Measured variation of wake-up intervals


#1

I don’t think that there’s a specification for the variance of actual wake-up times, but I’m wondering if there are any measurements that have been done. I’m finding devices with, say, a wake up time of 7200 seconds not waking up for for > 7500 seconds.

Is this size of error expected? If it is, do I have to characterise devices based on my own calibration, or are manufacturers compelled to provide measurements as part of the compliance testing?

I can imagine that it’s a pita trying to test for such variance (you’ve got to get statistically significant measurements of data with an effective measurement rate of, say, one data point per day).

For background, I’m trying to establish the probability that a device is dead/not contactable, so that, for instance, I don’t expect a motion detector to detect motion if it’s effectively offline.

tc


#2

Yes you are right, there are no tolerances specified for the Wake-Up Periods.
I would not be surprised to see something like 10% drift, this is more or less what is tolerated for the S2 bootstrapping process.

As a rule of thumb, the Z/IP Gateway considers that something is wrong when it did not hear from the device in longer than 2 Wake-Up Periods. e.g. the device may have woken up but could not reach the controller at time t and went to sleep again.

If the deviation is really too large compared to expected values, it should also be caught at the product certification.


#3

Dear All,

A device may sleep for 5 minutes before it wakes up again. The application will now examine if it is time to send the Wakeup Notification e.g. need to sleep 10 times (10 x 5 min) before sending the Wakeup Notification, or go back to sleep. But if the device is woken do to a button pressen or motion detected, it won’t know how much of the 5 min it as slept. The application may decide to use an average of 2.5 min in this case. But several actions within the configured Wakeup Notification periode, will make the Notification drift. So, I would also recommend 2 wake-up periodes for detecting device office.

BR
Carsten