LDC1000 application as an angular position sensor.

Hello Golem,

For a rotation application in which you need full 360 degree measurement, you will need multiple sensors and either a multi-channel LDC or multiple LDC1000s. If you use 2 sensors, they should be 90degrees separated.

As for the target shape, consider a target shape in which you have a width=k*angle (the red area corresponds to copper):

Regards,

ChrisO 

Hi Chris,

This helps me a lot, thank you.

Also, the resolution about this sensor is 16bits on the datasheet, but for our application, the accuracy and repeatability are both important.
Is there any reference about the accuracy of this sensor. In my thinking the accuracy may also depend on the alignment or the tolerance of width.
So maybe under the best situation, is there some information on the datasheet to let me calculate the accuracy? Or there are more parameters will effect the sensor's reading.

Thanks.

Have a nice day.
Golem

Hello Golem,

There are several considerations that are present - first, is this a static measurement (i.e. the measurement is performed while the disk is NOT rotating) or with a continuously rotating object?

For higher resolution, you will want to have 4 LDC channels spaced at 90degrees, in a Q+/Q- and I+/I- orientation. The LDC can provide 16bits of resolution, but you need to carefully map that electrical resolution to the mechanical system resolution.

You also need to determine what sample rate you need for your system.

Regards,

ChrisO

Hi Chris,

I want to use it as a position feedback control for motor(located after gear set). It require about 0.1 degree accuracy+repeatability for our application.
So the measurement is not static, with acceleration and deceleration also.
And about the sampling rate, I think 10kHz is enough for feedback control.

Thanks for aware the alignment issue. It might be a problem for manufacture.

Thanks.

Have a nice day.
Golem

Hi Golem,

0.1degree accuracy may not be easy to attain. I expect you will need to use at least 4 channels of LDC in a differential quadrature arrangement, along with some software calibration.

Regards,

ChrisO

Hi Chris,

I have an idea about the calibration. Correct me if I'm wrong.

If I use one LDC sensor and the pattern you posted. Also I use an absolute encoder with 12-bit or higher resolution that can achieve 0.1 degree accuracy for calibration.I can get a table between the actual degree and sensor reading. Then the question is about the repeatability and hysteresis.(if the alignment is perfect)

About the hysteresis part, more sensors will improve it. But if the hysteresis is also around 0.1 degree, I think it just fine.
If there is no alignment error, I think more sensors will not effect the repeatability. So, is the repeatability also around 0.1 degree or higher?

Is this a possible way to get enough accuracy in an angular application? Or just enough for 0.5 or 1 degree accuracy application?
If not, is there more information about repeatability or hysteresis that we have to find another way to narrow it down.

Thanks.

Have a nice day.
Golem

Hi Golem,

We will have a reference design for rotational sensing posted in the next few weeks.

Regards,

ChrisO

Hi Chris,

Is this Reference Design out?

Thanks,

Michael Raleigh