PCB Coil?

[bbsailor]

Here is an easy way to measure coil self resonance

bugwhiskers,

Using the equipment that you have, assuming you have a scope, you can measure a coil's self resonance.

1. Connect an another existing PI coil to either a PI TX circuit or a signal generator.

2. Position that coil (in 1 above) near the coil under test.

3. Connect the coil under test to the scope input using a 10X probe.

4. Turn on the PI TX circuit or signal generator and adjust the frequency so you can see the induced ringing in the coil under test. This induced ringing will stay constant even as you adjust the frequency of the PI TX circuit or the signal generator.

5. Enlarge your view of the ringing pattern and measure the distance between two peaks. This distance represents the period of the self resonant frequency.

The PI TX circuit or signal generator is only used to induce energy into the coil under test. If the frequency is too low, not enough energy will be induced in the coil under test to be able to observe the ringing. Work in the 1KHz to 10 KHz range and you will see a clear set of exponentally declining oscillations that represents the coil's self resonant frequency. The accuracy of you results will be based on how well you can interpet the distance between two adjacent peaks.

Your 150uH coil, with 4.8 ohms of resistance has a 31.25 uS Time Constant (TC). If your TX pulse is 19.2 uS, it does not have enough time to allow the current to build sufficiently before shutting off. In one TC the current builds to 63.2 percent of the max; in two TCs 86.5 percent of the max and in three TCs 95 percent of the max. That is why some PI circuits place a resistor in series with the coil to lower the TC so TX pulses in the 20 to 50 uS range have a chance to build sufficiently before shut off. If you added a 10 ohm resistor in series with your coil, your new TC would be 10.13 uS and a TX pulse of 33 uS would allow the pulse to build to 95 percent of its maximum value.

To make an accurate coil TC calculation add the MOSFET on-resistance to the coil resistance and divide the inductance in uH by the total coil resistance + MOSFET resistance to obtain the coil's TC in uS.

If you put a fractional ohm resistor in series with your coil to ground connection, and place your probe across this resistor, you can watch the coil current on your scope and see what I am talking about.

Here is another coil making tip. If you make a coil that has a very high self resonance, you have less capacitance which allows you to potentially sample early. However, when a little bit of capacitance is placed in parallel with the low capacitance coil, that coil's self resonance drops quickly. It is like looking at two parallel resistors in that a high value resistor in parallel with a low value resistor has a combined value closer to the lower value resistor. Low coil capacitance really makes more of a difference when other capacitance contributors are also minimized. The more capacitance you eliminate, the higher the value of damping resistor can be set. So as you change things, keep a good record of the damping resistor values. It is a good quality indicator for all the capacitance related things that are happening in the circuit.

Once you shield the coil, add the coax or what ever coil lead you use, and adjust the damping resistor value, add op amp input circuit loading, and MOSFET COSS, you have brought that new super high self resonance coil back down near the performance of the other coils. The op amp input resistor is effectively in parallel with the damping resistor while the diodes are conducting (above 0.7V). One way to see the effect of this is to increase the op amp input resistor to about 2.2K and then readjust the value of the damping resistor to a higher value for critical damping. MOSFET COSS at the operating voltage of your driver circuit also adds a good chunk of capacitance to the mix.

When you do these little experiments, as I suggested above, you can see how it is not just the coil but the things that are connected to the coil that contribute to the total performance of your PI machine.

I would like to see some inventive folks come up with a way to embed the first amp stage right in or on the coil housing. With a PC coil, like you are doing, this could be done with surface mount components, 1/8 watt resistors and tant. caps. What you need to do is to select components that will not become targets and have eddy currents generated in them while sampling is happening. With a gain of 10 to 25 this circuit might create a new breed of active coils for PI machines, something like Minelab did with their Sovereign coils that have an embedded preamp. By placing the first amp in the coil, you lower the impedance and lessen the effects of RX cable capacitance on coil performance. Cable capacitance is another large chunk of capacitance that potentially limits performace.

These are just some ideas to keep you awake at night.

I hope you can try the coil self resonance measuring method that I suggested above. Let us know if it works?

bbsailor

[Tinkerer]

Bbsailor,

Like everyone of your posts, this one is full of wisdom.
Allow me to tap your vast experience with a few questions.
Question #1: Coil current.
There seem to be 2 tendencies, the one uses the uses DC resistance of the coil to limit the current. High resistance makes for a fast TC, thus it is possible to make a pulse ramp that has a flat top, 3 TC or more. The rational is to minimize the eddy currents generated by the TX.
The other tendency uses a low DC resistance and limits the current with the impedance. Some times the TX ramp is less than one TC.
What are the pro and con’s of the two tendencies?

Question #2: Coil response:
I have noticed that different types of coil windings produce a different shaped RX response field. Some windings produce a very sharp center coil response. This is useful for pinpointing the target. However, the response of the coil is relatively less strong near the outer rim.
Other windings seem to have a response that is more diffused at the center and have a more evenly spread response across the whole width of the coil.
Do you have more info about this?

Question #3: Input resistor:
Using an inverting opamp configuration, the resistor before the diodes defines the gain of the opamp. In the non inverting configuration not. A high value resistor helps pushing the flyback to a high voltage that reduces the first delay.
What is the disadvantage of a high value resistor, lets say 4k7?

Tinkerer

[bbsailor]

Tinkerer,

Like everyone of your posts, this one is full of wisdom.
Not wisdom, but setting up a lab and winding a lot of coils and reading all of Eric Foster's posts and seeing the results with "hands on" or should I say "eyes on" the scope signals. I also went back to basic inductor theory to make sure I really understood and could internalize what is happening when a pulse is first applied to a coil and then what happens when the pulse is removed. It is important to fully understand where and how the processing is being done to extract a signal from a target in the presence of noise and how different targets respond best with different types of stimulating pulses and timing signals.

Allow me to tap your vast experience with a few questions.
Question #1: Coil current.
There seem to be 2 tendencies, the one uses the uses DC resistance of the coil to limit the current. High resistance makes for a fast TC, thus it is possible to make a pulse ramp that has a flat top, 3 TC or more. The rational is to minimize (?) the eddy currents generated by the TX.

I believe the rationale is to maximize the eddy currents in the target by allowing the coil current to fully or near fully build. As you change the TX pulse width you are changing the peak flyback voltage assuming you have not exceeded the MOSFET voltage rating. Different targets need different types of stimulation. Small gold responds better to earlier sampling with less current and higher PPS while larger targets such as coins and relics respond better with longer pulses and more coil current.

The other tendency uses a low DC resistance and limits the current with the impedance. Some times the TX ramp is less than one TC.
What are the pro and con’s of the two tendencies?

As Eric Foster stated on his PI forum, "start designing your coil based on the types of targets you mainly seek". While a coil optimized for small gold will also find coins, but it may not be optimized for coins. This is why many PIs have external controls to vary the Pulse Width, Pulse Frequency, Pulse Delay, Second Delay and RX Sample Window Width to allow for some range of target optimization. However, a high frequency low power gold machine will not make a deep seeking relic finder that typically needs a larger coil with more coil current operating at a lower frequency.

Question #2: Coil response:
I have noticed that different types of coil windings produce a different shaped RX response field. Some windings produce a very sharp center coil response. This is useful for pinpointing the target. However, the response of the coil is relatively less strong near the outer rim.
Other windings seem to have a response that is more diffused at the center and have a more evenly spread response across the whole width of the coil.
Do you have more info about this?

Some coils are very small and use a ferrite core to concentrate the pulse field. Other coils are a variety of mono coil shapes that are designed to penetrate deep for larger targets, or cover a wider area at a medium depth such as for beach hunting, while other coils (DD design) are designed to help neutralize mineralized ground or electrical interference. Coil design depends on what you are seeking and what the conditions are where you are hunting.

Question #3: Input resistor:
Using an inverting opamp configuration, the resistor before the diodes defines the gain of the opamp. In the non inverting configuration not. A high value resistor helps pushing the flyback to a high voltage that reduces the first delay.
What is the disadvantage of a high value resistor, lets say 4k7?

Reg and Eric should jump in on this answer. I'll take a swing at it. Higher input resistors contribute more noise in a high gain circuit than resistors with lower values. Also, high gain op amps have a bandwidth limitation. This may be why Eric uses two stages to obtain the desired gain.

When a DD coil is used, people have reported obtaining a 2 to 4 uS delay improvement over a mono coil. This is due to the fact that the RX coil is nulled or close to nulled and does not saturate the input stage of the first amplifier stage so sampling can be done a few uS sooner. Also, the RX coil (in a DD design) does not see the capacitance of the MOSFET so the RX coil will typically critically damp at a higher damping resistor value than the TX damping resistor.

Here is a practical problem for Tinkerer to figure out. The formula for calculating the value of the critical damping resistor is Rd = 2 X the Square Root of L Divided by C (with both L and C under the Sq Root sign). (L in micro henries and C in microfarads, example a 500 uH coil with a C of 400 pf = 2236 ohms). Draw the formula out for clarity. Calculate the damping resistor value for one of your coils by measuring the coil's self resonance at the end of the coax and use the coil's known inductance.

Compare the calculated value of Rd to the actual value of Rd adjusted on your PI machine. Why are the values different? What does the calculated value of Rd not take into account? This exercise will help you get your head around the invisible things that are happening in the circuit that relate to your questions as well as my answers.

Have Fun!

bbsailor

[bugwhiskers]

Answers for bbsailor

Hi bbsailor,

I performed both tests and the TC of the coil in question is 1.2 uS.

Its not the fastest PCB coil I have but it just works better than any of the others. Another one tested was .58uS, a spiral circle with 20* 2mm tracks.

I am preparing a spreadsheet for Andy with info on a range of capsule shaped PCB coils for Andy to come up with a formula to predict inductance for various configurations.

I also tried the series resistor trick and observed the rise, its not unlike the voltage rise on a charging capacitor... ever diminishing return for input voltage.

regards

bugwhiskers

[Tinkerer]

There seem to be 2 tendencies, the one uses the uses DC resistance of the coil to limit the current. High resistance makes for a fast TC, thus it is possible to make a pulse ramp that has a flat top, 3 TC or more. The rational is to minimize (?) the eddy currents generated by the TX.

I believe the rationale is to maximize the eddy currents in the target by allowing the coil current to fully or near fully build. As you change the TX pulse width you are changing the peak flyback voltage assuming you have not exceeded the MOSFET voltage rating. Different targets need different types of stimulation. Small gold responds better to earlier sampling with less current and higher PPS while larger targets such as coins and relics respond better with longer pulses and more coil current.

bbsailor,
thanks for the info.
of course you are right. What I meant is to minimise the eddy currents generated byt the Tx ramp not the cut off.

Tinkerer

[bbsailor]

bugwhiskers,

The coil TC is the inductance divided by the total coil resistance including MOSFET on-resistance and any series resistor.

The 1.2 uS I believe is the time between the adjacent peaks depicting the coil's self resonance of 833KHZ. The .58uS represents a 1.724MHz self resonance. Is this what you mean? Do these uS measurements include the coil coax or coil lead?

bbsailor

[bugwhiskers]

Reply to bbsailor

Hi bbsailor,

The figures include cabling, 300 ohm TV ribbon cable.
My CRO (Tektronix 2246) has on screen measuring and it was between the peaks.

regards

bugwhiskers

[Andy]

PI coil theory?

BBSailor: I wrote this up earlier before reading your posts here, I'll go search for Eric Foster's posts, thanks for the great information. Do you know if Minelab put their circuit in the center of the coil? Or on the outer edge?

Bugwhiskers,

After a lot of thinking on my previous posts, I decided to review some of the basic principles on metal detectors and in particular the pulse induction type. Coming from the other end of the spectrum, it appeared to be as easy as just changing frequencies in the formulas. What I found was that this is not correct, PI's don't work on “normal” rf principles, they work on magnetic principles. Yes, these two are related but they are not the same, and E-Field [electrical] is not the same as an H-field [magnetic]. The PI coil works on magnetic principles, but the feed leading up to the coil and all the remainder of the circuitry around rf-based.

Why do a review? Understanding how the PI really works is key to building better coil, and a better PI. It can also set the expectations of just what a coil will and will not do. In real terms, it means not wasting time and material building stuff that does have a snowballs chance in the Outback [pretty hot there right? :-)] of working. It also means we can accurately model PCB coils.

Basic Stuff

The primary purpose of a metal detector should be obvious, it is to detect metallic objects. Depending on the objective of the metal detector, the search coil needs to project either a small or large field. The field projection dictates that the search coil must be of a type that does not contain the field, this excludes designs such as a toroid, and can only include designs that insure that the field is radiated [external field]. Air wound or non-metallic-core coils provide the best external field or field projection [this includes wire-wound and printed-circuit designs].

Pulse Induction

A Pulse Induction [PI] detector is using an air cored coil not because it requires a tuned electronic circuit element with a specific inductance, instead, it is using the air coil as a “proximity” sensor for metallic objects. Again, the coil is not for the inductance used in an electrical circuit, but to create a region of space [magnetic field] having a definite magnetic flux density[1]. The magnetic flux density is primarily dictated by the amount of current flowing through the coil [Biot-Savart equation, not entirely accurate for here, but close] , it is also dependent on the cross-sectional layout of the coil [Sommerfeld, Maxwell, and Ampere] , a PCB coil excels here. The strength of the magnetic field is called “permeability” or in terms of detecting, how conductive an object is to magnetic flux.

Current flowing through the coil generates a magnetic field, as the magnetic field expands away from the coil, materials in the path react the magnetic field. It is probably safe to say that all materials will react to the magnetic field, however not all objects [material] have the same magnetic conductivity or permeability. Materials such as wood, plastic, air have low permeability [many metals also have low permeability, such as aluminum foil], while other materials, mostly those related to iron or ferrous have high permeability. A high permeability means a stronger reaction to the magnetic field produced by the coil.

The PI works by “firing” a pulse of current through the coil, in the absence of a certain threshold of permeability material, the pulsed magnetic field will peak and then start to decay, eventually going to zero. The length of time it takes for the field to peak and then decay to zero, again - in absence of a certain threshold of permeability material, is calculated and saved during calibration of the PI detector. In actual use, if a “high” permeability material is in the path of the coils magnetic field, that material will alter the decay time, the change in decay time is detected and reported by the PI.

Technically the reaction to the magnetic field is such that the magnetic field from the coil induces a secondary magnetic field(s) [also called eddy currents] in a material [aka target], the secondary field of the target then influences the field produced by the coil. The strength, reach, and [hopefully inductive influence :-)] of a coils' magnetic field depends on the size of the coil, the coils' cross-sectional area, the current through the coil and the current carrying ability of the coils' conductor.

Detecting the change in the magnetic field decay as opposed to creating the field, is another story, as mentioned earlier, magnetic circuits are not the same as electrical circuits, however they are closely related, a collapsing magnetic field also means a collapsing electrical “component”, in this case, voltage. Another way to look at it, is to see the coil voltage representing the rate of change, or differential of the magnetic flux in the coil. As the magnetic flux changes, so does the voltage. It is the voltage component that is monitored in PI circuitry. This is also critical to understanding a PI coil design.

PCB Coil

What it is the purpose of a PCB coil or even an accurately wound copper-wire coil? Based on the above it's sole purpose would be to create an accurate and repeatable magnetic field. If the magnetic field generated is accurate, that is in terms of magnetic properties, then the detection of secondary fields can be more accurate, the former is through the coil, the latter is through electronic circuits.

Magnetic Properties

In terms of magnetics, most materials have well known magnetic properties that is characterized by a magnetization or a B-H curve. These curves are separated into three [really two] categories, a reference material curve [vacuum, can be thought of as “zero”], diamagnetic material curve [ always lower than the reference, can be thought of as “negative”], and finally a paramagnetic material curve [always higher than the reference, can be thought of as “positive']. ... well actually this can be extended to include a breakdown on ferrous and non-ferrous metals too.

A B-H curve or plot or graph is very simple. It is an x-y graph where the horizontal or “x” axis represents the magnetic field strength, and the vertical or “y” axis represents the magnetic flux density.

A PI more or less tracks or plots a B-H curve, this is done via an integrator circuit. However this where “eye-ball viewing” a B-H plot and a detector part ways. The detector circuitry is looking for any change to the magnetic field, and is not concerned what side of the reference point it falls on, or at.

Expectations

In any case the purpose of this, is to set the expectations of what will be accomplished by a PCB coil design. What a coil by itself will not do is discern between sizes of material, say gold nuggets. What it will do is allow for a highly repeatable coil design that will give the electronics a very stable and accurate base to work with. If the coil properties are rigidly fixed [can not get much more rigid than a PCB], then the amount of current and voltage can be more precisely tracked. This precise tracking in turn allows for a more accurate representation of the magnetic flux changes, which will show up as better depth and ability to “target” smaller objects [like very small nuggets or flakes].

I agree with BBSailor in moving some of the components to the coil. Ideally the information from [leaving] the coil should be in a digital format, this would remove most of the weak spots from the circuitry [temperature differences, current and voltage drift]. First things first, get a repeatable design, this includes signal feeds to the coil.

For the modeling part, I have started on some design tools, to make them usable I am writing everything in Java, which should run just about everywhere. My expectations are to display the coil on the screen and to accurately save it in a easy to use format that can be imported by CAD programs. The software package will include the full source code with the code documented.

Andy

References:
[1] Clarke, R, 'Producing Wound components', http://www.ee.surrey.ac.uk/Workshop/...ils/index.html

[bbsailor]

Coil Preamp

Andy,

The Minelab Sovereign RX coil preamp is inside one of the D shaped coils. It derives it's power from a diode bridge off the TX circuit with a gain of about 25.

I was thinking that maybe they found a null spot, but it appears that they just located it inside one of the coils near one of the "D" corners. There is a photo of an open Sovereign coil housing and a hand-drawn preamp schematic somewhere on this forum. Use the search feature.

bbsailor

[bugwhiskers]

PCB coil info for Andy

Hi Andy,

The following are capsule shaped PCB coils I have made and characterized. The circular spiral coils maths is the same as wirewound.

All the capsule shaped coils are 285mm Long 150mm Wide.

Single sided 1oz copper.

33 Turns, 260uH, 5.7 Ohms TC 1.15uS, 1mm Tracks, 0.5mm between tracks.

36 Turns, 200uH, 7.8 Ohms, TC .97uS, 1.3mm Tracks, 0.5mm between tracks.

10 Turns, 35uH, 3.1 Ohms, TC .49uS, 1.3mm Tracks, 0.5mm between tracks.

Double sided 2oz copper.

26 Turns(total) 150uH, 4.8 Ohms, TC 1.2uS, 1mm Tracks, 3mm between tracks (each side)

44 Turns(total), 310,uH, 6.7 Ohms, TC 1.75uS, 1mm Tracks, 2 mm between tracks(eash side)

The last coil detailed is the latest (made today) and is the best performer.

Looking forward to seeing your efforts.

regards

Bugwhiskers
PS If you need to know about blank space size in the middle of the coil let me know.

[bugwhiskers]

Yearnings

A PI more or less tracks or plots a B-H curve, this is done via an integrator circuit. However this where “eye-ball viewing” a B-H plot and a detector part ways. The detector circuitry is looking for any change to the magnetic field, and is not concerned what side of the reference point it falls on, or at.

Andy,

I dont suppose you have access to a 1MHz Spectrum Analyzer ?

I have a hope to one day make up a coil and very fast pre-amp and pulse the coil at about 500 Hz.

Hook up a spectrum analyzer and see what parts of the decay waveform change with different metals/distance/size.

Make bandpass and notch filters(gyrators) to boost wanted segments and reject unwanted parts of the decay waveform. The output would then be fed straight to Headphones and the training of the ear/brain can commence.

regards

Bugwhiskers

[bbsailor]

PCB Coil Goals

bugwhiskers,

Here is a tip to help you compare coil performance.

You said:
"Double sided 2oz copper.

26 Turns(total) 150uH, 4.8 Ohms, TC 1.2uS, 1mm Tracks, 3mm between tracks (each side)

44 Turns(total), 310,uH, 6.7 Ohms, TC 1.75uS, 1mm Tracks, 2 mm between tracks(eash side)

The last coil detailed is the latest (made today) and is the best performer".

The coil that is the best performer my not be the absolute best for all targets, but could be the best based on the decay time of a particular target relative to the coil characteristics.

The "rise time" Time Constant (TC) of a coil is calculated by dividing total resistance of the coil circuit (Coil + MOSSFET on-resistance and any series resistor). In your case, the 44 turn 310 UH, 6.7 ohm coil and assume the MOSFET on-resistance is .5 ohms for a total of 7.2 ohms... 310/7.2=43uS. In one TC (TX pulse width 43 uS) your current will grow to 63 percent of max and in two TCs (86 uS pulse width) 86 percent of max and in three TCs (129 TX pulse width) 95 percent of max. Keeping the TX pulse on longer gives you only a little additional current growth to reach maximum. Adding a 10 ohm series resistor will lower the TX pulse widths to 18 uS, 36 uS, and 54 uS respectivly. So, a TX pulse width of 54 uS will get your coil current to 95 percent of it's max. Now you are trading off pulse width and target response for battery life.

Your 310 uH coil has a self resonant frequency of 0.5714 MHz or 571.4 KHz which is pretty good for a coil with 44 turns. Your coil has a capacitance of 250 pf. It would be impossible to obtain this high of a self resonance using traditional bundle winding techniques. My 268 mm diameter coil, using 19 turns of AWG 30 Teflon insulated wire has an inductance of 317 uH, resonating at 1.25 MHz with 51 pf of capacitance, close to your coil's inductance value but has less capacitance and would be less sensitive due to fewer turns. If I were to make my coil have 1585 uH, it would have 44 turns and it's self resonance would be about 350 KHz and would have more capacitance than your coil.

Now, the "fall time" Time Constant is the characteristic that really matters and is related to why you are seeking to make a low capacitance coil. Let's see what is happening. The Time Constant of the fall time current is what makes a coil fast because you want the coil to turn off 5 times faster than the decay time of the particular target you are seeking. The fall time TC is based on the value of the damping resistor, so lower capacitance coils have larger damping resistors and results in a faster current turn-off time. Let's assume that the damping resistor (Rd) for your coil is 1K ohm, then your 310 UH coil's turn-off TC is 0.31 uS thus making a target with a decay time of 1.55 uS the smallest target that you could easily detect. Now add the value of the op amp input resistor (assuming a 1 K value) that is effectivly in parallel with Rd during the conduction period of the clamping diode and you might have an effective Rd value of 500 ohms which moves the minimum target decay time now to 3.1 uS. Put your own circuit parameters into my example and see potential speed of your coil.

What you want to measure, to see the differences in various coil performance, is the current fall time of each of your coils which will tell you the minumum target decay time that you could detect. As you can see, the time constant ranges below 5 uS tells you alot about the potential sensitivity of a particular coil to a particular type of target. This 5 uS does not take ino account the op amp settling time which can add a few more uS to the total delay time. The trade off that yields the best result is having the largest number of coil turns that allows the PI circuit to operate in the desired target decay range. That is why getting coils to operate below 10 uS is balancing act between coil capacitance, coil inductance, MOSFET voltage rating, MOSFET COSS, damping resistor value, peak coil current, coil coax or coil lead length, and op amp settling time. To make the best coil comparison, you should try to only change one variable (or one set of related variables) while keeping the others the same. In some cases that will be difficult but looking at the speed of the current decay curve will tell you much about the coil's potential performance. Many circuit variables will be reflected in the value of Rd so keep a good set of records about this value with each change you make. This technique will be most helpful when trying to fully understand what is going on with the more subtle characteristics that are more difficult to isolate independently.

Let us know the following values of your coils and PI circuit? This will provide valuable information for those doing experiments and will result in some good technical exchanges on this forum.

Coil Inductance
Coil Resistance
Coil Resonant frequency (1 divided by the period between resonant peaks)
Damping resistor value
Op Amp input resistor value
Op amp gain
Current fall time (as see by measuring througn a 0.1 ohm resistor in series with the coil)
MOSFET used
Peak coil current
Pulse frequency (PPS)
TX Pulse width
Value of any series resistor

These tips should help your focus your coil research. You will soon see where the physical and electrical limits of coil design are as they relate to particular targets that you want to detect.

bbsailor

[bugwhiskers]

Data for bbsailor

Coil Inductance 310uH
Coil Resistance 6.8 Ohms
Coil Resonant frequency 571.4 kHz
Damping resistor value 470 Ohm
Op Amp input resistor value 470 Ohms to back to back 1N914 diodes then
220 ohms to inverting input of LM318
Op amp gain 4545 (220//1M)
Current fall time ~ 2uS
MOSFET used IRF9640 (P channel 200 V)
Peak coil current 2.48 Amps
Pulse frequency (PPS) 500Hz
TX Pulse width 73uS
Value of any series resistor nil
bbsailor

regards
bugwhiskers

[bugwhiskers]

Attached pic is my PIMD, the 9V Toshiba battery is for size comparisons only.

The LCD screens really aid programming the Atmel micro-controller.

The unadjusted screens from left to right are, sample 1, sample2, sample3, Differential op amp from sample 1 and sample 2, difference between sample 1 and sample 2, and finally, raw battery voltage reading.

Things to note, pot core for SMPS generating 19.5v for the coil and +5/-5 for op amps and micro.

The PCB coil shown is referred to as "most recent" in previous posts.


regards

bugwhiskers

[Andy]

Coil Math

Bugwhiskers,

Excellent information on the coils. I'll work through the math to see if I have something that makes sense. On the radial track path, you picked the best layout possible [to me anyway]. I was wondering about how you handled the "corners" and was hoping you had a radius as opposed to edges or angles. Is the track layer underneath in-between the upper layer tracks? You mentioned this on the previous coils as being the case.

I am not as fast as you on implementation, I wanted go through the theory, get a rough idea of the design and expectations, collect the test gear and then methodically proceed to test. I'll still stay on this route, and post everything here as you are doing. I do think that PCB is the way to go, once we map out all the variables it's very easy to replicate/duplicate.

BBSailor pretty much summed it up overall about the decay-side, it's where the information is at. That decay-field [off the coil] is interesting unto itself, it makes me wonder if taking a sample off the coil directly is the best way to get the information. There are two interesting things in that coil, the voltage change [decay] and also the current change. Right now we are only looking at voltage, maybe something interesting is happening on the current-side too, and it may not be apparent, such as lead/lag or phase angle in relationship to each other when something influences the field. I intend to monitor voltage and current, there might be something useful that we can exploit.

I do have a spectrum analyzer, however the BH curve is easier seen on a scope. If your detector is not "private", perhaps you could post just the analog part that is handling the coil. I ask, because I can probably tell you what two points to monitor with a scope to get the BH curve. If it's a touchy subject, then I'll figure out another way to get that information to/for you. I am assuming that you are using the PIC just for generating the pulse and everything else is handled by analog circuitry.

Waiting for some new board material so I can print out some boards to test wtih, that way this won't be one-sided.

BBSailor, do you have information about the different leads being used and what the problems are? The coil is only part of the sytem, getting the current pulse to the coil is another part. Depending on how one looks at the lead/feed to the coil, it [the feed] could just be something delivering current and accepting voltage, or it could be an active element, as in being part of a tuned circuit. My take is that if any AC component is traveling down it, then it takes on the characteristics of a tuned circuit, at least from the basics, impedance and capacitance. I would call a 10usec current pulse an AC component, and that rapidly decaying field on the return one too. You ideas on the active coil might be significant jump forward. At least get all the noisy stuff down there, including the PWM generator, doing this would pretty much eliminate issues with extra [unwanted] noise and capacitance.

That twin-lead that Bugwhiskers is using is interesting, it's looks like it's composed of smaller stranded wire to make a single conductor, makes a nice path for both components [AC + DC]. Not sure if it's optimized for deliverying that current pulse to the coil, figure that out later. Bugwhiskers, don't change it! I suspect your onto something with it [intentionally or otherwise :-)]. It would be nice to know the length of it though, from board-to-board.

Andy