ADXL357 V1P8ANA and V1P8DIG Decoupling

Hello Philip90 

The 1 uF is not needed anymore as we have discarded its use after some testing subsequent to Rev A release. (you should be fine with the 0.1 uF caps with relatively stable ground)

Hope this solves your doubt,

Mario SM

Thank you for your answer MarioSM. In that same area of the Rev C datasheet, there is also some text about a 200 ms restart time that isn't in the Rev A datasheet:

"When power cycling the ADXL356/ADXL357/ADXL357B, it is highly recommended to fully discharge the device to ground level (VSUPPLY and VDDIO = 0 V), for at least 200 ms, on each power cycle"

Is it known what failure mode can happen if there is not a 200 ms restart time at cold temperatures? I have an existing design without that kind of POR circuitry. I'm wondering how big of a deal it is and if I need to change it.

Thank you for your answer MarioSM. In that same area of the Rev C datasheet, there is also some text about a 200 ms restart time that isn't in the Rev A datasheet:

"When power cycling the ADXL356/ADXL357/ADXL357B, it is highly recommended to fully discharge the device to ground level (VSUPPLY and VDDIO = 0 V), for at least 200 ms, on each power cycle"

Is it known what failure mode can happen if there is not a 200 ms restart time at cold temperatures? I have an existing design without that kind of POR circuitry. I'm wondering how big of a deal it is and if I need to change it.

Hi Philip90 

Based on previous clients experience some axis may start making wrong measurements as you get close to -40 ºC. (problems related with how the internal LDOs handle somo load values)

If you experiment this kind of issue and the circuitry can not be handle I would suggest to develop a security script that; forces the reset of the part to default mode as many times as needed until the measurements makes sense, or your desired behavior is reached. 

Hope this helps.

For the wrong measurements, would they be close to right, or not close at all?

Hi Philip90 happy new year

I cannot guarantee anything, but since we would not be following the datasheet guidance, we can expect the part to perform out of the featured performance specs or simply get stuck.

If it works, how close to right or wrong will depend on your application and the source of the non-idealities, such as a noisy GND.

Hope this was useful.

Mario SM

Hi MarioSM happy new year to you too!

Thank you for your reply.

When the datasheet says

Philip90 said:

(VSUPPLY and VDDIO = 0 V), for at least 200 ms,

Is there a tolerance on the 0V requirement? It takes my VDDIO pin a few hundred milliseconds to drop to less than 300mV, but takes a few seconds to reach 0V. Am I ok letting VDDIO and VSUPPLY only dropping to 300mV before reapplying power? Or do I need to wait the multiple seconds required to truly reach 0V?

I have a lot of capacitance for other functions on the power rail that I can't get rid of

Hello Philip90 

Previous experience from other customers in a similar situation using a 10k resistor during 1 second where able to successfully (in most of their attempts) power cycle besides V1p8dig not dropping further than 0.4 V.

Again, this depends on several factors outside the datasheet limits, but I hope it can work as a (half) reference.

Regards,

Mario SM

Ok, should I change the 100K resistor mentioned in the datasheet to 10K? This is the one from the LDO outputs to ground right?

 Hi Philip90 

I would stick with the 100k datasheet suggestion, which is the value that came out of testing and discussion within our team, but if you find a problem, you know you have the lever to lower the resistor further.

"This is the one from the LDO outputs to ground right?"
yes.

Let me know if you need anything else.

Mario SM