What effective resolution can be achieve at 10 kHz sample rate PSoC® 4700 MCU for inductive sensing?

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
MiVa_4722436
Level 1
Level 1
First reply posted First question asked Welcome!

Hello there,

I am new here and I found interesting solution for my application based on PSoC® 4700 MCU. I need to build inductive sensor based on eddy current.

I understand that the PSoC® 4700 MCU measures amplitude of oscillation which changes with gap.

I need to sense

  • range 0.2-2 mm
  • sample rate 10kHz
  • resolution at least 12 bits
  • minimum latency (<50 usec)

I know that resolution and precision in general is a function sensor diameter, frequency and other parameters but before I design anything...

  • What is achievable effective resolution of inductive sensor based on PSoC® 4700 MCU if running at 10 kHz sample rate?
  • What would be the latency caused by measurement (how much time is used for measurement period and for data post processing)?

Thank you very much,

Michael

0 Likes
1 Solution
Hari
Moderator
Moderator
Moderator
750 replies posted 500 replies posted 250 solutions authored

Hi miva_4722436

You can find out the scan time (the time required to scan once) in the Scan Order tab in MagSense configurator.

pastedImage_0.png

The resolution, which is termed number of subconversions must be set according to the design. As mentioned in previous post, higher number of sub-conversions, the better will be the signal. It is recommended to start with 100 and increase it till you obtain an SNR of atleast 5:1.

Thanks,
Hari

View solution in original post

0 Likes
4 Replies
VenkataD_41
Moderator
Moderator
Moderator
750 replies posted 500 replies posted 250 solutions authored

Hi Michael,

If you are already aware of Cypress' mutual capacitance based capacitive sensors, the the concept of inductive sensing would be almost same.

First of all, we would like to tell you that the Inductance sensors from us are not linear sensors like for example ADC.

The inductive sensing IP measures the inductance of the sensors and converts them into raw counts. This conversion is non linear with distance from the target. Also there will be several environmental parameters, sensor design which will effect the inductance and therefore rawcounts. So it will be difficult to achieve the resolution you are looking for.

Coming to the sampling frequency, there is always a tradeoff between the rawcounts and the scantime. More the scan time, more is the rawcount value and lesser is the sampling rate. The process of achieving the perfect trade-off s called tuning. cypress provides a GUI tool for tuning.

Please go through the Inductive sensing design guide from Cypress Semiconductor an Infineon Technologies company which gives you an idea how Cypress' Inductance sensing IP works.

https://www.cypress.com/file/427771/download

Your questions on resolution and scan time will be answered in the next post after internal discussion.

Thanks

Ganesh

0 Likes
Hari
Moderator
Moderator
Moderator
750 replies posted 500 replies posted 250 solutions authored

Hi miva_4722436

You can find out the scan time (the time required to scan once) in the Scan Order tab in MagSense configurator.

pastedImage_0.png

The resolution, which is termed number of subconversions must be set according to the design. As mentioned in previous post, higher number of sub-conversions, the better will be the signal. It is recommended to start with 100 and increase it till you obtain an SNR of atleast 5:1.

Thanks,
Hari

0 Likes

Hello Ganesh and Hari,

thanks for quick answer. I am looking into further explanation.

I understand that the output is nonlinear since the characteristic depends on target distance. The closer the coil to target the stronger response can be observed and also corresponding resolution which might change based on distance.

For my usage if I would set conversions for 100 times it gives 33.33 usec latency for sensor frequency 3MHz.

sample_time_100conv = 1 / Fsensor * number of sub_conversion = 1 / 3 MHz * 100 = 33.33 usec.

My system can live even with 2-3 bigger delay so even 200-300 sub_conversion might be still okay (66-99 usec).

The question is whether the signal will be still clean enough to be used - SNR and stability which defines repeatably of such a sensor.

Thank you,

Michael

0 Likes
Hari
Moderator
Moderator
Moderator
750 replies posted 500 replies posted 250 solutions authored

Hi MiVa_4722436

You are not supposed to keep the Lx clock frequency based on the timing requirement. The Lx clock frequency is determined by

pastedImage_0.png

Please make sure that the Lx clock frequency satisfies this. Otherwise, the SNR that you obtain will be very poor and if incorrectly tuned, might give out a negative result as well.

Once the correct resonance frequency is obtained, you can check the timing constraints by altering the N = Number of subconversions. This can be lowered till you obtain an SNR of 5:1.

Thanks,
Hari

0 Likes