DRV5053-Q1: Basic questions

Hello Takumi-san

There are many factors that can affect accuracy and resolution. Some of these factors include: The accuracy/resolution of the sensor, your level of averaging/filtering, the level of calibration you are willing to do, the noise of your system, the type/size/shape of magnet you select, variability within your own system, the ADC you use, etc.

With that being said, if your primary concern is accuracy then I would recommend using the DRV5055 instead of the DRV5053. And if you are only using one pole of the magnet (which I suspect you are), then I would highly recommend using the DRV5056 instead (-Q1 versions are also available).

The DRV5056 is a unipolar device, so it will give you twice the resolution/voltage range as the DRV5055 if you are only using one magnetic pole.

So, to your specific questions:

Q1)

This requested accuracy/resolution with the DRV5053 may be very challenging to do, so lets look at the DRV5056A4 (which has the smallest output referred noise):

With a 5V supply, the linear output range of the device is 0.6V-4.8V, which is a full scale range of 4.8 - 0.6 = 4.2V. For a resolution of .5um (or .0005mm), you would need to be able to detect a voltage change of 4.2/(2/0.0005)=0.00105V, or 1.05mV. The noise floor alone of the DRV5056A4 is 3mV. This means that in order to achieve the desired accuracy, you would have to use an RC filter on the output and probably also do some averaging (based on your available bandwidth), in addition to system calibration (note that the magnetic field of a magnet does not change linearly with distance, but the sensor changes linearly with magnetic field, the output would need to be calibrated based on your specific magnet).

The DRV5056 is also designed to compensate for the temperature drift with a typical neodymium magnet, so if yo use a neodymium magnet, then you will not need to do as much temperature compensation yourself, if any.

Q2)

All 3 of the above mentioned devices only detect magnetic field in a single axis. If you want multiple sensing directions, you would need to use multiple sensors. Although there will be variability between individual devices, accuracy will meet the specifications given in the datasheet.

Q3)

From the diagram, I believe you are correct, this appears to compensate the bias current to the Hall sensor. As mentioned above, the DRV5056 has temperature sensitivity compensation designed to drift with a typical neodymium magnet, so you would be better off using that part for your tight specifications.


NOTE:
Also please note that the above advise are ideas to help you meet your requirements. The actual realization of meeting those tight requirements will depend on many factors, as mentioned before. I recommend that you look over the datasheet values and use similar methods as used in my answer to your first question to see how much they will affect your accuracy/resolution, and see if there are any ways to calibrate those variances out.

Also, I recommend using the "DRV5056 Distance Measurement Tool" which can be found at: http://www.ti.com/product/DRV5056/toolssoftware 

Hello Mitch-san,

Thank you so much for your kind reply and suggestion !!

I understood well.

Best Regards,