This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LDC1314: LDC1314 coin type metal detection at maximum distance

Part Number: LDC1314
Other Parts Discussed in Thread: LDC1614, TIDA-00509, LDC1101, TS3USB30E

Hi,

I'm developing a inductive sensor using LDC1314.

The sensor should be able to detect a coin type metal (10mm minimum diameter, steel, 2mm thickness) object presence on a flat surface (25x100mm shape).

The trouble is to increase the distance of detection up to 10-15mm.

I tried with a single coils rectangular 4 layers but the frequency shift is limited to few percentual points and the detection distance is few millimeters.; resonant frequency is around 400kHz.

I'd be grateful if you can send me some suggestions on coil design:

- frequency: does lowering the frequency increase the detection distance?

- number of coils to cover the surface: am I supposed to use more than one coil? I tried a single 4 layer 25mm diameter coil with limited but sufficient performance. 3 coils side aligned?

- multiplexing: I can use multiple multiplexed coils (other built-in channels are already used in my application); shall I switch off the current before switching the coils (external mux on a single LDC channel)

I refer to the example:

www.ti.com/.../TIDA-00509

Thanks in advance

  • Hi Andrea,

    You can try LDC1614 to get better resolution and extend the detection range.

    Your sensor coil has a large inductance value but the AC resistance may also be quite large. You can use the calculator tool below to estimate the electrical parameters of different sensor configurations. Generally it is best to optimize for Q factor. Hundred kHz range should work well. 

    http://www.ti.com/lit/zip/slyc137

    When using a mux, it is not necessary to turn off the device before switching.

    Regards,

    Yibo

  • Hi Yubo,

    thank you for your help.

    I will test also the LDC1614; how does it increase the sensing range?I have found some help here now:

    https://e2e.ti.com/blogs_/b/analogwire/archive/2015/06/17/inductive-sensing-how-far-can-i-sense

    https://e2e.ti.com/blogs_/b/analogwire/archive/2015/11/16/inductive-sensing-target-size-matters

    I would reach a higher frequency shift (percentual of steady)  when target is present.

    Sorry for not having put here all the details; adding some comments I would read the presence of  coins (euro project) in a box which base is around 25x100mm. If possible I would detect a single coin presence. So this is a empty box coin sensor; do not mind to the small sensor (as full sensor):

    The note that I've found in many TI applications and data-sheets around distance and relative size of coil and target seems not applicable or I'm wrong in some interpretation.

    I mean that in case of rectangular coil shape not only the lowest side measure is the matter. So If I want to detect at 10/15mm distance the presence of a 10mm diameter coin which is the suggested  coil shape? Is multiple coils the easiest solution?

    I will use the spreadsheet you've shared (thanks); I've not exactly understood the right Q, L, R e.g. parameters of the coil are the optimized ones.

    Thanks in advance

    Andrea Regis

  • I would also add a question. I've verified (as obvious) by oscilloscope that also the amplitude of the LC tank is affected by the proximity of a conductive target.
    As from during current calibration is possible to read the amplitude on internal register (LDC1x14 family...)? Or am I supposed to add an external differential analog amplifier if significative?

    Regards
  • Hi Andrea,

    You had a few questions I would like to clarify:
    1) Sensing range of the inductor is not a function of the sensor frequency (ignoring the fact that larger sensors have a higher parasitic capacitance and cannot operate at as high a frequency).
    2) With your sensor having a SRF of 400kHz, you should not exceed 300kHz as your operating frequency.
    3) sensing 10-15mm of target range with a 25mm sized sensor will require a higher resolution device - especially since the target is smaller than the sensor (this reduces the signal also).
    4) When switching sensors using an external mux, you can switch while the sensor is operating, but you may have a voltage spike that can damage some muxes.
    5) For your application where you wish to identify the presence of a coin, you may want to use an additional channel as a reference to compensate for environmental shifts, as your signal is quite small.
    6) I do not recommend adding any active components into the sensor signal path. While you could use the automatic current drive mode of the LDC161x, you will see that there will not be a large sensor amplitude change when a coin is 15mm away.
    7) Use of multiple sensors is an option - you could make 4 25mmx25mm sensors which would have a higher SRF; you would need to mux between them, but they would have pretty much the same sensing range as the single coil, although it might be tricky to place the sensors adjacent to each other.

    Regards,

    ChrisO
  • Hi Chris

    thank you very much.

    Concerning your useful notes:

    1) Sensing range of the inductor is not a function of the sensor frequency: I have plastic parts on the sensor; probably the frequency is involved in different absorbtion (and propagation) because of different materials proximity.


    2) With your sensor having a SRF of 400kHz, you should not exceed 300kHz as your operating frequency: OK I have to verify the SRF; I think is more than 400kHz. This is the frequency I set by a 270pF capacitor.


    3) sensing 10-15mm of target range with a 25mm sized sensor will require a higher resolution device - especially since the target is smaller than the sensor (this reduces the signal also). LDC1614; probably S/N is not good enough. Anyway, I will verify testing the suggested LDC.


    4) When switching sensors using an external mux, you can switch while the sensor is operating, but you may have a voltage spike that can damage some muxes.I understand; I've not particular fast timings. I probably will put 0 current or switch off by I2C (if possibile) before muxing. I do not know how this is performed in the application TIDA-00509


    5) For your application where you wish to identify the presence of a coin, you may want to use an additional channel as a reference to compensate for environmental shifts, as your signal is quite small. What is the qualitative frequency shift target in case of normal enviroment noise? 1 to 10%?


    6) I do not recommend adding any active components into the sensor signal path. While you could use the automatic current drive mode of the LDC161x, you will see that there will not be a large sensor amplitude change when a coin is 15mm away. I agree with you. Yes amplitude variation it is not so wide; but also frequency shift is not. I've tested the device with 1M resistor decoupling without particular interferences, with some hundreds pF capacitor tank (probably it is necessary to add another simmetrical resistor to the other terminal of the coil or using a differential amplifier...):

    Something like  an AM demodulator....(RC constant of the demodulator is not so critical because of slow reading timing°; peak detector is more difficult because of frequency shift). So no hidden registers are available for amplitude measure?

    7) Use of multiple sensors is an option - you could make 4 25mmx25mm sensors which would have a higher SRF; you would need to mux between them, but they would have pretty much the same sensing range as the single coil, although it might be tricky to place the sensors adjacent to each other. Ok

     

    Thank you for your support

     

    Andrea Regis

     

     

  • Hi Andrea,

    I would like to mention that the LDC1614 is very high resolution, and is able to pull out very small frequency shifts - shifts of less than 1000ppm can be measured, but it is often better to look for relative changes than absolute changes in the output.

    As for your sensor, you should not need to use a 4 layer sensor; at the size you have, a 2 layer sensor will work fine and have lower parasitic capacitance. I would then recommend using a larger sensor cap than 270pF - try to go larger than 500pF if possible (a 2 layer sensor should be better in this regard).

    Finally, I would not recommend using a difference amplifier on the INAx/INBx lines - the tempcos of all of the different aspects of the system will be complex.

    Regards,
    Chris O
  • Hi Chris,

    thanks.

    I agree; ii's better the relative reading. This also depend on how fast the target is moving in the hopper.

    Concerning 2 layers I will take this into account when starting mass production; and I'll increase the capacitor value also in 4 layers version.

    I will not use the difference amplifier but a simple single ended active peak detector. The question is concerning if the amplitude reading can be an help. Do you also have example of metal sorting (in this case coin validation) applications?

    I've got another question concerning the use of flexibile pcb as in this case it should be useful to follow hopper's lateral profile.

    Regards

    Andrea Regis

  • H Andrea,

    If you want to obtain amplitude information, consider the single channel LDC1101, which has an Rp measurement mode that effectively measures the amplitude (more accurately, it measures the sensor signal loss). You will need to use a sensor frequency above 500kHz with this device.

    Regards,

    ChrisO
  • Dear Chris,

    thank you for your explanation.

    I would also ask you one opinion on TI part number use; I'm developping LDC1314 (or LDC1614) board but I need more channels. I will use TS3USB30E multiplexer instead of TS3USB30E as indicated in :

    www.ti.com/.../tidu954a.pdf

    Do you think I could have some troubles when the signal is too low due to target proximity (target in contact) in particular when restarting the oscillations (target going away)?

    Regards