[MMA8452Q] - HF disturbances & Interrupts

[anon-ef9907ee3c]

Hello everybody,

We are developing an application using a MMA845Q for waking up the whole system through transient detection interrupts.

Our application also embeds a GSM chip for communication.

The fact is the GSM creates HF disturbances of 900 MHz that propagates everywhere in the circuit.

We followed RF design guidelines for the GSM, and also design guidelines for the accelerometer layout + decoupling.

However, we got a strange behavior happening very rarely, but that is problematic: it appears that the HF disturbance makes the chip generating a false transient interrupt (no real motion at all).

Here are my questions:

Is there any chance that these HF disturbances can disturb the sensor behavior ?

Does anybody already encounter that ?

How can we become robust against that problem ?

Thank you very much for your help.

Have a nice day,

Simon

[Joshevelle · Accepted solution]

Hello Simon,

Yes, it is possible that the high frequencies are affecting the MMA8451Q, are you using the internal High Pass Filter (HPF)?? Since you said that is "happening very rarely", the HPF should fix the problem:

http://cache.freescale.com/files/sensors/doc/app_note/AN4071.pdf

Please notice that you would need to have a very good RF decoupling design to not affect the electronics around the radio, if the decoupling is very poor, a high pass filter might not do the job either. I would also recommend you to contact the GSM chip manufacturer and ask them to verify your RF design.

Have fun!

Josh

[anon-ef9907ee3c]

Hi Josh,

Thank you for your reply.

Yes, we're using the internal HPF for transient detection, and we're having the unexpected interrupts triggered anyway. As RF noise is high frequency noise, I would think that the HPF would have no effect on it at all (it let pass high frequencies)...

By the way, we also contacted our GSM provider to review our RF design, we'll see how to improve decoupling if possible.

I'll close this thread once we got the feedback from our GSM provider, to make sure the problem comes from the GSM side.

Thank you again, and have a nice day !

Simon

[Joshevelle]

Simon,

Yes you're correct! I meant LPF, it might work :smileyhappy:.

Low Pass Filter is set in Register 0x0F

Josh

[anon-ef9907ee3c]

Hi Josh,

Thank you for your confirmation, and sorry for this late reply, we were performing additional tests.

From what I understood from the datasheet, the LPF seems to be usable only by the pulse processing function.

We're using interrupt through transient detection, that can only disable/enable a HPF. But disabling the HPF there does not seem to remove the HF disturbance. It's just that you add the low freq signal (including gravity).

Unfortunately, our app needs to get rid of gravity. I think the best thing we could do would be to design a band pass filter to remove the HF noise....

Anyway, we're trying to remove the GSM 900 MHz glitch using small caps around the accelerometer.

I'll hope we'll manage to do so...

Thank you again.

Best regards,

Simon