Coil as sensor

Cenlasoft · 2009-02-01 16:45

Hello,
I have been experimenting with coils and resonance for a while. My concept is to use a coil as a sensor to test metals. That part I can do. I want to measure the voltage change in a coil with the AD0831 and a BSII or SX28. The ADC and BSII is pretty easy. I researched resonance circuits and have constructed several coils. I have a scope and have tried a parallel tank circuit. It worked, I think. I sent the appropriate frequency (square wave) from a 555 astable circuit into a 44uH and .22uF circuit. I used a compass to see if the coil was producing a magnetic field and it was. I adjusted the frequency until maximum compass deflection, but I don't think I am at resonance. I have ordered a function generator, but its not here yet. I cannot get a series resonance to deflect the compass, it does nothing.

square wave>

| |(.22uF)

Coil (44uH)

>Gnd from square wave generator.

I tested the coil with a scope and I get a ringing. Any help or suggestions would be appreciated.

Thanks,
Curtis

$WMc% · 2009-02-01 21:33

Cenlasoft

You really, already have a function generator.The BS2 and the SX28 with a PULSOUT command will do the trick.I'm not and engineer and I don't want to miss lead You, But .22uF seems a little high? for Your Tank circuit.

GE uses a ~18" long coil with a 1/2" steel rod passing through the coil for throttle position on their Steam Turbines. This "rod" is the Position feed back from the V1 valves on the steam turbine. This is a very simple circuit and is vary reliable. It can also be easily adapted to Your application. I don't have it here at home But Monday (at work), I'll scan it and post it here in this thread

_____________$WMc%____________

sorry I cant send!!! PAT.#

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The Truth is out there············································ BoogerWoods, FL. USA

Post Edited ($WMc%) : 3/1/2009 12:37:25 AM GMT

Carl Hayes · 2009-02-01 21:42

What property of metals are you trying to test?

The resonant properties of a tank circuit affect only its alternating-current behavior. The compass you mentioned, on the other hand, was affected by the DC component of the current in the coil. Since the square wave you presented was switching between ground and (one supposes) +5 volts, it had both an AC and a DC component. The compass told you about the DC component, but was unaffected by the AC component (because it was changing too fast for the compass to move in response).

Probably when you varied the frequency you inadvertently varied the duty cycle too, which would of course change the average value of the DC. That's why the compass responded differently.

Using a DC detector (the compass) is surely the poorest possible way to try to measure AC resonance effects.

If I were you I would rethink this process a bit.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
· -- Carl, nn5i@arrl.net

Cenlasoft · 2009-02-01 23:37

Thanks to everyone for the suggestions, I will read more about the DC part that's moving the compass. I am an archaeologist and I am trying to do non-destructive testing on metal artifacts (eddy current analysis). I tried using the BSII, but the output frequency was too low. The SX28 could get the right frequencies, but the margin of error was +/- 100 hertz. What can I use to see the AC part?

Thanks,
Curtis

Beau Schwabe · 2009-02-01 23:44

Cenlasoft,

With the values that you have chosen ( 44uH and .22uF ) , your resonant frequency should be about 51kHz.

As it has been indicated, a compass should have the least amount of deflection while at resonance because of the AC nature.

There are many ways to read the inductance changes of a coil, your application will depend on which solution will be the best. If you can provide more details, such as schematics, pictures, etc. of what you currently have and give us a better idea of what you want to accomplish, then we might be able to help.

If you have a Scope, you should be able to measure and look at the square wave produced from the 555 with some degree of accuracy near 51kHz.

You should know that not all coils are created equal. You can have an air-core coil that has the same inductance value as a metal-slug coil and they will have very different characteristics when it comes to how they will respond to the presence of any additional external metal. The spacing of the windings as well as the wire thickness can also play a part that will affect how the coil responds.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Cenlasoft · 2009-02-01 23:54

Hello,
Thanks again, My 555 circuit is adjustable and I can get very close to the resonant frequency. I tested the coil with an lcr meter and got very close to the calculated value of the coil. Beau, you had sent me a circuit with a series lc as in the posting. It also had a diode to convert to dc and a voltage divider going to a capacitor and then an ADC to read the result and sending it to the SX28 or BSII.

I want to put the coil near a metal artifact (on it) and read the voltage change as I scan it.

I'll work on posting some pictures and schematics.

Thanks,
Curtis

Carl Hayes · 2009-02-03 02:58

Reading between the lines, it looks as if what you're actually trying to build is a metal detector. Using a coil as a metal detector is actually a rather mature technology. I've never built one, or even used one, but there have been many articles. Basically, what you do is have one oscillator whose frequency is controlled by the inductance of your sensing coil, and one whose frequency is controlled only by the position of a knob. You set the one controlled by the knob very close to the one controlled by the sensing coil, and listen to the beat frequency between the two. When the sensing coil approaches a ferromagnetic object, its inductance increases ever so slightly, so that oscillator goes down in frequency and the beat note in your ear changes. When the sensing coil approaches a nonmagnetic metal object, its inductance is decreased ever so slightly, so that oscillator increases its frequency slightly, and the beat note in your ear changes.

That's how I understand it anyway, but I haven't actually studied these devices.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
· -- Carl, nn5i@arrl.net

Cenlasoft · 2009-02-03 04:00

Hi,
Thanks Carl,
In a sense the metal detector is close to what I want to build. I want to use it to test for cracks or irregularities in copper, aluminum, bronze, and other metals as well. I want to scan an object with a coil (probe) and measure eddy currents. I want to use an ADC to numerically represent the smooth areas and irregular ones. Sort of a metal scanner. I will post some details of my progress when my function generator comes in. I will also purchase the parallax usb scope because I need two channels.
Thanks,
Curtis

Cenlasoft · 2009-02-08 21:18

Hello,
I want to thank everyone for your past help. I have a few
questions.
1. Using a simple coil and capacitor in series, I
inserted a square wave using a function generator. The coil had a value of 36 uH and the
capacitor was .22 uF. I also tried .1 uF, and .01 uF capacitors I could not get the coil to resonate.
2. The parallel tank circuit worked and resonated and I could see it with my scope.

I need the series type because I want the low impedance in order to have enough current and voltage to send to an ADC and then sx28 for analysis. Here is a typical circuit I used.

Square wave >

|coil|

|cap|

>ground

3. My test points for the scope was before and after the coil. Am I correct?

I researched resonance and AC theory and I am stumped as to why I cannot get a series LC circuit to work. I am no expert, but maybe I missed something?
Any help would be appreciated.
Curtis

$WMc% · 2009-02-08 23:19

Cenlasoft

There are some other methods to detect flaws in metals that are nondistructive

1 X-RAY or RT, But its not easy to get a radioactive sorce and the 80 LB. depleated Uranuim case to store the sorce in is a pain to work with.

2 UT Ultra Sonic Testing. UT has been around for years. Used in Boilers to measure Tube thickness,detect cracks, and flaws in welds. It quickly gained use in the raceing field. As cylinder walls could be measured for thickness and rollcage welds checked for integrity. This is the easyest N.D.T. going. I have a UT meter at work that is 4"wx6"hx3/4"d, It works on all metals and it even works on plastic ,PVC, ETC.

______________$WMc%_______

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The Truth is out there············································ BoogerWoods, FL. USA

Tracy Allen · 2009-02-09 00:28

Cenlasoft, I concur with $WMc% about the ultrasonic testing, maybe suitable for your application. There is a very nice tutorial on the subject at,
www.ndt-ed.org/EducationResources/CommunityCollege/Ultrasonics/cc_ut_index.htm

As for the resonance, I think you must be doing something wrong or have a misconception about the measurement. It is the whole circuit (=coil + capacitor + signal source) that resonates, not just the coil. Be sure your ground leads are not effectively shorting out the signal.

I just hooked up a 1mH coil in series with a 0.01uF capacitor, just like your diagram, with a function generator feeding the series combination. The function generator has an output resistance of 50 ohms. Channel A of the 'scope looks at the signal generator output, while channel B measures the voltage across the coil, and both channels have 1MOhms x10 probes. There is a strong resonance at close to 50 kHz, as there should be. The voltage across the coil shoots up to a high level, much higher than the input signal level. At that same time, the output from the generator measured across the resonant circuit drops to a much lower level, which is due to the low resistance of the series resonant circuit reacting with the 50 ohm output resistance of the generator. The resonant circuit is absorbing and dissipating energy. Is that clear? The voltage across the capacitor is also high. At frequencies much lower than resonance, the coil acts more like a short circuit, and the entire voltage=signal source simply appears across the capacitor. At frequencies much higher than resonance, the opposite is true, the capacitor acts like a short circuit, and the entire voltage=signal source appears across the coil. The waveforms are more complicated when you drive it with a square wave. When you drive it with a sine wave, the input and output are both sine waves that change in phase relation to one another. But when you drive it with a square wave, both the input (after the source resistance) and output waveforms are ugly. One way to look at that is that the resonant circuit extracts only one frequency from the square wave and leaves the rest with various phase shifts.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

steve_b · 2009-02-09 14:43

I've seen some testing where they take a current and pass it through the DUT (Device Under Test)....they would then sprinkle iron filings and see where they congregated...if it was along a line, it meant a crack (as the current would set up a N-S pole type of attraction).

Not sure how this would work with brass (non-magnetics) and it's certainly harder to do this with complex forms (ie. engine blocks).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
<FONT>Steve

What's the best thing to do in a lightning storm? "take a one iron out the bag and hold it straight up above your head, even God cant hit a one iron!"
Lee Travino after the second time being hit by lightning!

parsko · 2009-02-09 16:35

Cenlasoft,

I'm literally doing the same exact thing at home right now, for a different project. On Friday, i had the same issue. I had a coil-cap combo I was driving with a square wave off a Propeller pin. I could not get it to resonate. I used a keyboard input (using the up/down/pgup/pgdn keys) to adjust the frequency of the square wave. I used steps of 10Hz and 500Hz (fine and course). No resonance. My next step is to try to drive it using a 1Hz step (which will take longer to find the resonant freq, if there is one).

Beau has done this, and you can see a list of good references at my LVDT link here:
LVDT (Linear Variable Differential Transformer)

From what I can tell, a resonant circuit shouldn't be that sensitive. Meaning, you should be able to get it to resonate if you are "close" to the resonant frequency.

Tracy, can you repeat your experiment with a square wave? Can you keep note as to when you start to see a resonance, when the peak hits, then when you start to see it drop off? Aka, if resonance is 50k, does it start at 49,999, peak at 50,000, and fall off at 50,001? Or is it more like 49500, 50000, 50500? I can understand that the exact design will affect this, but I've been curious to know how sensitive these resonant frequencies are.

-Parsko

Beau Schwabe · 2009-02-09 17:20

parsko,

What do your coils look like? If the 'Q' is low or the coil is already saturated with metal to achieve the desired inductance, then you will see low output results. The 'best' coil in terms of high 'Q' is going to be an air core... even better an air core which has the same depth (or length) as the diameter. This minimizes parasitic capacitive effects within the coil.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

parsko · 2009-02-09 17:41

Beau,

I'll post my efforts to my topic post. I'd like to not hijack this one. But, I thought my question above was relevant to both of us. I'll try to post later with a more detailed question and information about my problem. I have to get the exhaust manifold back in my car first though...

I'm still curious, though, how sensitive these are. It wasn't mentioned in your objects that you posted (that I link to in my post).

-Luke

Tracy Allen · 2009-02-09 21:00

Parsko,

I used a ferrite core shielded inductor, 1 milliHenry in series with a 0.01 mF X7R capactor and test leads from the signal generator. The measured resonant frequency was 50.25 kHz, with half power points (0.707 in voltage) of 49.5 -- 51 kHz. That 1.5 kHz bandwidth corresponds to a Q factor of 33. The series resistance (not counting the generator) was about 12 ohms, as determined from the depth of the resonant notch, and from the measured Q.

A 10 Hz frequency scan step should be able to detect that. At a higher Q, say Q=500, a good air core coil would have a bandwidth of 100 Hz and the 10Hz step followed by 1Hz should be able to find it. A very low Q coil on the other hand might have so much series resistance that the dip for resonance would be so shallow and wide as to escape detection.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Cenlasoft · 2009-02-09 23:05

Parsko and everyone,
Thanks for your suggestions.
Yesterday, I used another coil that had the same diameter as the length of the coil (my Q was 25). I got slightly better results with a series LC circuit. Still, the parallel circuit works the best, But I need a series one. I used a sine wave instead. I checked my grounds as Tracy had suggested. Here is what my circuit looks like. are the grounds OK?

sine wave>

|36 uH|

|.1 uF|--->Gnd from Function generator
| |
test point test point
+ scope -
Tonight, I will build a bigger coil 2 mH and experiment again

Thanks,
Cenlasoft

Cenlasoft · 2009-02-09 23:09

Hello,
The formatting is off on my last reply

sine wave>

|36 uH|

|.1 uF|--->Gnd from Function generator
| |
test point test point
+ scope -

Thanks,
Cenlasoft

Cenlasoft · 2009-02-09 23:10

Hello again,
The test points are supposed to be across the coil
Thanks,
Cenlasoft

Tracy Allen · 2009-02-10 01:33

Well, if the function generator ground and the scope (-)=ground are hooked together through the instruments, then in your circuit they would effectively short circuit the capacitor--> no resonance. Maybe. If the instruments are battery powered, or if the scope probe is differential, or the instuments are ungrounded, then maybe not. To be sure, flip it so that one side of the coil connects in common to both the function generator ground and to the 'scope ground, then no worries. If you have dual channel 'scope, put the second channel (+) on the function generator output, and sync the 'scope to that.

sine wave>-----ch2----|0.1 uF|----ch1----|36 uH|------->Gnd from Function generator & scope
     ch1 = 'scope channel 1 probe
     ch2 = 'scope channel 2 probe

As you tune through resonance, you should see the voltage across the coil soar upward, and the voltage from the function generator will drop, respectively, to a maximum and a minimum at resonance, and the phase relation between the two will shift by 180 degrees very rapidly as you pass through resonance. The phase relation at resonance will be 90 degrees. The voltage from the generator drops because its source resistance forms a voltage divider with the reactance of the tuned circuit, and that becomes minimum as a pure resistance at resonance.

Be sure the 0.1uF capacitor is a film or at least an X7R type. Not a Z5U.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Cenlasoft · 2009-02-10 01:41

Hello,
Thanks Tracy,
I will try your suggestion and post. I believe that was the problem.
You all are great
Cenlasoft

Beau Schwabe · 2009-02-10 03:48

Cenlasoft,

I believe Tracy nailed it!

FYI) you can also exchange the Coil and Inductor Capacitor in a series resonant configuration.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 2/11/2009 4:30:52 AM GMT

Carl Hayes · 2009-02-11 01:29

Beau Schwabe (Parallax) said...
Cenlasoft,

I believe Tracy nailed it!

FYI) you can also exchange the Coil and Inductor in a series resonant configuration.

Um, you can exchange the coil (inductor) with itself?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
· -- Carl, nn5i@arrl.net

Cenlasoft · 2009-02-11 03:20

Thanks to everyone,
My series lc circuit works. I will build a few big coils tonight and try different values. I only have a single channel scope, but I will order the pc parallax scope tomorrow.
I post my results as soon as I can.
Cenlasoft

Beau Schwabe · 2009-02-11 04:29

Carl Hayes,

"Um, you can exchange the coil (inductor) with itself?" - lol, I meant the Coil and Capacitor.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Cenlasoft · 2009-02-15 19:40

Hello,
I have been trying all the great suggestions that everyone has given me. I read the post on the LVDT and also Beau's inductive joystick. I don't have a propeller kit yet, but I will order one soon. I am now getting resonance with many different coil and capacitor pairs in series. I want to continue to build my eddy current metal tester and upon reading about the inductive joystick, I believe that the method that will work for me. I downloaded what I could from Beau's post and have looked at the pictures included. My questions are:

What were the values for the coils in the joysticks?

Is there is schematic that I can look at to modify for a two coil setup?

My series LC circuit is:

|1.2 mH|

| |
>SG--||

>scope + |
|
>SG

|
|

>Scope -

The capacitor is 560 pF. I want to thank everyone. My previous problem was my ground connections. My inexperience was the problem also, but I want to learn more.
Any suggestions would be appreciated.
Thank you,
Cenlasoft (Curtis)

skylight · 2009-02-15 21:00

Im not a metallurgist but could sound such as ultrasonic be used? different densities /characteristics would each type of metal have its own unique signature?

oops I missed the previous post suggesting this

Post Edited (skylight) : 2/15/2009 9:11:14 PM GMT

Beau Schwabe · 2009-02-15 21:18

Cenlasoft,

"What were the values for the coils in the joysticks?" - Those coils are about 55uH and adjusted to resonate at about 3.9MHz with the 5pF to 30pF variable capacitor.

"Is there is schematic that I can look at to modify for a two coil setup?" -· Inside of the code, the coils are paired as L1/L2 and L3/L4... just use one of the pairs.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Cenlasoft · 2009-02-15 22:02

Hello,
Thanks for the information. I appreciate the suggestions about the ultrasonic. After I experiment with the coils, I will try the ultrasonics. This technique seems very good. Should I switch to the propeller platform or can I stay with the sx28 and an ADC? I will order the Prop anyway. Right now I am trying to wrap those small coils. I am using 30 gauge magnet wire around an 8 mm form as in the pictures by Beau.
Thanks again,
Cenlasoft

Beau Schwabe · 2009-02-16 01:11

Cenlasoft,

The coil as you have it using 30 gauge wire on an 8mm form will give you about 29uH of inductance if you are winding them the same way as I describe.

If your wanting to make an equivalent coil with the 30 gauge on the 8mm form, you will need about 70 turns as opposed to the 50 turns.

If you want to play with some numbers using various wire gauges as well as different coil forms, then you can use this Excel spreadsheet that I put together.· For your coils you want to use the lower section labeled Special Case Multi-Layer where l equals b ... the GREEN fields are all that you want to mess with.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Cenlasoft · 2009-02-16 02:03

Hello,
Thanks very much for the spreadsheet.
Cenlasoft

Coil as sensor

Comments