• TI Thinks Resolved

LDC1000: Moving Target Application

Prodigy 130 points

Replies: 15

Views: 398

Part Number: LDC1000

Hello TI Team,

We are using LDC1000 for detecting a moving target(ex : moving metal wheel) of heavy mass. Anyhow we are having two consecutively placed inductive coils from which we are reading Rp measurements. We have set the thresholds for with target and without target from our processor. We have managed to detect targets and achieve our application. But my product seems to fail detecting the target sometimes. We have diagnosed that the Rp value changes along with temperature and various others environmental factors. Hence we have designed our application as to change my default value of threshold along with the realistic change. But after our product works for more than 24 hours in a field environment sometimes the LDC is reaching a point where the RP measurement maybe stuck above the target detecting threshold value. Under what circumstances does the Rp change alot like this? How do we control this? Under What circumstances does the Rp value get struck at a particular region? Our target is a moving wheel probably at a speed of 250 km/hr whose detection has been achieved by our R and D team. The problem is the fault which is happening prior to working of more than 24 hours in field open environmental condition. If you prefer we can send the log of Rp change along with temperature which we have measured in house climatic chamber and also in environmental field conditions. We request you to kindly suggest us the resolving method for our design. I would kindly request you to help us by sending a technical team if possible who can help out our design. 

Please Note - a) We have tried using LDC1101 which is not suitable for our application since it not able to provide us comparator interrupt output at fast moving target condition.

b) We have tried measuring inductance/ freq count from LDC1000 which does not seem to have much variation for my metal target. 

  • Hi Venkatesh,

    The LDC1000 is currently not recommended for new devices, and we cannot guarantee the part's future availability. I strongly recommend using a different part. We have the LDC Device Selection Guide which may help, but it sounds like the LDC161x could be a good fit. The LDC161x is our highest resolution inductive sensing device, so it may be able to detect shifts in inductance that the LDC1000 cannot. It also has an interrupt pin (INTB) like the LDC1000, which can notify you as soon as a new data conversion is ready. In general we do not recommend Rp sensing, precisely because of its high susceptibility to temperature drift.

    If you could describe more about your system requirements I'd be happy to discuss the feasibility of using a different inductive sensing device for your application. It would be helpful to know:

    1. The target's material

    2. The size of the material

    3. The distance between the target and the sensor coil

    4. The type of sensor coil -- wire wound or PCB?

    5. The required sample rate

    6. What type of detection is required? Event counting to determine the wheel's angular velocity? Detection of the presence or absence of the wheel? Something else?

    Best Regards,

    Kristin Jones

    Sensor Signal Conditioning

    Applications Engineer

  • In reply to Kristin Jones93:

    Dear Miss Kristen,
    Thanks a lot for your reply.

    We have tried to achieve the sensing range of my target using LDC1612 but however could not achieve the same using the evalulation board. Anyhow we understand that the LDC1000 is a product which might get stopped, hence we might have to redesign the same using LDC1612.
    I understand that you recommend reading of inductance rather than Rp, Here are my questions related to the same
    1. Can we set comparator thresholds for inductance to give INT output from any device just like how we do for Rp?
    2. Will there be proportional variation in inductance along with Rp drift when my target is present?


    Further to this my answer to your questions are as below

    1. The target's material - Stainless Steel

    2. The size of the material - Thickness of 26mm with a diameter of one metre. ( Wheel)

    3. The distance between the target and the sensor coil - Max - 35mm

    4. The type of sensor coil -- wire wound or PCB? - Wire Wound ( Insulated Litz Copper Wire used)

    5. The required sample rate - I would have to get data for every 250 micro seconds(Rp/Inductance).

    6. What type of detection is required? Event counting to determine the wheel's angular velocity? Detection of the presence or absence of the wheel? Something else?
    - I would require the detect the presence or absence of wheel through which i need to achieve my application. Anyhow we have achieved the same and succeeded but the working in field environment has become an issue.


    I would like to explain in private message if anything more to this require. Please revert back with your valuable advise.
  • In reply to venkatesh murthy:

    Hi Venkatesh,

    In answer to your questions:

    1) No, the other LDC devices do not have the comparator threshold output available. We do have some inductive switch devices that have similar abilities, but they are not as sensitive as the LDC161x and would be less suitable for this application.

    2) Yes, both the inductance and the Rp will change when the sensor interacts with the target. I recommend reading this application note for more information about how the coil's inductance will change when interacting with the target. Section 1.2 includes an equation that relates Rp and L to each other, and section 2.4 shows some sample data for how Rp and L change based on target interaction.

    Based on the data you provided, it looks like the LDC161x could not meet your sample rate requirement. The LDC1101 would be a better option as it has a higher maximum sample rate, but does not have an interrupt pin like the LDC161x and the LDC1000. Unfortunately, I don't believe we have a device that meets all of your needs. If either the sample rate or the interrupt pin are preferences and not absolute requirements, it may be possible for you to get a working solution with either the LDC161x or the LDC1101

    In general, these things could improve your inductive sensing system:

    - The more conductive the target is, the stronger the inductive response will be. If it's possible to add a copper or aluminum coating or layer around the stainless steel wheel, this could improve the system's sensitivity. For more detailed help, please see LDC Target Design.

    - We generally recommend PCB sensors instead of wire-wound inductors because they have less part-to-part variation and are more stable across temperature. For guidelines on sensor size and design, please see LDC Sensor Design.

    Best Regards,

    Kristin Jones

    Sensor Signal Conditioning

    Applications Engineer

  • In reply to Kristin Jones93:

    Dear Miss Kristen,

    We have done the climatic chamber testing  for LDC 1312 using EVM, The acquired  plot from GUI is attached for your reference ,we have a drift of 1uH inductance. My target also results in a drift of approximately 1uH . Kindly advice how do i resolve this.

    For your reference i have attached the plot with my target moving at 250KMPH.LDC1312_Temp_test details.zip

  • In reply to venkatesh murthy:

    Hi Venkatesh,

    Could you clarify how long you measured the data? Without knowing several of your register setting I am unable to calculate the time frame from the number of samples provided. Could you also clarify if your target is present during the whole test, part of the test, or none of the test? 

    The easiest way to remove temperature drift from the system is to use a reference coil that is identical to the sensor coil. This reference coil should be placed such that it cannot sense the target but is still exposed to the same changes in temperature. The output data from the reference coil can be subtracted from the output data from the sensor coil, which removes the temperature drift from the sensor coil data.

    Best Regards,

    Kristin Jones

    Sensor Signal Conditioning

    Applications Engineer

  • In reply to Kristin Jones93:

    Hi Kristin,
    The data was measured for a period of 6 hours. The temperature cycling severity was -10 degree Celsius to +70 degree Celsius at a rate of 1 degree per minute and fixed temperatures for 1 hour at the lowest and highest severities. The target was not applied at the time of temperature cycling. But we have attached one more file(250.csv) having the data of my target moving over the sensor coil at a speed of 250 km/hr.

    We cannot use reference coil for attacking the temperature drift because the below reasons
    1. We have space constraint in our product.
    2. We will not able to match the detection at higher speeds by reading both the sensors reference values using I2c which can run at maximum 400 Khz only.


    By looking at the plot could you please confirm if this drift is realistic for our cycling or are we making a mistake in the configuration/hardware level.

    Our configuration in for temperature cycling using EVM GUI(LDC1312- EVM) was as follows

    * Channel - 0
    * Fin Select - 1
    * Parallel Capacitance - 330pF
    * Inductance - As shown in GUI - 54 uHenry.
    * Input Gain - 16(4 bit shift)
    * Input Deglitch Filter - 3.3 Mhz
    * Register Update Rate - 5 Seconds
    * Channel Sequencing Mode - Sequence Channel Measurements
    * Reference Clock Source - Internal Oscillator ( 43.4 Mhz)
    * Measurement Timings - Chennal 0 - Divider Code - 1
    Current Drive and Power
    * Power Mode - Enable Low Power Sensor Activation Mode

    Sensor Initialization current - Code 17 -/ Current(uA) - 196
  • In reply to venkatesh murthy:

    Hi Venkatesh,

    Typically, L-measurement shifts about 30ppm/degree C for a PCB sensor coil. Using the data you sent me I estimated your temperature drift to be roughly 1kHz across the full temperature range (if you used a PCB sensor coil). It looks like you are seeing closer to 10kHz drift across your full temperature range. I think this is plausible for a wire-wound coil. I would suggest using a PCB coil to reduce the temperature drift. This application note should help the design process, as well as this LDC Excel Calculator Tool. The "Spiral_Inductor_Designer" tab would be especially helpful for you.

    Best Regards,

    Kristin Jones

    Sensor Signal Conditioning

    Applications Engineer

  • In reply to Kristin Jones93:

    Hi Kristen,
    We are reading from LDC1612 through i2c in the following methods.
    1. Using microcontroller at 400 Khz continuously
    2. Using FPGA at 400 Khz continuously.

    We are not facing any problem in method 1, but when we are reading using method 2 there is alot of non target variation in the raw value. We have tried using all configuration recommended in datasheet also. Could you please help us out on this?

    We want to finalize our design by reading through FPGA itself such that it will act as a comparator for with/without target conditions.
  • In reply to venkatesh murthy:

    Hi Venkatesh,

    Is there a difference in the cable length from the MCU to LDC1612 and from the FPGA to LDC1612? Are the MCU and the FPGA using the same supply voltage and using the same I2C communication voltages?

    Best Regards,

    Kristin Jones

    Sensor Signal Conditioning

    Applications Engineer

  • In reply to Kristin Jones93:

    Hi Venkatesh,

    I haven't heard from you in a while, so I'm assuming you were able to solve your issue. If this is not the case, please feel free to post a new thread.

    Regards,

    Kristin Jones

    Sensor Signal Conditioning

    Applications Engineer