indication timeout reliability

RE: indication timeout reliability

Petter Myhre — Wed, 01 Feb 2017 22:46:08 GMT

No worries :) Great that you figured it out. Maybe you can accept my answer by clicking the grey circle next to it?

RE: indication timeout reliability

639zw — Wed, 01 Feb 2017 20:06:43 GMT

Petter, I figured it out. It had to do with how ble_bas.c implements 'ble_bas_battery_level_update()'. Specifically it only sends the notification/indication if the value is different than the previous reported value... I was sending the same value all the time. I'm not sure why 2-3 indications would successfully go out before stopping. However, that's a different problem. Thank you for your help, and I'm sorry for wasting some of your time. :)

RE: indication timeout reliability

Petter Myhre — Wed, 01 Feb 2017 17:37:27 GMT

I guess there is a very small chance, but then it would be a bug. Do you have a sniffer log? Also, could you upload your complete project(s) so I can test them here? What SDK and SoftDevice versions are you using?

RE: indication timeout reliability

639zw — Wed, 01 Feb 2017 17:27:54 GMT

Petter, Thanks for the reply! I am sure the indication is actually sent since I get the Indication on the RF52 running as the Central. I have print messages to say when the Central receives an indication, and when it sends the confirmation. The Peripheral never receives a 'BLE_GATTS_EVT_HVC' event. Is there a possibility that the softdevice receives the indication but never sends the 'BLE_GATTS_EVT_HVC' event? That seems to make sense, since I would not get a timeout in that case. Maybe if the indication confirmation is corrupted somehow?

RE: indication timeout reliability

Petter Myhre — Wed, 01 Feb 2017 17:02:43 GMT

Sounds very strange that it misses the confirmation. It should either get a BLE_GATTS_EVT_HVC event or a BLE_GATTS_EVT_TIMEOUT event, see this MSC. Are you sure that the indication is actually sent? Are you checking the return value of sd_ble_gatts_hvx()?

Yes.
30 seconds.
Correct.
Correct.