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.

WEBENCH® Tools/LDC1614: distance sensing

Part Number: LDC1614

Tool/software: WEBENCH® Design Tools

Hello!

I designed the coil below with your tool and was not sure if I will be able to meassure the distance of a tiny metal disc (6mm diameter) from the pcb. The disc will have a distance between 1 and 5 mm away from the pcb and I want to be able to tell the distance with a difference of 0.1mm per step (40 steps in total). With the specs given below, would you think that would be possible? Any help will be much appreciated.

  • Hello Eveblaw,

    The target and sensor are a bit small for 5mm sensing range. Is it possible to increase the size of both? For robust 5mm sensing range, you generally want to have a sensor that is larger than 10mm in diameter.

    You may also need a reference sensor to compensate for environmental shifts.

    Regards,

    ChrisO
  • Hello ChrisO,

    Thank you for your quick response.

    The system is inclosed in a metal box that will be placed in an office, so temperatures will always be around ~20C degree with varying humidity. Would a reference sensor still be neccessairy in this case?

    Due to mounting holes on the pcb, the coil can only be placed with a cilindrical shape with a minimum diameter of 160 mil (~4mm) and max diameter of 300 mil (7.6mm).

    With the given design, at which distance do you think it will be possible to read at least 40 steps? Would there be another way of improving the result other than increasing the number of layers from 4 to an even higher number?

    I look forward to hearing from you.

  • Hi Eveblaw,

    Once you can drive the sensor, there is no advantage to increasing the number of layers. With a 4 layer sensor design, you should be ok. Note that you actually need to design using a Racetrack sensor design, not a circular design. Also, your sensing range is controlled by the shorter side, not the longer side of the inductor. I would recommend that you review the Sensor Design App note ( http://www.ti.com/lit/snoa930) for more details on this.

    You can use the LDC Tools spreadsheet tool (www.ti.com/.../slyc137) to calculate the characteristics of a racetrack sensor.

    I believe that you will be able to resolve 40 steps in the system you propose, but you there will be significant noise and you may see drift which will affect your performance. I strongly recommend a reference sensor, or if possible a differential sensor configuration.

    Regards,

    ChrisO
  • Chris Oberhauser said:
    Once you can drive the sensor, there is no advantage to increasing the number of layers.

    Does this mean that if a 2 layer design falls in line with the ldc1614 specs, that there won't be an benefit of using 4 layers over 2 layers? I changed 4 to 2 layers in my initial design and increased the sensor capacitance (from 250pF) to 500 pF. Image of the settings and results is available below.

    What is the benefit of a racetrack sensor over a circular design if the shorter side is equally large for both designs? The mounting holes on the pcb restrict me to an surface of up to 300x400 mil. That would mean that the racetrack design will be only 100mil longer on the long side compared to the circular design. Sorry if I sound ignorant, but I couldn't find anything about it in the app-note. I did find the quote below which kind of says the opposite of what you said:

    The inductorshapeis an importantcharacteristicof the inductorbecauseit determinesthe shapeof thegeneratedmagneticfield.A circularspiralgeneratesa moresymmetricalmagneticfield thanothershapes,and is the optimumshapefroman inductancevs. RSconsideration. In general,it is recommendedto use acircularinductorfor the highestpossiblesensingcapabilities,unlessspecificsystemrequirementswarrantthe tradeoffsof an alternativeshape.

    The quote above comes from page 2.1, page 5.

    Chris Oberhauser said:
    I strongly recommend a reference sensor, or if possible a differential sensor configuration.

    If I am not mistaken; a reference sensor is an unused sensor design with identical dimensions somewhere else on the same pcb that can is being compared with the actually used sensor?

    As always, I look forward to hearing from you!

    Kind regards,
    Eveblaw

  • Hi Eveblaw,

    Yes, if the LDC can drive the sensor, then you are fine. The 2 layer design is rather close to the minimum RP spec of the device - 1.038k, while the spec is 1k, so you could be a bit below, which can increase noise. You do have the option of the higher current drive on channel 0, which is set by HIGH_CURRENT_DRV.

    The racetrack shape is useful for applications where you may want to measure a lateral movement of a target and not have sensitivity in the response to orthogonal movement (e.g. let's say you wanted to measure movement in the X direction but ignore movement in the Y direction),.

    Another use for racetrack shapes is for very small sensors, such as button replacement. Instead of adding more layers for the sensor, you can increase the sensor area with the racetrack shape. It doesn't improve sensing range, but it can add enough inductance to make driving a smaller sensor easier.

    The reference sensor is indeed a second channel which matching dimensions on the same PCB if possible.

    Regards,
    ChrisO
  • hey ChrisO,

    Perfect, thank you for all this useful information!

    The HIGH_CURRENT_DRV bit won't help me out, since I will be using channel 1 and 2 as well, which don't have this feature. In addition to that, I would love to keep the pcb as cheap as possible, so if I can get away with a 2 layer design, that would be perfect. Another possible way to get away from the 1k RP value, is to lower the sensor capacitance from 500pF to 440pF. This means that the RP value gets away from the spec limit, but the sensor frequency is closing in on the 10MHz limit as you can see in the image below. Since the RP value was several thousand ohms away from the limit for the 4 layer design, I am not sure if I can consider this change of ~120 ohm as a noticable improvement. What is your take on this? Should I stick to the 4 layer design or would you think that 2 layers might be doing well enough?

    https://i.imgur.com/CE0ZrKT.png

    Since I am only interested in measuring movement in the Y direction, I have one question left about the coil shape:
    The max coil diameter is 7.6mm. The diameter of the tiny disc is 6mm. You said that the sensor area would be increased if I used a racetrack shape. Does this improve the SNR, even if the measured object's diameter is already smaller than the coil diameter?

    kind regards,
    Eveblaw

  • Hello Eveblaw,

    Using a racetrack shape will not improve SNR, but will add inductance and so that can improve the Rp. If you can use a racetrack shape, something like 7.62mm X 10mm should have a reasonable Rp. Another area to investigate is using a 5mil gap and a 5mil trace - you would be able to add several more turns, which would be a good 2 layer sensor.

    I still have concerns that the 5mm distance will be quite far for a 7.6mm diameter sensor.

    Regards,

    ChrisO