[Bug] SPEEDYBEEF405WING enters CMS/OSD state from stick movement on clean 8.0.1 config, freezing servo control

[qyg9qxyy8v-boop]

On a SPEEDYBEEF405WING target with INAV 8.0.1, moving the sticks quickly left/right can cause the aircraft to enter a CMS/OSD state and the servo outputs stop following stick input.

Receiver input is still alive and the FC does not hard-freeze, but the servo control behaves as if CMS/OSD has been entered.

This happens even on a completely clean configuration after Full Chip Erase, with no old diff restored.

Steps to Reproduce

1. Flash INAV 8.0.1 to SPEEDYBEEF405WING
2. Use Full Chip Erase
3. Do not restore any previous configuration
4. Only configure the receiver
5. Move sticks quickly side to side / aggressively through corners
6. Servo outputs stop reacting normally to stick movement

Expected Behavior

Stick movement should never enter CMS/OSD behavior during normal operation, and servo outputs should continue to follow RC input.

Observed Behavior

Before the trigger:

Arming disabled flags: CAL COMPASS ACC HWFAIL CLI

After the trigger:

Arming disabled flags: CAL COMPASS ACC HWFAIL THR CLI CMS OSD

So CMS and OSD flags appear after the stick movement trigger.

Additional Details

• FC target: SPEEDYBEEF405WING
• Firmware: INAV 8.0.1
• Clean flash: yes
• Full Chip Erase: yes
• Restored old config: no
• Only configured: receiver
• Receiver still reacts after the event
• FC stays alive
• Tasks continue running
• I2C errors remain 0

CLI status before trigger

INAV/SPEEDYBEEF405WING 8.0.1 Mar 28 2025 / 09:52:26 (ae47bcb)
GCC-13.2.1 20231009
System Uptime: 145 seconds
Voltage: 16.11V (4S battery - OK)
CPU Clock=168MHz, GYRO=ICM42605, ACC=ICM42605, BARO=SPL06, MAG=QMC5883
Sensor status: GYRO=OK, ACC=OK, MAG=OK, BARO=OK, RANGEFINDER=NONE, OPFLOW=NONE, GPS=NONE
SD card: Startup failed
I2C Errors: 0
System load: 11, cycle time: 523, PID rate: 1912, RX rate: 50, System rate: 9
Arming disabled flags: CAL COMPASS ACC HWFAIL CLI
OSD: MAX7456 [30 x 16]

CLI status after trigger

INAV/SPEEDYBEEF405WING 8.0.1 Mar 28 2025 / 09:52:26 (ae47bcb)
GCC-13.2.1 20231009
System Uptime: 207 seconds
Voltage: 16.10V (4S battery - OK)
CPU Clock=168MHz, GYRO=ICM42605, ACC=ICM42605, BARO=SPL06, MAG=QMC5883
Sensor status: GYRO=OK, ACC=OK, MAG=OK, BARO=OK, RANGEFINDER=NONE, OPFLOW=NONE, GPS=NONE
SD card: Startup failed
I2C Errors: 0
System load: 10, cycle time: 502, PID rate: 1992, RX rate: 50, System rate: 9
Arming disabled flags: CAL COMPASS ACC HWFAIL THR CLI CMS OSD
OSD: MAX7456 [30 x 16]

Tasks after trigger

Task list rate/hz max/us avg/us maxload avgload total/ms
0 - SYSTEM 9 783 0 1.2% 0.5% 13433
1 - PID 1964 253 172 50.1% 34.2% 73616
2 - GYRO 3984 20 10 8.4% 4.4% 7805
3 - RX 50 36 28 0.6% 0.6% 282
4 - SERIAL 99 506 16 5.5% 0.6% 89
5 - BATTERY 49 23 15 0.6% 0.5% 147
6 - TEMPERATURE 98 8 1 0.5% 0.5% 25
7 - BEEPER 99 8 4 0.5% 0.5% 83
9 - COMPASS 9 327 321 0.7% 0.7% 653
10 - BARO 62 241 232 1.9% 1.9% 2944
11 - ADSB 0 3 1 0.5% 0.5% 0
15 - TELEMETRY 499 19 2 1.4% 0.5% 328
18 - OSD 249 2153 7 54.1% 0.6% 633
19 - CMS 49 79 2 0.8% 0.5% 11
22 - VTXCTRL 4 2 1 0.5% 0.5% 0
23 - PROGRAMMING 9 11 8 0.5% 0.5% 16
25 - AUX 99 11 3 0.6% 0.5% 55
26 - SPORT MASTER 498 8 0 0.8% 0.5% 65
28 - GIMBAL 49 3 1 0.5% 0.5% 6
29 - HEADTRACKER 49 2 1 0.5% 0.5% 5

Notes

A similar older report exists for SpeedyBee F405 Wing on INAV 7.1 where stick movement could freeze servos and involve CMS behavior (#9931), but this report reproduces on a clean Full Chip Erase setup on 8.0.1 as well.

[sensei-hacker]

Yes, entering stick commands while disarmed will -- execute stick commands. By design.
Before flying / driving, arm your craft.

Same as you start your car before you drive to the store.
Pushing random buttons and such with your car off will do things too - like reprogram your key, or reset the transmission calibration.

Entering random inputs does random things, yes.

[qyg9qxyy8v-boop]

Thanks, I understand that stick commands while disarmed are expected by design.

However, in my case the problem is not limited to disarmed bench testing. I have also experienced the same behavior in flight, which is the real safety issue, because servo control freezes as if the FC entered CMS/OSD state.

I also tested this on INAV 9 after a clean flash with Full Chip Erase and no old configuration restored, and the same behavior is still reproducible there as well.

Also, I have used other flight controllers with INAV before and I have never seen this happen there, which is why this behavior on SPEEDYBEEF405WING stands out to me.

So from my side this does not appear to be only normal disarmed stick-command behavior.

[qyg9qxyy8v-boop]

Would it be possible to completely disable CMS/OSD stick activation on SPEEDYBEEF405WING?

I am asking because this still happens on my setup even though OSD/CMS-related options are already disabled on my side.

[sensei-hacker]

This might sound random, but are you by chance using DEBUG over MSP?

If you can reproduce an issue with it armed, that will be interesting. That could indicate a bug that needs to be fixed. I would want to actually fix the bug rather than hope that maybe it's related to OSD commands and just start disabling features on one specific FC, potentially leaving the bug to continue on the other 226 flight controllers.

Fixing a bug (or even confirming there is a bug) requires being able to reproduce it.

Ps - I noticed you said 8.0.1. There have been about 100 different things fixed since then.

[qyg9qxyy8v-boop]

I’m seeing this even on version 9.1. The OSD menu is disabled for joystick access, yet it can still be triggered by stick movements. When that happens, the aileron and elevator servos get stuck in a random/extreme position. This is happening on the SpeedyBee F405 Wing flight controller. I haven’t seen it on any of the other FCs I use. The same issue is also reported by another user in Issue #9931 on the SpeedyBee F405 Wing.

[sensei-hacker]

The same issue is also reported by another user in Issue #9931 on the SpeedyBee F405 Wing.

Please read the comments on the issue you linked. Thanks.

[breadoven]

However, in my case the problem is not limited to disarmed bench testing. I have also experienced the same behavior in flight, which is the real safety issue, because servo control freezes as if the FC entered CMS/OSD state.

Is there any indication it entered CMS in flight ? If not then there may be some other issue causing the servos to freeze. I assume this isn't a permanent freeze otherwise the plane would have crashed.

Why not test on the bench when armed, just disable the motor/remove prop obviously.

[qyg9qxyy8v-boop]

Closing this with my findings from additional testing.

At first I suspected a power/BEC problem, because I had a previous crash on another setup that was caused by a weak BEC. That made this issue confusing and sent me in the wrong direction for a while.

After more bench testing, I found that the servo “freeze” I am seeing here is related to CMS/OSD menu activation from stick movement, not a full FC lockup. I was able to trigger the behavior by moving the sticks aggressively, and the outputs behaved like CMS/OSD had been entered.

Important detail: even with CMS/OSD menu disabled in configuration, it could still be triggered in my setup.

What I observed on my side:

• INAV 7.1.2 on SpeedyBee F405 Wing
• when disarmed, I could not reproduce it anymore in my latest tests
• when armed, CMS/OSD could still be triggered from stick movement
• receiver/FC stayed alive; this did not look like a hard freeze

I also had a real crash during a test flight while using Autotune in ANGLE mode and making large stick inputs. Because of that, I cannot say with certainty that the crash itself was caused by the CMS/OSD trigger. It may have been a separate issue or a bad combination of factors.

So my current conclusion is:

• the CMS/OSD trigger from stick movement appears to be real on this hardware/setup
• but I cannot prove that my crash was caused directly by this specific bug

I am closing this from my side because I found a plausible explanation/workaround for the bench behavior, but I wanted to leave this note here in case it helps someone debugging similar servo freeze / CMS activation on SpeedyBee F405 Wing.

[sensei-hacker]

As noted previously, it would be very helpful if you could attach the log from any issue that occurred while armed.

Also note while disarmed, turning off the INAV OSD does not turn off stick commands.

[breadoven]

I tried to trigger this on this specific FC and it of course enters CMS when disarmed but not when armed no matter how much the sticks are moved or placed in the CMS trigger position.

• when armed, CMS/OSD could still be triggered from stick movement

Can you explain exactly what you did to achieve this please.