A
Anonymous
Guest
I can
Find's Treasure Forums
Welcome to Find's Treasure Forums, Guests!
You are viewing this forums as a guest which limits you to read only status.
Only registered members may post stories, questions, classifieds, reply to other posts, contact other members using built in messaging and use many other features found on these forums.
Why not register and join us today? It's free! (We don't share your email addresses with anyone.) We keep email addresses of our users to protect them and others from bad people posting things they shouldn't.
Click here to register!
Need Support Help?
Cannot log in?, click here to have new password emailed to you
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
A
Anonymous
Guest
Eric
This is where my lack of practical PI experience holds me back. So far the only kind of target I have modeled is the simplest case such as a thin ring of non ferrous metal where the skin depth is greater than the thickness of the target. I would also like to be able model other common objects, but I don't know what their signals look like.
It looks like most of the signals in your plot could be reasonably approximated by [k*e^(-t/tc1) + (1-k)*e^(-t/tc2)] where the first term gives the steep decay at the beginning and the second term gives the long tail. K and tc2 could be calculated from the straight line part of the curve on the log graph, and tc1 could be picked to best fit the beginning of the decay. The copper plate does not even straighten out in the first 400 usec, and it would be harder to fit.
But what about iron? I don't know anything about its waveform. Does anyone have a library of decay curves for common iron objects? I don't see how anyone can even consider discrimination without having good models for different classes of targets.
And then there are the ground signals. I don't know what they look like either.
Robert
This is where my lack of practical PI experience holds me back. So far the only kind of target I have modeled is the simplest case such as a thin ring of non ferrous metal where the skin depth is greater than the thickness of the target. I would also like to be able model other common objects, but I don't know what their signals look like.
It looks like most of the signals in your plot could be reasonably approximated by [k*e^(-t/tc1) + (1-k)*e^(-t/tc2)] where the first term gives the steep decay at the beginning and the second term gives the long tail. K and tc2 could be calculated from the straight line part of the curve on the log graph, and tc1 could be picked to best fit the beginning of the decay. The copper plate does not even straighten out in the first 400 usec, and it would be harder to fit.
But what about iron? I don't know anything about its waveform. Does anyone have a library of decay curves for common iron objects? I don't see how anyone can even consider discrimination without having good models for different classes of targets.
And then there are the ground signals. I don't know what they look like either.
Robert
A
Anonymous
Guest
Hi Robert,
Here are some log plots of ferrous objects. Starting from the left is a 3in x 4in piece of tin plated steel. Next is a 4in diameter x 3in high coffee tin. 3rd is a 2in masonry nail. 4th a 4in long adjustable spanner and 5th is a 3in nail. As you can see, ferrous objects give a wide variety of decays. The tin plated steel and coffee tin behave more like non-ferrous objects because of their large surface area and being directly under the coil. When the TX field is at right angles to the plane of the surface, the predominant response is the conductivity of the steel. If the steel plate was tipped vertical to the coil so that the field was parallel to the plate the response would change to be more like 4 and 5. A possible method of object identification and ferrous/non-ferrous discrimination is to have a time constant readout on the detector. With non-ferrous objects, such as coins and rings, the TC stays the same as you scan over it and also from whichever direction you scan. With a nail, or even a flat bit of steel, the TC changes as the orientation of the coil field with respect to the object changes. The TC of a nail will go from long to short to long, as you scan along its length. Scan at right angles and you get one short TC due to the nail
Here are some log plots of ferrous objects. Starting from the left is a 3in x 4in piece of tin plated steel. Next is a 4in diameter x 3in high coffee tin. 3rd is a 2in masonry nail. 4th a 4in long adjustable spanner and 5th is a 3in nail. As you can see, ferrous objects give a wide variety of decays. The tin plated steel and coffee tin behave more like non-ferrous objects because of their large surface area and being directly under the coil. When the TX field is at right angles to the plane of the surface, the predominant response is the conductivity of the steel. If the steel plate was tipped vertical to the coil so that the field was parallel to the plate the response would change to be more like 4 and 5. A possible method of object identification and ferrous/non-ferrous discrimination is to have a time constant readout on the detector. With non-ferrous objects, such as coins and rings, the TC stays the same as you scan over it and also from whichever direction you scan. With a nail, or even a flat bit of steel, the TC changes as the orientation of the coil field with respect to the object changes. The TC of a nail will go from long to short to long, as you scan along its length. Scan at right angles and you get one short TC due to the nail
A
Anonymous
Guest
Eric
What is the vertical scale on these plots? It looks like it is somewhere from 2 to 3 decades full scale, but I am going to try to match these curves so I would like to know more precisely.
Robert
What is the vertical scale on these plots? It looks like it is somewhere from 2 to 3 decades full scale, but I am going to try to match these curves so I would like to know more precisely.
Robert
A
Anonymous
Guest
Eric
I don't think the fact that thin flat iron looks like a conductive target when you are directly above it should cause any more problem than it does for a VLF. A tin lid can easily look like a coin to a VLF detector when the coil is above it. It's when the coil is approaching or leaving it that you get the iron signal.
It does bother me that this signal 5 looks so much like the thick copper plate.
It seems to me that the spread in time constants during the sweep would be more telling than the shape of one decay curve. Also, it only takes two samples on the curve to get an estimate of the time constant, but it takes 3 to measure the curvature and the third sample is going to have a worse signal to noise ratio.
Now I am starting to see why you PI guys don't do much discrimination. In VLF we measure the signal while the transmitter is running. A conductive target gives a signal that is out of phase with the transmitted signal. A ferrous target gives a signal that is in phase with the transmitter. That is a wide spread that is easily detected. In PI you measure the signal when the transmitter is off and you lose that information. I just realized that while I was writing this post.
I thought it was a really neat trick that you read the signal while the transmitter was off, and that you were getting a better signal to noise ratio for free. But it was not free, you paid a big price for it. I may not cross over to the dark side after all.
Robert
I don't think the fact that thin flat iron looks like a conductive target when you are directly above it should cause any more problem than it does for a VLF. A tin lid can easily look like a coin to a VLF detector when the coil is above it. It's when the coil is approaching or leaving it that you get the iron signal.
It does bother me that this signal 5 looks so much like the thick copper plate.
It seems to me that the spread in time constants during the sweep would be more telling than the shape of one decay curve. Also, it only takes two samples on the curve to get an estimate of the time constant, but it takes 3 to measure the curvature and the third sample is going to have a worse signal to noise ratio.
Now I am starting to see why you PI guys don't do much discrimination. In VLF we measure the signal while the transmitter is running. A conductive target gives a signal that is out of phase with the transmitted signal. A ferrous target gives a signal that is in phase with the transmitter. That is a wide spread that is easily detected. In PI you measure the signal when the transmitter is off and you lose that information. I just realized that while I was writing this post.
I thought it was a really neat trick that you read the signal while the transmitter was off, and that you were getting a better signal to noise ratio for free. But it was not free, you paid a big price for it. I may not cross over to the dark side after all.
Robert
A
Anonymous
Guest
PI has its advantages and its disadvantages like alot of things. Some of the advantages I like are that it doesn't reguire the ridge geometry especially monoloops, that IB requires. Don't have to get the coils nulled perfectly (which saturates the integrators in sensitive IB) since transmitter is off. Easier to get more power put into the whole system and look for eddy current decay of objects, verses turning up the power on IB/vlf and basically lighting up a whole reactive hillside.
On the downside as Dave Johnson keeps indirectly pointing out is the reactive signal isn't there for doing discrimination. Only the resistive. This means you can't do things like the arctangent of the reactive over resistive and get phase. Your time constant chart kinda looks like an arctangent.
Keeping this short if you look at the market, PI wins underwater, on the salty beachs, big gold nuggets deep down or in Aussieland, big deep treasure cache. Until something breaks VLF/IB wins on small gold, coin shooting, iron rejection, discrimination in general. So depends on what you are up to.
JC
On the downside as Dave Johnson keeps indirectly pointing out is the reactive signal isn't there for doing discrimination. Only the resistive. This means you can't do things like the arctangent of the reactive over resistive and get phase. Your time constant chart kinda looks like an arctangent.
Keeping this short if you look at the market, PI wins underwater, on the salty beachs, big gold nuggets deep down or in Aussieland, big deep treasure cache. Until something breaks VLF/IB wins on small gold, coin shooting, iron rejection, discrimination in general. So depends on what you are up to.
JC
A
Anonymous
Guest
It's possible to use an IB loop with PI and demodulate the reactive for whatever purpose you want it. I and other people have done it; I think that Eric Foster at one time marketed such a machine. Of course, it's a compromise-- you no longer have the freedom to make a searchcoil by winding lamp cord around the lid of a laundry basket.
The worst thing about free lunches is that one always has to pay something to get them.
--Dave J.
The worst thing about free lunches is that one always has to pay something to get them.
--Dave J.
A
Anonymous
Guest
If you can get the reactive information then you can discriminate. Problem solved. So now what is wrong with PI?
JC
JC
A
Anonymous
Guest
What yourself and others have been doing the last couple weeks is taking PI out of its historical niches and putting it into the mainstream where it can compete heads up in areas where VLF/MF has previously reigned supreme.
It's been a lot of fun, whaddaya say?!!
--Dave J.
It's been a lot of fun, whaddaya say?!!
--Dave J.
A
Anonymous
Guest
Dave
OK, now that I understand why discrimination is difficult in PI, I need to know some things about iron that I never cared about before. In VLF there is a very large effect that allows discrimination between ferrous and non-ferrous, but that effect is only observable when the coil di/dt is non zero. With PI if you do the sampling while di/dt is 0 you do not see that large effect, and you have to look for more subtle things. I wondered why you were so obsessed with magnetic properties, now I can see.
If I have an IB coil, for test purposes, and drive a triangular current through it, and put a lossless piece of ferrite near it, I will see a square wave on the receive coil. The received signal will be proportional to di/dt and it will switch quickly and cleanly when di/dt changes. With iron it will not be as simple. There will be another signal from the eddy current that I want to ignore for now. I just want to focus on the reactive part of the signal. For reasons that you have stated in other posts, the magnetic field from the iron will not exactly follow the coil field. What I want to know is what kind of time frame are we talking about for common iron objects at normal PI field strengths and room temperature? I never had to worry about this before so I am completely in the dark. If I could separate out the reactive signal and look at it on a scope, what would I see for a 5 kHz triangular current?
Robert
OK, now that I understand why discrimination is difficult in PI, I need to know some things about iron that I never cared about before. In VLF there is a very large effect that allows discrimination between ferrous and non-ferrous, but that effect is only observable when the coil di/dt is non zero. With PI if you do the sampling while di/dt is 0 you do not see that large effect, and you have to look for more subtle things. I wondered why you were so obsessed with magnetic properties, now I can see.
If I have an IB coil, for test purposes, and drive a triangular current through it, and put a lossless piece of ferrite near it, I will see a square wave on the receive coil. The received signal will be proportional to di/dt and it will switch quickly and cleanly when di/dt changes. With iron it will not be as simple. There will be another signal from the eddy current that I want to ignore for now. I just want to focus on the reactive part of the signal. For reasons that you have stated in other posts, the magnetic field from the iron will not exactly follow the coil field. What I want to know is what kind of time frame are we talking about for common iron objects at normal PI field strengths and room temperature? I never had to worry about this before so I am completely in the dark. If I could separate out the reactive signal and look at it on a scope, what would I see for a 5 kHz triangular current?
Robert
A
Anonymous
Guest
With IB wound coils, continuous current through the transmit coil, demodulation of the reactive and resistive components of the "return" signal we are getting close to inventing a VLF metal detector.
I guess I didn't really answer Robert's question very well, here is the bright side of PI. With IB wound coils (real ones, not theory) it will be very hard to null them (unless wound in one plane, and there is still capacitive coupling between the transmit and receive coils to deal with {different phase}) at the currents of strong PI and at the high receiver gains of PI, the receiver is saturated. Therefore, alot is sacrificed, true enough, but the advantage is stronger transmitter and higher gain in receiver (since xmitter is off) leading to greater detection range.
Monos and double d's are also easier to build. And null on DDs doesn't have to be perfect either as long as you protect the receiver as usual.
So if you can figure out true discrimination and ground balance with "normal PI" then you can have your cake, free lunch, etc.
JC
I guess I didn't really answer Robert's question very well, here is the bright side of PI. With IB wound coils (real ones, not theory) it will be very hard to null them (unless wound in one plane, and there is still capacitive coupling between the transmit and receive coils to deal with {different phase}) at the currents of strong PI and at the high receiver gains of PI, the receiver is saturated. Therefore, alot is sacrificed, true enough, but the advantage is stronger transmitter and higher gain in receiver (since xmitter is off) leading to greater detection range.
Monos and double d's are also easier to build. And null on DDs doesn't have to be perfect either as long as you protect the receiver as usual.
So if you can figure out true discrimination and ground balance with "normal PI" then you can have your cake, free lunch, etc.
JC
A
Anonymous
Guest
Robert:
The reactive part of the signal is the part which is in phase with the triangular current; which, when differentiated by the (separate, preferably IB) receive coil, becomes a rectangle.
If you could subtract that rectangle out, then what would be left is the resistive components which are at 90 degrees with respect to the current. Since we're not talking a single sinusoid here, but a superposition of sinusoids, the signal you'd see in the time domain would a bit of a mess. And, it would be highly dependent on the shape and orientation of the iron.
One way to answer this question, is to get hold of a Fisher Impulse, which transmits pulses which have a triangular current waveform, make an air coil several inches in diameter, hook it up to a 'scope with a suitable damping resistor in parallel, slide it over the edge of the Impulse loop until you find a position of induction balance, and then bring a piece of iron into the field.
This technique is not limited to the Impulse, but I mention the Impulse because your question related to triangular current waveforms.
If you were thinking of a continuous triangular current waveform, Fisher CZ's do that. Same trick can be used with a CZ to see its response in the time domain on a 'scope.
--Dave J.
The reactive part of the signal is the part which is in phase with the triangular current; which, when differentiated by the (separate, preferably IB) receive coil, becomes a rectangle.
If you could subtract that rectangle out, then what would be left is the resistive components which are at 90 degrees with respect to the current. Since we're not talking a single sinusoid here, but a superposition of sinusoids, the signal you'd see in the time domain would a bit of a mess. And, it would be highly dependent on the shape and orientation of the iron.
One way to answer this question, is to get hold of a Fisher Impulse, which transmits pulses which have a triangular current waveform, make an air coil several inches in diameter, hook it up to a 'scope with a suitable damping resistor in parallel, slide it over the edge of the Impulse loop until you find a position of induction balance, and then bring a piece of iron into the field.
This technique is not limited to the Impulse, but I mention the Impulse because your question related to triangular current waveforms.
If you were thinking of a continuous triangular current waveform, Fisher CZ's do that. Same trick can be used with a CZ to see its response in the time domain on a 'scope.
--Dave J.
A
Anonymous
Guest
Dave
I did not word my question well.
Say I put an iron object under the coil and record the result. Then I take a non-ferrous object of the same size, shape, orientation, and conductivity as the iron, put it same place under the coil and record the result. Then I appropriately scale the second result and subtract it from the iron signal. That gives me the difference I am looking for. The largest part of that difference should be the rectangular reactive part. But that part would not be observable if I wanted to level off the current somewhere to take a quiet measurement. So also subtract off the rectangle. Now what is left? I assume that if the iron has some losses beyond the I^2 R loss that there will be some other difference in the signals that will show up at this point. Is there anything distinctive that might be used for discrimination purposes, or is it going to be insignificant enough to get lost in the noise? My fear at this point is that it will not be significant enough.
I chose a triangular wave because I thought it was easy to describe and visualize, and it has a lot of frequencies. If I use a sine wave the reactive part is a sine wave in phase with the coil current. After I subtract that off I get a phase shifted sine wave that looks just like a conductive object. Unless there is a frequency dependence I can use, I will not be able to tell the difference between a ferrous and non-ferrous object.
The only coils I have are White's with the built in capacitor. They are not suitable for this kind of test because I can only use them at one frequency.
Robert
I did not word my question well.
Say I put an iron object under the coil and record the result. Then I take a non-ferrous object of the same size, shape, orientation, and conductivity as the iron, put it same place under the coil and record the result. Then I appropriately scale the second result and subtract it from the iron signal. That gives me the difference I am looking for. The largest part of that difference should be the rectangular reactive part. But that part would not be observable if I wanted to level off the current somewhere to take a quiet measurement. So also subtract off the rectangle. Now what is left? I assume that if the iron has some losses beyond the I^2 R loss that there will be some other difference in the signals that will show up at this point. Is there anything distinctive that might be used for discrimination purposes, or is it going to be insignificant enough to get lost in the noise? My fear at this point is that it will not be significant enough.
I chose a triangular wave because I thought it was easy to describe and visualize, and it has a lot of frequencies. If I use a sine wave the reactive part is a sine wave in phase with the coil current. After I subtract that off I get a phase shifted sine wave that looks just like a conductive object. Unless there is a frequency dependence I can use, I will not be able to tell the difference between a ferrous and non-ferrous object.
The only coils I have are White's with the built in capacitor. They are not suitable for this kind of test because I can only use them at one frequency.
Robert
A
Anonymous
Guest
The decay curve of metallic iron in a PI is fairly close to that of a "medium-conductivity object". The phrase in quotes is a weasel-word: what it means depends on the context of the particular machine in question.
The actual response of iron is highly dependent on its shape and orientation relative to the coil. However, in a given machine, there will normally be a curve below which you find low conductivity nonferrous with very little iron, and another curve above which you find only high conductivity nonferrous.
Many years ago I built a discriminator PI which knocked out everything in the pulltab range and below, while finding coins (except nickels), and it was balanced for the "red stuff" too. In high mineralization where VLF discriminators were nearly useless, with this unit I didn't dig any trash I didn't want to. Weighed about 2 1/2 pounds, ran on a zinc-carbon 9 volt battery-- a real sweetheart of a machine. Not the basis of a product though, because it didn't have enough "air sensitivity". Since it knocked out medium and low conductivity stuff, it wouldn't have been any good for relic hunting or for coinage other than U.S.A. If I'd have rectified the signal, then the iron would have come back.
Anybody who plays with PI for a while runs into this same set of problems. You can get really fancy analyzing decay curves, but as you point out, whatever systematic differences may exist between iron and nonferrous are so small that they tend to get lost in noise and ground interference. You can demodulate flyback time to get information from that, but again that's a weak signal which is affected by ground.
Of course you can use an IB loop and demodulate the imbalance signal to get the reactive; then you have a nice strong iron ID signal which has been discombobulated by soil minerals.
I'm not saying that there is no satisfactory solution to the problem of iron. I post what I can, but because I'm in this business there's some stuff I can't talk about.
--Dave J.
The actual response of iron is highly dependent on its shape and orientation relative to the coil. However, in a given machine, there will normally be a curve below which you find low conductivity nonferrous with very little iron, and another curve above which you find only high conductivity nonferrous.
Many years ago I built a discriminator PI which knocked out everything in the pulltab range and below, while finding coins (except nickels), and it was balanced for the "red stuff" too. In high mineralization where VLF discriminators were nearly useless, with this unit I didn't dig any trash I didn't want to. Weighed about 2 1/2 pounds, ran on a zinc-carbon 9 volt battery-- a real sweetheart of a machine. Not the basis of a product though, because it didn't have enough "air sensitivity". Since it knocked out medium and low conductivity stuff, it wouldn't have been any good for relic hunting or for coinage other than U.S.A. If I'd have rectified the signal, then the iron would have come back.
Anybody who plays with PI for a while runs into this same set of problems. You can get really fancy analyzing decay curves, but as you point out, whatever systematic differences may exist between iron and nonferrous are so small that they tend to get lost in noise and ground interference. You can demodulate flyback time to get information from that, but again that's a weak signal which is affected by ground.
Of course you can use an IB loop and demodulate the imbalance signal to get the reactive; then you have a nice strong iron ID signal which has been discombobulated by soil minerals.
I'm not saying that there is no satisfactory solution to the problem of iron. I post what I can, but because I'm in this business there's some stuff I can't talk about.
--Dave J.
A
Anonymous
Guest
I traced the curves of some of the ferrous and non ferrous objects then re-scaled them horizontally to normalize for time. The red curves are iron, the other curves are nonferrous. The leftmost curve is the copper plate.
As far as shape goes the irons fall in-between the non ferrous curves.
Robert
As far as shape goes the irons fall in-between the non ferrous curves.
Robert
A
Anonymous
Guest
In general, when you plot iron, it will resemble the high-conductivity nonferrous extreme more closely than it does the low-conductivity nonferrous extreme. Some PI's have a timing sequence that reaches a curve limit at a particular level of high conductivity, and then "folds back" some as you go even higher in conductivity: this can complicate target identification.
I believe, though I have not proven it, that if you really stretch the receive period out, you'll find the tail on iron targets hanging in there better than most ordinary high-conductivity nonferrous, such as large coins. Big masses of nonferrous metal, however, would probably hang in there right with the iron.
Somebody out there has probably already done this and is free to talk about it. Next post, please?
--Dave J.
I believe, though I have not proven it, that if you really stretch the receive period out, you'll find the tail on iron targets hanging in there better than most ordinary high-conductivity nonferrous, such as large coins. Big masses of nonferrous metal, however, would probably hang in there right with the iron.
Somebody out there has probably already done this and is free to talk about it. Next post, please?
--Dave J.
A
Anonymous
Guest
If we could stretch the receive time out to Alice Springs and back, then we could tell all kinds of stuff about what kind of target. There are two reasons why this is not practical.
First of all, if you don't have very many samples, you've either sacrificed sensitivity, or you have to integrate over a long time, resulting in really sluggish response.
Second, if you help out the high-conductivity stuff by stretching out the transmit time (here we go again on that transmit vs. flyback stuff?), you're driving the coil into its low-Q region, and keeping it there longer, so power consumption skyrockets.
In VLF detectors, US coinage other than nickels is best seen at frequencies about 2-3 kHz and below. But, if you run a beeper down in that range, the need to have reasonable battery life means you sacrifice transmit horsepower, and the air sensitivity isn't very good. That's the single most important reason why most U.S.A. coinshooting VLF machines run in the 4-8 kHz range.
Summarizing:-- the sluggishness of high-conductivity and iron targets runs contrary to our need to gather plenty of data quickly and to keep power consumption reasonable. This means that our ability to detect and ID the big stuff almost always gets compromised.
--Dave J.
First of all, if you don't have very many samples, you've either sacrificed sensitivity, or you have to integrate over a long time, resulting in really sluggish response.
Second, if you help out the high-conductivity stuff by stretching out the transmit time (here we go again on that transmit vs. flyback stuff?), you're driving the coil into its low-Q region, and keeping it there longer, so power consumption skyrockets.
In VLF detectors, US coinage other than nickels is best seen at frequencies about 2-3 kHz and below. But, if you run a beeper down in that range, the need to have reasonable battery life means you sacrifice transmit horsepower, and the air sensitivity isn't very good. That's the single most important reason why most U.S.A. coinshooting VLF machines run in the 4-8 kHz range.
Summarizing:-- the sluggishness of high-conductivity and iron targets runs contrary to our need to gather plenty of data quickly and to keep power consumption reasonable. This means that our ability to detect and ID the big stuff almost always gets compromised.
--Dave J.
A
Anonymous
Guest
Dave
Yes, that agrees with this data. These pieces of iron fell between the slowest coin in Eric's data and the thick copper plate. The next faster nail (which I did not show) approximately matched the slow coin. Unfortunately I could not read what the coin was, so I don't know what size we are talking about. And I still don't know the vertical scale on the log plots, so I could not calculate the time constant.
Robert
Yes, that agrees with this data. These pieces of iron fell between the slowest coin in Eric's data and the thick copper plate. The next faster nail (which I did not show) approximately matched the slow coin. Unfortunately I could not read what the coin was, so I don't know what size we are talking about. And I still don't know the vertical scale on the log plots, so I could not calculate the time constant.
Robert
A
Anonymous
Guest
Dave Robert and all,
All non-ferrous objects decay as a single exponential after one time constant, whereas the viscous signal in iron decays as log t. This means that large, conductive non-ferrous items, like the copper plate in my plots, may not reach 1TC for several hundred microseconds, or even milliseconds. If you compare the decay of a 3in nail, to that of the copper plate from 50uS to 400uS, they look much about the same. However, if you were to go to later times, the log plot of the copper would end up as a straight line, while that of the nail would still be curved. They would stay like this right into the noise level. The other problem is that objects with long TC
All non-ferrous objects decay as a single exponential after one time constant, whereas the viscous signal in iron decays as log t. This means that large, conductive non-ferrous items, like the copper plate in my plots, may not reach 1TC for several hundred microseconds, or even milliseconds. If you compare the decay of a 3in nail, to that of the copper plate from 50uS to 400uS, they look much about the same. However, if you were to go to later times, the log plot of the copper would end up as a straight line, while that of the nail would still be curved. They would stay like this right into the noise level. The other problem is that objects with long TC
A
Anonymous
Guest
I tried the same basic method that Eric used in his PPD1 detector. When a balanced coil such as a dual D is used, the mostly reactive signal received during the transmit period will go sharply positive and then negative for a conductive target. A negative only signal results from a ferrous target.
I used a sample and hold circuit with a 1uS sample time which sampled at the positive edge of the transmit pulse. A second 1uS sample was taken slightly later, but still during the transmit period.
I used a more linear method than a comparator to control the feedback speed of this reactive or "X" signal. This provided a very smooth operation and a good discrimination ability.
OK, The signal during the transmit pulse although mostly reactive does contain some resistive component. The signal taken after a delay from the transmit pulse is almost purely resistive.
This is true for a beach detector. A detector designed for use in heavily mineralized ground will need a ground balance control which will require more work but is quite possible to do.
OK, it must be obvious to the VLF guru's reading this that we now have a demodulated "R" or resistive component and a seperate "X" or reactive component available the same as a modern VLF detectors synchronous detectors provide .
By varying the ratio of "R" and "X" to a summing or a difference amplifier, you can vary the discrimination point the same way a VLF/TR does. It is even possible to add notch discrimination!
The detector worked fine. Discrimination was OK in both the air and beach sand. I have started work on the ground balance control to allow discrimination in highly mineralized ground.
Standard VLF motion disc circuits should also work OK. This is due to the fact that we have seperate "R" and "X" signals to work with.
Due to some strange happenings at my work, I have not been able to do anything on the detector for the last few weeks. I am in the process of setting up a second lab in my home.
I just bought a dual channel 200MHz oscilloscope, a synthesized function generator, A Lambda triple output, variable lab supply, and a bunch of other good stuff today. I will soon be back in operation. All the best, Dave. * * *
I used a sample and hold circuit with a 1uS sample time which sampled at the positive edge of the transmit pulse. A second 1uS sample was taken slightly later, but still during the transmit period.
I used a more linear method than a comparator to control the feedback speed of this reactive or "X" signal. This provided a very smooth operation and a good discrimination ability.
OK, The signal during the transmit pulse although mostly reactive does contain some resistive component. The signal taken after a delay from the transmit pulse is almost purely resistive.
This is true for a beach detector. A detector designed for use in heavily mineralized ground will need a ground balance control which will require more work but is quite possible to do.
OK, it must be obvious to the VLF guru's reading this that we now have a demodulated "R" or resistive component and a seperate "X" or reactive component available the same as a modern VLF detectors synchronous detectors provide .
By varying the ratio of "R" and "X" to a summing or a difference amplifier, you can vary the discrimination point the same way a VLF/TR does. It is even possible to add notch discrimination!
The detector worked fine. Discrimination was OK in both the air and beach sand. I have started work on the ground balance control to allow discrimination in highly mineralized ground.
Standard VLF motion disc circuits should also work OK. This is due to the fact that we have seperate "R" and "X" signals to work with.
Due to some strange happenings at my work, I have not been able to do anything on the detector for the last few weeks. I am in the process of setting up a second lab in my home.
I just bought a dual channel 200MHz oscilloscope, a synthesized function generator, A Lambda triple output, variable lab supply, and a bunch of other good stuff today. I will soon be back in operation. All the best, Dave. * * *