• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 207 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

LDC1614EVM: Changes in the standard deviation

LDC1614EVM: Changes in the standard deviation (continuation)

Gerald
Genius 9570 points
Part Number: LDC1614EVM
Other Parts Discussed in Thread: LDC1101

Hi Team,

May I ask for help? This concern was a continuation of the E2E post below:

https://e2e.ti.com/support/sensors-group/sensors/f/sensors-forum/1037487/ldc1614evm-changes-in-the-standard-deviation

Customer have done some more research now, especially regarding the Settling time, Cycling through measurements and so on.

Kindly see below details:

"I am using the LDC1614EVM, corresponding OneUI Application and have connected my coils and capacitors to CH2, where CH0 and CH1 are connected to the PCB-coil that comes with the card.
I keep testing with different number of equal coils in series and equal capacitors in parallel (10nF).

The default value for Settle count is 1024.
I cannot get the standard deviation to change by setting the Settle count to different numbers.
I cannot get the standard deviation to change by changing to “Repeat single channel measurement” from “Sequence channel measurements”.
I have done tests in the range 500kHz to 1,4MHz and the result shows that at higher frequency, the measurements look more stable and it shows in the standard deviation.

Approximately:
700kHz -> Standard deviation of 0,00065
1000kHz -> Standard deviation of 0,00025
1400kHz -> Standard deviation of 0,00006

At the moment, or at least from what I have understood, the proposed change of settle count did not solve the issue.
I would really appreciate further help."

We hope to received further assistance on this. Thank you so much in advance.

Kind regards,
Gerald

  • 0
    TI__Mastermind 18545 points

    Hello Gerald, 

    With the change in frequency of the sensor, the number of samples the device can get during the sampling window also decreases. You can change the sampling time by changing the RCOUNT setting in the device. By increasing the sampling time, you can get a finer resolution and lower standard deviation. More information on the RCOUNT setting can be found in the data sheet and the Optimizing L Measurement Resolution for the LDC161x and LDC1101 application note. 

    Best Regards, 

    Justin Beigel

  • 0 in reply to Justin Beigel
    TI__Genius 9570 points

    Hi Justin,

    I have received a response from our customer as shown:

    "When I run the OneUI Application, default values of Reference Count is set to 65535 ( =26214,10 us) which according to my understanding is the maximum value.

    That means that when I change the Reference Count value, I get the Standard Deviation to change, but only to worse values.

    I read the “Optimizing L Measurement Resolution for the LDC161x and LDC1101” = SNOA944 which gives some

    Do I understand correctly?

    • OneUI Application standard setting of Reference Count = 65535 gives the best standard deviation numbers.
    • Higher frequencies give less standard deviation (at least in the range 300kHz – 1400kHz).
    • If low standard deviation is a target, one shall work at high frequencies. According to the document, up to about 6MHz when fREF = 40MHz.

    I have now also started to wonder if the voltage to the coils is of too low amplitude. The amplitude is somewhere around 400mV peak-to-peak in current setup. I have read that it optimally should be somewhere in the range 1.2 to 1.8V. The only solution is to work with smaller capacitors, right?

    I will keep on testing and in the mean time try to keep you informed so that you can guide me in case you see something wrong in my approach."

    Kind regards,
    Gerald


  • 0 in reply to Gerald
    TI__Mastermind 18545 points

    Hello Gerald, 

    As you observed, the Reference Count of 65535 is max so any change would decrease the standard deviation so it is best to leave it as the max value for this case. 

    As the frequency of the sensor increases, the number of samples that can be gathered during the sampling window (set by RCOUNT) increases as well. This is why the higher frequencies also have a lower standard deviation. However, as you increase in frequency, the frequency resolution also slowly increases. This does not exactly translate to a resolution change in the target application but should be noted. Changing the capacitor is an easy way to change the sensor frequency as long as the Q factor of the sensor is still appropriate. The Inductive Sensing Design Calculator Tool (Rev. G) can help with the calculations for this. 

    The make sure the amplitude is between the 1.2 and 1.8V range, you can change the IDRIVE settings on the device. The device can use an automatic IDRIVE setting with the RP_OVERRIDE_EN bit as a tool to help set the IDRIVE registers. For best resolution, it is recommended to only use the automatic feature while determining how to set the IDRIVE. From there, it is best to hard set the register setting. 

    Best Regards, 

    Justin Beigel