This post is older than 2 years and might not be relevant anymore
More Info: Consider searching for newer posts

How does connection synchronization account for drift?

Hello,

I am currently trying to find out how the BLE devices keep their connections in sync. According to the spec, the peripheral can have a slaveLatency of up to 32s (Vol 6B 4.5.2). While being asleep, the slave will use his sleep clock with an accuracy of 250ppm if calibrated. As both the master and the slave will drift, this drift accumulates to 1ms for both, when accounting for minimum and maximum drift for both. If the connection is held asleep for 32s, that would make an extended listening time of 32ms for the slave. I find that this is a rather long time to keep listening and it somehow alleviates the power savings when using larger connection intervals.

Now I understand that there is both clock jitter and clock drift and I am wondering if clock drift is a phenomenon that is rather constant over time and only changes every now and then based on temperature and other physical influences.

Also, how can I calculate the average consumtion of the nRF51422 while receiving or while transmitting? Do I have to add up all the peripherals and stuff from the product specification to get the full consumption or is there a summed up value available? Is there a similar graph to this one e.g.? www.powerguru.org/.../CC2540-Bluetooth-Low-Energy-System-on-Chip.jpg

Are my calculations above right or am I missing something?

Thanks, Marius

Parents
  • There is a similar graph available in the S110 Softdevice specification, "BLE power profiles" chapter.

    Your assumption is correct, when the device is sleeping for a long time between BLE connection events, the receiver on the peripheral device needs to be enabled before the actual transmission to compensate for worst case clock drift. However, choosing long connection interval does not increase the average current consumption in comparison with short connection interval. The average current consumption should remain the same.

    Further advise on current consumption can be found on this thread

Reply
  • There is a similar graph available in the S110 Softdevice specification, "BLE power profiles" chapter.

    Your assumption is correct, when the device is sleeping for a long time between BLE connection events, the receiver on the peripheral device needs to be enabled before the actual transmission to compensate for worst case clock drift. However, choosing long connection interval does not increase the average current consumption in comparison with short connection interval. The average current consumption should remain the same.

    Further advise on current consumption can be found on this thread

Children
No Data
Related