LIS3DH and LSM9DS1 accelerometer comparison

[Mich1]

Posted on April 13, 2018 at 14:16

Hello,

I am trying to compare LIS3DH and LSM9DS1 (accelerometer part) and I am suprised by the result:

The LIS3DH has the configuration: CTRL1: 0x67 (ODR = 200Hz, LPen=0, ) CTRL4: 0xA8 (BDU, 8g, HR)

   => sampling frequency = 200Hz and BW=ODR/9=22Hz

The LSM9DS1 has the configuration: CTRL5:0x38 (no decimation), CTRL6: 0x98 (ODR = 238Hz, 8g, bandwidth determined by ODR, ) CTRL7: 0x80 (HR, filsters bypassed) CTRL8: 0x44 (BDU, IF_ADD_INC)

   => sampling frequency = 238Hz and BW=105Hz

But when I capture data with both PCBAs tightly attached, the LIS3DH signals have more amplitude and seems to have a higher bandwith.

Did I missed something or is there a problem?

Do you suggest another configuration for the LSM9DS1?

time in sec.

Best regards

Mich

[Miroslav BATEK · Accepted solution]

Posted on April 16, 2018 at 14:23

The reduced BW of the LSM9DS1 is cause by the HR (high-resolution) mode. 

If you enable the HR mode the LP filter is added in the signal chain. You can configure the cutoff frequency using DCF bits. By default the cutoff frequency is ODR/50 so in your case 4.76Hz.

You can set the DCF [1:0] bits to 10, the cutoff frequency will be 26Hz and the results should be similar to LIS3DH.

[Mich1]

Posted on April 16, 2018 at 15:20

Thanks a lot for your answer.

A note inside Table 70 could be appeciated to indicate that DCF configuration is mandatory when HR is enabled, only figure 8 indicates it and LPF2 is not mantioned inside the accelerometer registers. DCF is neither mantioned inside figure 8 nor in 7.12.