns out, some gold signals will be very similar to the ground signal, so this subtraction process can effectively reduce the response from some gold objects.

In the case of larger gold objects, the subtraction process can actually cause the signal to change from an increasing response to a decreasing response, meaning a piece of gold would create a negative signal. This negative signal can easily be "rectified" much like the rectifier in any other circuit. The rectification process will then make the large gold also respond with a positive signal, rather than a negative signal.

However, as mentioned before, some gold will respond much like the ground so there will be some gold objects that will be cancelled much like the ground signal To overcome this problem, different length pulses can be used and multiple subtraction processes incorporated.

If a longer pulse is also used, the longer pulse alters the ground signal characteristics. This alteration is sufficient that it requires a different ratio of subtract signal to cancel the ground response. This change in subtraction level then changes which gold might be eliminated. As such, any gold that might be eliminated by only using a short pulse will produce a strong signal when using a long pulse, and visa versa. The result is most gold will be detected quite strongly if pulses of different duration are used. However, in all cases, many of the larger nuggets that have a decay lasting longer than the time when the ground signal sample is taken will also be reduced in signal strength.

Another technique can be used for ground balance but it is generally used with a DD coil. In this case, a sample is done during the pulse on time as well as the pulse off time. The two different samples produce different signals, which then can be combined to minimized the ground response. This type of sampling can also be used to produce a better form of discrimination which would be much more accurate.

One technique that can be used is a variation of the first ground balance technique. This method just minimizes the ground response using the subtract method. By doing this, the ground signal is minimized significantly but gold responses are not eliminated. Some nugget responses are, however, reduced in signal strength and as such, there is some depth loss. This normally occurs with nuggets greater than 2 gram or so.

Finally, Discrimination on a PI

Due to the nature of the signals caused by different objects, it is extremely difficult, or put another way, almost impossible to build a good discriminating PI.

Since the time it takes for a target signal to decay can vary because of the size, shape, and chemical makeup of the object, then any type of later sampling will not produce a reliable form of discrimination.

Many PI's rely on the ability of an adjustable delay whereby the operator can simply adjust the delay longer to see if an object is a piece of gold or not. If the delay is increased and the signal from an object disappears, then the operator can assume the object is made of a lower conductive material such as gold. This is acceptable for those hunting something like gold rings, but does not work well on gold nuggets. larger gold nuggets can produce a much longer delay, so any attempt to use this delay technique will result in one thinking a large gold nugget to be junk.

Another concept used on a PI for discrimination is to sample during the "on" time of the pulse. Any target will produce a slight change in the signal seen at that time as well as a change when the normal target sample is taken.

If the analysis is done correctly, then one can use both the "pulse on" and "pulse off" signals and get a better analysis of a target. This type of design can lead to a better form of discrimination. However, few if any PI's are actually using this technique.

Regardless of the technique used, no form of discrimination is perfect, and, most likely, never will be. Some techniques are better than others, but all can be fooled, and this is true of both PI's and VLF's.

By Reg Sniff
Copyright 2003 The author reserves all distribution rights.