Just a point of information for those folks who, like myself, choose to pulse-power the ADXL355 in order to save power.
I found that when I left a long (>5s) interval between powerings of the ADXL355, I got sane results on the accelerometer output. But when I repeated this at 1s intervals, the output would remain roughly stuck at some arbitrary value (varying very slightly between cycles).
I surmised that what was actually happening was that the POR reset circuit was being activated after 5s of power down, but not after 1s (after 1s, not all the charge on the decoupling capacitors - and on capacitance within the ADXL355 - had been drained out). Writing 0x52 to register 0x2F (Reset) every time I power up solves this problem.
My question is: whilst I wait for 10+ms before attempting to read data (to allow the ADXL355 to settle), do I have to wait for this length of time until I start writing to the SPI/i2c interface? I've been able to get reliable results by waiting 2ms after powering up before first writing to the SPI/i2c interface and then waiting for another 1ms (having issued the reset command) before setting up the accelerometer's operational parameters. There is little in the datasheet (other than the blanket typ. <10ms Turn On Time) to guide me in this regard. Is MAnthony's reply to ADXL355 Reset Command @I2C indicating a delay > 170us based on some information that I've missed? Some insight into the operation of the SPI/i2c interface after power-up/reset (and the POR operation, reset sequence) would be much appreciated.
Many thanks in advance and kind regards,