Connections drop and don't come back (and "SoftDevice Controller ASSERT: 53, 296")

RE: Connections drop and don't come back (and "SoftDevice Controller ASSERT: 53, 296")

David Ormand — Sat, 15 Mar 2025 01:23:38 GMT

Thanks. Moved to 2.9.1 (wow, lots of differences esp. with "sysbuild"), not seeing the problem any more. More testing will tell.

RE: Connections drop and don't come back (and "SoftDevice Controller ASSERT: 53, 296")

Einar Thorsrud — Wed, 12 Mar 2025 13:17:19 GMT

I see. Then you are most likely seeing the issue discussed in this thread. This is fixed in SDK 2.6.3, so I suggest you move to that (or a newer SDK release for that matter, but this is the same as 2.6.0 just with some important issues fixed).

RE: Connections drop and don't come back (and "SoftDevice Controller ASSERT: 53, 296")

David Ormand — Fri, 07 Mar 2025 17:13:46 GMT

Yes, thank you, I had seen that post, so I was aware the assert was due to missing a scheduled event _down in the SoftDevice_. What is different/problematic about my situation is I only see this (and the connection callback errors, and the characteristic read callback errors, and the occasional device resets) when the signal strength is low (due to antenna or generally RF problems). It works fine when the Central's antenna is facing the Peripherals.

I've also taken measures to cause as much processing as possible in the main() thread context or kernel workqueue contexts, rather than in callbacks, so it wasn't obvious where the "overstay" would be occurring.

RE: Connections drop and don't come back (and "SoftDevice Controller ASSERT: 53, 296")

Einar Thorsrud — Thu, 06 Mar 2025 12:42:13 GMT

Hi,

The assert is triggered by an "overstay" event in the SoftDevice controller scheduling related to BLE central operation. Please see this thread for more details.