Skip to content

Conversation

@gemenerik
Copy link
Member

@gemenerik gemenerik commented Feb 20, 2025

A short boot delay was initially introduced in commit 1f498f1 to properly initialize the GAP8 on the AI-deck. In PR #102, the delay was removed when no AI-deck was detected to optimize boot time. However, it now appears that the system has started to rely on this delay, as removing it sometimes causes the radio to fail at startup. The exact reason remains unknown, but this commit restores the delay to maintain stability.

A short boot delay was initially introduced in commit 1f498f1 to properly initialize the GAP8 on the AI-deck. Previously, no delay was needed. However, it now appears that the system relies on this delay, even without the AI-deck, as removing it sometimes causes the radio to fail at startup. The exact reason remains unknown, but this commit restores the delay to maintain stability.
@gemenerik gemenerik added the bug label Feb 20, 2025
@gemenerik gemenerik requested a review from ToveRumar February 20, 2025 14:22
@gemenerik gemenerik merged commit b0a1600 into master Feb 20, 2025
@gemenerik gemenerik deleted the rik/bootdelay branch February 20, 2025 14:29
@jonas-eschmann
Copy link

Hi @gemenerik, do you have plans to investigate this further (since "The exact reason remains unknown")?

I've run into a lot of issues with 2025.02 leading to "Too many packets lost" when integrating my own controller. It is quite hard to debug but after lots and lots of trying I found that flashing 2023.02 through cfclient first, then flashing my modified firmware (minimal modifications on top of 2025.02 + OOT controller) is completely stable. So I'm pretty sure now this issue is with the radio firmware (which remains unchanged when I flash the modified firmware). In my experience:

radio: 2023.02 & stm: 2025.02 => stable
radio: 2024.10 & stm: 2025.02 => unstable 1/5 => 1/10 reboots lead to issue
radio: 2025.02 & stm: 2025.02 => unusable ~1/2 reboots lead to issue

In the last case: once it happens is usually not recorvered by a simple reboot but only by removing the battery and letting it rest

Btw. this is all in combination with the FlowV2 deck, without it, I found it to be more stable in general

I would also like to run this on the CF Brushless but since 2023.02 predates it, I hope you might be able to find the root cause and fix it.

@ToveRumar
Copy link
Contributor

Hi @jonas-eschmann !

Im sorry to hear you are experiencing issues with this release. Thank you for supplying such detailed information about your issue!

To properly flash this version you need to have an updated cfclient. Have you made sure you are on the latest cfclient and cflib versions?
This radio fw introduces a delay which needs to be handled by the client/lib side. So if you are seeing this issue more with 2025.02 radio FW, this is likely to be the issue.

Could you try this and see if it helps your issue? If the issue remains, please start a thread on our support forum.

@jonas-eschmann
Copy link

Yes, I'm using cfclient 2025.2 and cflib 0.1.28

@bitcraze bitcraze locked as resolved and limited conversation to collaborators May 26, 2025
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants