TESORO GOLDEN SABRE

[der_fisherman]

In answer to your interesting comments, ideas and questions

The schematic in the PDF TGSL_101 _PART1.pdf seems not quite correct, the output of U107b goes nowhere so it serves no purpose. Just curious.
Compare it to the schematic in this post concerning the area mentioned. While it would add more wires to the cable between the search head and the control box I can
see the improvement putting blocks 1,4 and 9 in the search head (coil assembly). Obviously block 9 because it is part of a dual IC in block 4. Block numbers being
named in the PDF mentioned.

The schematic you posted, even when I blow it up, is almost impossible to read clearly, may I suggest you post a copy in a better resolution, maybe zipped to keep the
size down......

Also in answer to a question here the capacitors C1,2 and 6 should be at the coils in the search head for maximum TX and RX signal. This would aid operation in more than one way. Increasing the Q of the LC (+R) in both not only narrows the BW of the signal it also increases the resonant rise in the RX signal improving sensitivity and depth, as well as increasing the strength of the TX signal.
As an added bonus the S/N ratio will improve for the processing circuitry, giving greater noise immunity overall. I have read much talk on shielding but no consideration to eddy currents set up which would reduce depth by interfering with the field of the TX coil.
It appears most shield the entire coil assembly yet the TX coil is not the one needing it. By not shielding the TX coil an increase in radiated field would be seen increasing detection depth since eddy currents in the shield will not be interfering with the TX field strength.

I am interested in your knowledge on why having the caps in the head is better, we have discussed it here, but up to now none of the arguments, for or against, hold water better than the other. The foremost problem I see that one in the search coil, you are stuck with them. So to my mind, the effects must improve the search depth/stability by at least 10% before I would consider doing it that way.....
Also, increasing the "resonant rise" may, or may not be a good thing with an "off resonant" detector? I would need some better information than that to make me try it.....

To me shielding only the RX coil while considering the type of metal used as well as it's thickness is a good place for improvement. So is picking a metal very slightly diamagnetic such as silver foil for the RX coil shield. However expense likely rules this out. Pb is toxic, Be both toxic and too hard, Bi and Cu are too diamagnetic, Ag too expensive, so I would have to say Zn foil is just right. Cheap and not too diamagnetic, given a thin foil which is just enough to shield noise but not interfere with alternating magnetic fields at 15 KHZ. Has anyone tried these ideas?

I also shield the Tx coil, but I can follow your thoughts that maybe it is not needed, I can certainly see a reason to not shield it in my next search head. Has anyone else done this?
By the way, Don B. uses thin foil emergency blankets for the foil shield, even thinner than Aluminium foil I am told and works better. Such a blanket costs only about a 1US$ and will supply more than enough foil for several search heads, also on my list for the next coil.

As an added bonus the S/N ratio will improve for the processing circuitry, giving greater noise immunity overall. I have read much talk on shielding but no consideration to eddy currents set up which would reduce depth by interfering with the field of the TX coil. It appears most shield the entire coil assembly yet the TX coil is not the one needing it. By not shielding the TX coil an increase in radiated field would be seen increasing detection depth since eddy currents in the shield will not be interfering with the TX field strength.

My understanding is that as the foil is very thin AND the loop is not completed, eddy currents are therefore only produced IN THE THICKNESS of the foil and not in the length, therefore only at a very tiny level, not enough to worry about. Anyone know a simple method to measure if that comment of mine is true or not?

You did not mention Aluminium foil, which I have used on several coils and gold foil, which is available so thin that the cost is reasonable.....I found this website for edible (bacteria free) gold foil, normal gold foil will also be even cheaper:-

http://www.goldleafsupplies.co.uk/ac..._Transfer.html

Where for Price:£15.70 (Excluding VAT at 20%) you get 600 Sq.cm..Which must be enough for at least the Rx coil if used carefully and overlapping just enough and having
laid a single copper wire to make sure all parts are at the same potential.The leafs when laid end to end (no overlapping are 120cm long, if cut in half to reduce the width to 2.5cm, you would get 240cm or 94" at 1" wide, which would be more than enough....

Gold being highly conductive, I would expect that any eddy currents produced would be severely reduced (isn't that why some PI metal detectors have problems in finding
gold objects?), if not completely non-existent. Also Gold leaf leaf is only around 4 to 5 millionths of an inch thick, thinner than any other foil I have seen (which of course helps to keep the price down as well!!). I will experiment with this as sometime in the future unless you feel that it will not be a good idea.

Silver leaf in books of 24 leaves, each 80mm x 80mm (Gold was 50 x 50mm), gives a length of nearly 4 meters (3.84) when cut to 4cm wide (halved) and are considerably
cheaper at only:-Price:£10.75 (Excluding VAT at 20%) per book. According to your thoughts, this should also be a good foil to use.......

Is there anyone building a coil at this moment that is willing to try a) Only shielding the Rx coil and b) Using gold or silver leaf? It would be good to have some feedback on both subjects....

With regard to costs, I see no problems with using either Gold or Silver leaf, if placed on a stable coil (super glued coils), already shaped if needed and covered with insulating tape and a single thin copper wire for example as both leaf types stick well to slightly charged surfaces without glue, but of course glue could still be applied. Then a further coating of insulating tape on top. Not forgetting to leave a gap in both the copper wire and the Gold/Silver leaf.....a bit of practice will probably help!!

I have to admit that I had never ever thought of Gold and Silver till I read your comments, where you just mentioned Silver in passing......thanks for the inspiration....

Regards

Andy

[golfnut]

Tx getting into Rx

Theres a interreaction between both TX/RX which seems to pick the crap up from the TX

Sure nulling is step 1 (coarse tune out)

A layout issue on IGSL is the the colpitts osc is co-located with th Rx pre amp!

This direct conversion architecture hates that.


Loads of Tx osc signal the Rx preamp..

S

[Dr Vel]

der_fisherman,

Sorry found the diagram on this site and do not have a better one. I did not mention Al because everyone already uses it so nothing new to add there. As to caps in the head there is no room for debate this is simply basic fundamental theory. Both coils function as parallel resonant circuits. Looking at each side (L and C) you have a coil in series with a resistance, in parallel with a series capacitor and resistor. The R of the L side is simply the resistance of the wire. The R on the cap side is termed ESR (most especially dissipation in the dielectric). These two resistances sum and in effect transfer electron momentum into thermal vibration, meaning energy in the circuit is lost as heat. On the TX side this equates to a weaker field being broadcast. On the RX side this equates into a weaker signal being received. Both effects equate to a reduced penetration depth meaning less ability to find deeper targets. When you say 'off resonance detection' you are missing the point that what you are detecting is still within the bandwidth of the LC circuit and what I meant by resonant rise is merely the fact that the less energy lost as heat in the RX circuit equates to a greater signal amplitude for the circuitry to work with. If the 'off resonance' signal you were looking for is so far away from the TX frequency that it is outside of the BW of the RX LC circuit it would make more sense to remove the C and merely use a coil with a very broadband circuit. However this is not the case since you would be dealing with a very nonlinear medium for a target. A gas plasma such as mercury vapor in a light is and example of a nonlinear medium, a solid metal gold target in the ground is not. It would be easier for me to consider this subject if you state how far off resonance is the signal you are looking for. If for example the TX oscillator is running at 15 KHZ what is the frequency deviation in HZ of the signal you are trying to detect? Then one could calculate the desired BW required in the circuit knowing the actual Q of the LC circuit. Detecting a target by measuring a change of phase angle still means working with less than 2 Pi radians of change from the TX signal and in the best metal detector ever built I am sure the BW of both the TX and RX LC circuits is much greater than this. Meaning of course improving the Q in the search coil is not going to reduce your ability to detect your 'off resonant' desired signal. What it will do is increase the transmitted field and increase the desired received signal. As I stated this is a very fundamental concept in physics.

[der_fisherman]

der_fisherman,

Sorry found the diagram on this site and do not have a better one. I did not mention Al because everyone already uses it so nothing new to add there. As to caps in the head there is no room for debate this is simply basic fundamental theory. Both coils function as parallel resonant circuits. Looking at each side (L and C) you have a coil in series with a resistance, in parallel with a series capacitor and resistor. The R of the L side is simply the resistance of the wire. The R on the cap side is termed ESR (most especially dissipation in the dielectric). These two resistances sum and in effect transfer electron momentum into thermal vibration, meaning energy in the circuit is lost as heat. On the TX side this equates to a weaker field being broadcast. On the RX side this equates into a weaker signal being received. Both effects equate to a reduced penetration depth meaning less ability to find deeper targets. When you say 'off resonance detection' you are missing the point that what you are detecting is still within the bandwidth of the LC circuit and what I meant by resonant rise is merely the fact that the less energy lost as heat in the RX circuit equates to a greater signal amplitude for the circuitry to work with. If the 'off resonance' signal you were looking for is so far away from the TX frequency that it is outside of the BW of the RX LC circuit it would make more sense to remove the C and merely use a coil with a very broadband circuit. However this is not the case since you would be dealing with a very nonlinear medium for a target. A gas plasma such as mercury vapor in a light is and example of a nonlinear medium, a solid metal gold target in the ground is not. It would be easier for me to consider this subject if you state how far off resonance is the signal you are looking for. If for example the TX oscillator is running at 15 KHZ what is the frequency deviation in HZ of the signal you are trying to detect? Then one could calculate the desired BW required in the circuit knowing the actual Q of the LC circuit. Detecting a target by measuring a change of phase angle still means working with less than 2 Pi radians of change from the TX signal and in the best metal detector ever built I am sure the BW of both the TX and RX LC circuits is much greater than this. Meaning of course improving the Q in the search coil is not going to reduce your ability to detect your 'off resonant' desired signal. What it will do is increase the transmitted field and increase the desired received signal. As I stated this is a very fundamental concept in physics.

Thanks for the answer, would you also comment on the idea of using gold or silver leaf for the coil shielding as well with regard to their possible effectiveness......thanks.

I have also read on the pages of Geotech that to double the distance of sensing, a 64 fold increase in transmitted power is also needed, which did surprise me at the time.

Conversely, if true, a small reduction in Tx power will probably not affect the distance very much as to reduce the range by half, a drop in Tx power to 1/64th would be needed.....which implies to me that even a doubling of Tx power will only result in a very small, probably unmeasurable in the field, improvement in range....

What is your take on that please?

Regards

Andy

[golfnut]

TX oscillator is running at 15 KHZ what is the frequency deviation in HZ of the signal you are trying to detect? Then one could calculate the desired BW required in the circuit knowing the actual Q of the LC circuit. Detecting a target by measuring a change of phase angle still means working with less than 2 Pi radians of change from the TX signal and in the best metal detector ever built

This is a homodyne radar in effect, or a zero IF RX, the image lands at DC.

Its my view that the Tx signal phase is what is pulled by the metal target. Rather than listening to a true return from a target (A PI does clearly)

A signal (phase pull) is generated from the motion over a target, slow sweep = low phase component return.



This what you see at the mix output - and the frequency of the signal is within a cycle.

In fact with a lively sweep and srtong target pull, you see 1 sinusoid like a heart beat on a
ecg

Steve

[simonbaker]

Sure nulling is step 1 (coarse tune out)

A layout issue on IGSL is the the colpitts osc is co-located with th Rx pre amp!

This direct conversion architecture hates that.


Loads of Tx osc signal the Rx preamp..

S

I don't think crosstalk at the PCB is significant. The null signal is due to TX-RX coil-coil interaction. If you disconnect the RX coil, you won't see any significant signal in the preamp. Whatever is there is synchronous and the SD converts it to a tiny DC voltage.

Regards,

-SB

[simonbaker]

der_fisherman,

Sorry found the diagram on this site and do not have a better one. I did not mention Al because everyone already uses it so nothing new to add there. As to caps in the head there is no room for debate this is simply basic fundamental theory. Both coils function as parallel resonant circuits. Looking at each side (L and C) you have a coil in series with a resistance, in parallel with a series capacitor and resistor. The R of the L side is simply the resistance of the wire. The R on the cap side is termed ESR (most especially dissipation in the dielectric). These two resistances sum and in effect transfer electron momentum into thermal vibration, meaning energy in the circuit is lost as heat. On the TX side this equates to a weaker field being broadcast. On the RX side this equates into a weaker signal being received. Both effects equate to a reduced penetration depth meaning less ability to find deeper targets. When you say 'off resonance detection' you are missing the point that what you are detecting is still within the bandwidth of the LC circuit and what I meant by resonant rise is merely the fact that the less energy lost as heat in the RX circuit equates to a greater signal amplitude for the circuitry to work with. If the 'off resonance' signal you were looking for is so far away from the TX frequency that it is outside of the BW of the RX LC circuit it would make more sense to remove the C and merely use a coil with a very broadband circuit. However this is not the case since you would be dealing with a very nonlinear medium for a target. A gas plasma such as mercury vapor in a light is and example of a nonlinear medium, a solid metal gold target in the ground is not. It would be easier for me to consider this subject if you state how far off resonance is the signal you are looking for. If for example the TX oscillator is running at 15 KHZ what is the frequency deviation in HZ of the signal you are trying to detect? Then one could calculate the desired BW required in the circuit knowing the actual Q of the LC circuit. Detecting a target by measuring a change of phase angle still means working with less than 2 Pi radians of change from the TX signal and in the best metal detector ever built I am sure the BW of both the TX and RX LC circuits is much greater than this. Meaning of course improving the Q in the search coil is not going to reduce your ability to detect your 'off resonant' desired signal. What it will do is increase the transmitted field and increase the desired received signal. As I stated this is a very fundamental concept in physics.

Hi Dr Vel:

It is not so simple as that. The RX coil tank seems to be intentionally low Q. The cable resistance is very small compared to the RX coil resistance anyway. If you want a more resonant tank, then use thicker wire for the RX coil -- but that has not proved useful for the TGSL design.

There may even be reason to include the cable in the tank circuit (by putting the cap on the PCB) because it may help suppress noise in the cable from reaching the PCB -- just speculation. I'm sure putting the cap in the coil head is also a viable design, probably just easier or cheaper to make compatible coil MD designs without cap.

The RX tank circuit is an "off-resonance" design, meaning the RX tank is tuned differently than the TX frequency. If you make the RX tank more resonant (higher Q), you do not necessarily increase gain -- you might even reduce it or decrease the S/N ratio, because you are now picking up much more noise at resonance and less target signal at off-resonance. The frequency difference is approx 1.6 kHz I believe.

The bandwidth of the demodulated target signal quite small, probably 10 to 50 Hz (the pulse of the target due to sweep motion). So it is almost a single frequency compared to the bandwidth of the RX tank.

Part of the reason for the off-resonant design is to control phase shift. If you get near resonance, the phase shift of the target signal is very sensitive to component values and could not be easily mass-produced. That doesn't mean we need to follow manufacturing design in our personal MD circuits -- but you would need to redesign the synchronous detector sync pulse circuits if you introduce a large phase shift into the RX circuit.

I'm not sure what 2 pi radians has to do with anything. The phase shift is modulated somewhere in the 10 to 50 Hz range I believe.

Cheers,

-SB

[simonbaker]

The schematic in the PDF TGSL_101 _PART1.pdf seems not quite correct, the output of U107b goes nowhere so it serves no purpose. Just curious.

I think that is the battery check circuit which most people omit -- I guess they don't like the beep when turn on the MD -- makes 'em jump...

-SB

[simonbaker]

This is a homodyne radar in effect, or a zero IF RX, the image lands at DC.

Its my view that the Tx signal phase is what is pulled by the metal target. Rather than listening to a true return from a target (A PI does clearly)

A signal (phase pull) is generated from the motion over a target, slow sweep = low phase component return.



This what you see at the mix output - and the frequency of the signal is within a cycle.

In fact with a lively sweep and srtong target pull, you see 1 sinusoid like a heart beat on a
ecg

Steve

Yes, I believe it can be as low as a 1 Hz with a smooth sweep of a deep target. I figure some pulses (shallow targets) may have frequency components in the 10 Hz to 50 Hz range, since that is roughly the bandwith of the amplifier stages.

-SB

[simonbaker]

This is a homodyne radar in effect, or a zero IF RX, the image lands at DC.

Its my view that the Tx signal phase is what is pulled by the metal target. Rather than listening to a true return from a target (A PI does clearly)

A signal (phase pull) is generated from the motion over a target, slow sweep = low phase component return.

Steve

I don't think of it as Tx phase pull, because the synchronous detector is synced to the TX signal and would not notice any phase modulation of the TX signal.

Rather, I think the target eddy currents are transmitting a different phase signal to the RX coil at the same frequency. We detect that tiny signal, buried in the null signal, as a differential during the target sweep, since the null signal just results in a DC voltage.

Regards,

-SB

[Dr Vel]

2 Pi radians is a measure of phase between cycles, or in effect the measure of one cycle or 360 degrees in terms of angular frequency. Or rad = degrees times Pi/180. I probably should have just mentioned phase in terms of degrees. If the difference is greater than one cycle we are talking about a frequency difference not phase unless the RX signal is a non integer difference, said difference being measured in electrical degrees between cycles. Which puts us in line with the words of simonbaker and golfnut. This is how I see it for this method of detection.

I do doubt the need for a 64 fold increase in TX power being needed to produce a greater depth of detection. This is like saying if I can put one watt of energy into my search coil and detect a coin at say 15 inches that I would need to step up to 64 watts to go any deeper. While yes it would this does not mean I will not see improvement dissipating say 5 watts in the TX coil. To go to extreme power levels you do realize the size of the battery needed to operate the detector. Of course waveshape needs to be considered. If the TX signal is a fast pulse of high peak magnitude we can achieve greater depth while using less average energy from the supply.

In other words a PI detector. Yet the lack of phase information becomes the issue meaning inability to discriminate. Probably off topic to bring it up here but it seemed to fit the conversation. I would say proper TX waveshape, coil design, and fast processing will overcome this discrimination problem and to be honest reading what Carl and moodz are doing is to me the right track. I am not as versed in metal detector theory as many here having spent 50 years designing electronics in the field of radio but I am working on it.

On the subject of various materials for the coil shielding what I was trying to consider was the magnetic properties of the shield metal and whether diamagnetic properties could make a difference.

[kliner]

I think it was said that to double the detection depth a power increase of 64 times would be required

[simonbaker]

2 Pi radians is a measure of phase between cycles, or in effect the measure of one cycle or 360 degrees in terms of angular frequency. Or rad = degrees times Pi/180. I probably should have just mentioned phase in terms of degrees. If the difference is greater than one cycle we are talking about a frequency difference not phase unless the RX signal is a non integer difference, said difference being measured in electrical degrees between cycles. Which puts us in line with the words of simonbaker and golfnut. This is how I see it for this method of detection.

I think it just boils down to phase modulation, which does not change the center frequency regardless of the amplitude of the phase change. You can modulate .1 radians or 20 radians. However, any phase modulation, rate or amplitude, will spread the spectrum of the carrier signal proportionally.

On the subject of various materials for the coil shielding what I was trying to consider was the magnetic properties of the shield metal and whether diamagnetic properties could make a difference.

I'm interested in those physics questions too. Probably a physics dude could simulate our coils and get some answers that would be useful to know.

Regards,

-SB

[der_fisherman]

Further discussions

2 Pi radians is a measure of phase between cycles, or in effect the measure of one cycle or 360 degrees in terms of angular frequency. Or rad = degrees times Pi/180. I probably should have just mentioned phase in terms of degrees. If the difference is greater than one cycle we are talking about a frequency difference not phase unless the RX signal is a non integer difference, said difference being measured in electrical degrees between cycles. Which puts us in line with the words of simonbaker and golfnut. This is how I see it for this method of detection.

I do doubt the need for a 64 fold increase in TX power being needed to produce a greater depth of detection. This is like saying if I can put one watt of energy into my search coil and detect a coin at say 15 inches that I would need to step up to 64 watts to go any deeper. While yes it would this does not mean I will not see improvement dissipating say 5 watts in the TX coil. To go to extreme power levels you do realize the size of the battery needed to operate the detector. Of course waveshape needs to be considered. If the TX signal is a fast pulse of high peak magnitude we can achieve greater depth while using less average energy from the supply.

In other words a PI detector. Yet the lack of phase information becomes the issue meaning inability to discriminate. Probably off topic to bring it up here but it seemed to fit the conversation. I would say proper TX waveshape, coil design, and fast processing will overcome this discrimination problem and to be honest reading what Carl and moodz are doing is to me the right track. I am not as versed in metal detector theory as many here having spent 50 years designing electronics in the field of radio but I am working on it.

On the subject of various materials for the coil shielding what I was trying to consider was the magnetic properties of the shield metal and whether diamagnetic properties could make a difference.

Firstly I must thank you for your further comments and ideas, you are a much needed person here to provoke good discussions in this area and I am sure that everyone here will welcome your radio knowledge as being far more positive than negative. I hope others will join me in thanking you and our "other" experts, of which there are many, thank God.

I personally am a "Digital Electronics" person and I flounder dreadfully in analog areas, I can follow a good explanation, but I am loath to say too much, though I try and find bits of info, like the Gold and Silver leaf idea that can be tested by other members and discussed by the experts.....

I was not the expert that said that a 64 fold increase in power was needed to double the Tx range. I always worked with the idea of a 4:1 increase for some ancient reason (that I have completely forgotten, Radio transmissions??), but I am not clever/knowledgeable enough to know which is correct!! If it actually follows the rules of Radio transmission, then power increases will be more effective (I think!), so can you add something to the discussion please on that point and correct my thinking?

I do like your thoughts of NOT have a sine wave, but having a fast pulse with a fairly large gap to save on battery power, similar to a PI, but in the VLF (15 KHz) range but the return signal being examined in a more VLF/TR manner, not PI, it does sound as being possible.....what does everyone think? Battery power and transmitting power being both very important.

If the factor of 64 is truly correct (I have no idea!), then going from 1 watt to 5 watt will bring only extra detection of the order of 7.8 % !! With a small object being say detected at 5" with 1 watt, it will now be detected at around 0.39" deeper. The variation on height when sweeping may even just lose this mini "gain" completely. As I previously mentioned, simply its not worth to increase Tx power so greatly for such a mini win.....Not even 0.5"!!! For larger deeper objects maybe it will help better....

Please comment further with more detail with regard to gold and silver leaf shielding....If I understand correctly what others have posted, is that not only is the type of shielding material important (what is diamagnetic etc..), but the thinner the better (more important than material type is my impression from other posts) and what is thinner than gold and silver leaf? I would guess that even the graphite sprays some use are still thicker than such leaf.......please anyone comment......Simon, Eduardo, Golfnut and Don.B etc....all of you.....we need to get some testing done before the weather improves......

I did some further research on Gold only here in Germany and the price for 10 sheets of gold leaf on a plastic carrier between 4 and 5 microns thick is €1.50, or less than US$2.00
Each leaf is 4.5 cm square. It should be "stronger" as its on a plastic sheet and easier to apply because of that.

24 sheets would be around US$4.00....and be more than enough for even two coils in a search head of 8 - 10" diameter......though as with the emergency blankets, you have to watch which side you use to contact the ground wire! Check with an Ohmeter meter first.....
Please discuss all points further....

Have a great day

Andy

[der_fisherman]

I think it was said that to double the detection depth a power increase of 64 times would be required

I read it too, who originally said that?

Andy