MMA8451Q: Impact of high-pass cut-off frequency on data

[tricon]

Dear Community!

We currently want to realize a project using the MMA8451Q where the following happenes (very simplified):

• The accelerometer is in sleep-mode with a data-rate of 12.5Hz. Moreover, the transient acceleration detection is activated
• When a transient acceleration is detected, it wakes up and switches it's data rate to 800Hz. Also the µC gets an interrupt from the accelerometer and the µC starts collecting high-pass filtered data from the accelerometer at 800Hz to search for the highest (transient) acceleration detected

My question is the following: When the accelerometer wakes up it switches its data rate dramatically from 12.5Hz to 800Hz. According to the datasheet of the MMA8451Q the high-pass cut-off frequency relates on the data rate. So it is changed on an transient acceleration event as well.

My feeling tells me that this can impact the high-pass filtered data collected right after the transient acceleration event (even for frequencies well beyond the cut-off frequency). Is this correct? Does the change of the cut-off frequency invalidate the first high-pass filtered data from the accelerometer after the transient acceleration event?

It would very be interesting for me how the high-pass is implemented within the accelerometer. Unfortunately, I did not find any information in the application notes.

Thanks for any help!

[TomasVaverka]

Dear Johannes,

I do apologize for this delayed response, I have just sent you the answers via your SR, but I am including them here as it might help others.

1. The high-pass filter doesn’t reset from sleep to wake transition.

2. The high-pass filter implementation is independent from the ODR but the cut-off frequency is a fraction of the ODR.

3. The first data after sleep to wake transition will not have the correct frequency response since the sampling frequency has changed. It will last the filter settling time.

4. However, for fast dynamic acceleration event, the spectral response for the frequencies well above the HPF cut-off frequency remain correct.

5. To summarize: Yes, the change of the cut-off frequency invalidates the first high-pass filtered data from the accelerometer after the transient acceleration event.

Please let us know if you have any further questions.

Regards,

Tomas

[tricon]

Dear Thomas,

thank you for your very detailled description! It will help us (and for sure others as well) a lot during development!

Best regards!