Help: Sensor Design - to measure Electrical Conductivity

[xishveirx]

Hi Everyone,
This is my first post here, so have mercy on me

Project: Using a Non-Contact Probe (Sensor) To detect an unknown 1"x1" cube's (it can be steel, alumnium and plastic) Electrical Conductivity (EC) and hence decide its material. Measurement of EC must be in levels of HI MED or LOW

Theory to measure EC: Parallel Plate Capacitor with the presence of a dielectric (cube) would change the capacitance C of the parallel plate capacitor. Since Q=CV must remain true always V would change to compensate the change in C
It would be a non-contact probe, which means I cannot touch my plates on the cube. Therefore, I could have my plates connected 90degrees on one end ("L" shape) to get as close as possible to any two adjacent sides of the cube.

To show HI MED or LOW levels, I would have a comparator which would decide based on pre-determined voltages. Output would be 3 LEDs to display the 3 Levels.

Circuit & Implementation: Undecided, and need help here
1. What's the best material to be used for the plates?
2. Do I actually need a circuit to collect charge and form Capacitance or
3. I could provide Vac through the plates and measure any changes in Voltage by placing a cube in between the plates
4. Any other suggestions/comments please?

Since it is fairly early stage of design I would still consider any other designs/concept. Most importantly, I would appreciate a circuit diagram with the concept to conduct testing asap.

Thanks to everyone reading this post.

 

[chroot]

Sounds to me like you're just asking us to do your project for you. What have you figured out so far on your own?

- Warren

 

[berkeman]

Hi Everyone,
This is my first post here, so have mercy on me

Since it is fairly early stage of design I would still consider any other designs/concept. Most importantly, I would appreciate a circuit diagram with the concept to conduct testing asap.

Well, you won't get a circuit diagram from us at this stage, but you will get some hints and questions to motivate your thinking. Here are some initial questions to get you thinking and doing some research (fun project, BTW):

-- What are the differences in material properties between the three cubes? What are the best property differences to use to keep the detector circuit simple?

-- If you could use a standard laboratory instrument and your own passive probe setup, which lab instrument would be the best to use?

-- How do metal detectors work? (Use HowStuffWorks.com if you don't have a better source) How do LCR meters work?

-- What kinds of things do you want to try to do in your probe setup to get the best possible signal strength out? That will make your measurement the most reliable, and minimize the external circuit's complexity.

-- Extra credit question for later: What would you do differently if you were allowed only to have your non-contact probe on one side of the cube, instead of being allowed to put plates or whatever on two opposite sides...?

 

[xishveirx]

hey chroot & berkeman thanks for the replies.

Chroot,
I apologize if I gave the impression that I wanted a direct solution to my project. Up till now, I had done all the research and work (up to a point where I'm confused - intimidated)

In fact all the theory and implementation is my work. I had only been given the project title and requirements by Mr. Customer who has strictly indicated that he/she needs to know the EC and Magnetic Permeability (MP) of the three cubes. I am actually required to design two sensors - for EC and MP respectively. I had only mentioned EC here since I had a clear idea on doing the MP sensor (using two coils to measure difference in induced emf)

Berkeman,
Here are the answers toyour Q's

1. The material properties differ in terms of EC & .Steel is most conducting and has higher MP (increases flux in magnetic field). Aluminum comes 2nd. Plastic would not conduct or affect a flux in a magnetic field.
In terms of the best property to use: I don't have a choice as I have to meet the reuquirements placed by Mr. Customer to implement both sensors (to measure both the EC & MP). Hence, I came up with a parallel plate capacitor sensor and 2 coils sensor. I hope I answered your questions.

2. Oscilloscope would be best to observe the change in levels (Magnitude) of AC signal.

3. As for metal detectors: I understand now that they work on the concept of Vac which creates electromagnetic fields in transmitter coil and receiver coil. We don't have lcr meters in our lab though, its mostly DMM, function generator, oscilloscope, PS..etc

4. My probe setup should be easily handled in one hand (relatively small). I could convert AC to DC using a rectifier circuit. To get the best possible signal strength out I could amplify my Vout before running it through a comparator (I still am unclear how I pre-assign voltage levels on the comparator, but I will find that out soon)

My concernsnow are:
1. I am confused whether or not my initial approach was correct?
EC - using a parallel plate cap. and MP - using two coils and induced emf.

2. I am a bit more clear on using two coils for MP but still unsure on what concept should my design be for measuring EC.

3. I appreciate the help, but it does make me due to strict deadlines.

Thanks guys for the feedback and input

 

[berkeman]

It sounds like you are on the right track. The coil sensor arrangement will detect the steel, and the parallel plates will discriminate between the plastic and metal cubes. In both cases, you will be putting an AC voltage into your coil/plates, and measuring the AC current that flows. For simplicity, you can probably pick a single test frequency and use the same voltage source for both measurements.

BTW, the instrument that I was referring to is an Impedance Analyzer, like the HP 4194. It let's you put in AC excitation waveforms into a device under test, and observe what the AC impedance does versus frequency. Kind of like an LCR meter with extra features...

 

[xishveirx]

Yeah thanks berkemen

I have changed my design to detect magnetic permeability now
Instead of using two coils i resorted to 1 coil that detects the phase difference.
It took me a lot of time to make my own coil and pick the best RLC circuit setup. There was a change in phase (between reference input and output across the coil)
Only problem is now the parallel plate setup (a simple RC circuit) does not seem to be sensitive. Variations apparently are in mV range only
For example if Vin=5V with a 1k Resistor (most of the voltage drop occurs across the plates) The voltage drops from 4.69V to about 4.66V when I place the steel cube. For alumnium it drops to about 4.67V
This is giving me problem as my comparator will not be able to be that sensitive to set levels of HI MED & LO.

For the RLC circuit setup (for measuring magnetic permeability) I decided to use a phase detector: XOR gate with input1=Vin and input2=Voltage across the coil. The bigger the phase difference the wider the pulse out of the XOR output. Again I don't know how to distinguish a pulse's width to set levels of HI MED & LO. For example, a comparator would work for magnitude levels, but what will work for pulse width's?

 

[berkeman]

You can probably increase the sensitivity of the capacitance measurement by using a higher excitation frequency.

Also, you can measure a pulse width by driving an integrator with the pulse, and measuring the peak output voltage.

 

[xishveirx]

You can probably increase the sensitivity of the capacitance measurement by using a higher excitation frequency.

Also, you can measure a pulse width by driving an integrator with the pulse, and measuring the peak output voltage.

Thanks berkeman, I might increase my frequency if time allows me to do so.

However, do you by any chance have a circuit of a good integrator using an op-amp that needs about +/-5V?

 

[berkeman]

An opamp integrator circuit is pretty basic, and your textbooks should cover their design. If not, I googled opamp integrator tutorial, and got lots of hits. Here's a helpful one from National Instrument's "NIZone" website:

http://zone.ni.com/devzone/conceptd.nsf/webmain/BD2DA2B16894BBEC8625704C0072E4A2

BTW, in addition to the basic simple opamp integrator circuit, you'll need to put a resistor in parallel with the feedback/integrator cap to give you a bleed-off/recovery time constant. There could be other little hooks and handles that you will want to put around the circuit as well. The NIZone info will probably have some good suggestions for some opamps that will work for your supply range. Always pay attention to the input and output voltage ranges of opamps -- only the most expensive ones will handle rail-to-rail signals.

 

[xishveirx]

thanks berkeman,
ill let you know if this works or not