Does Arduplane switch to secondary sensor if primary fails?
I believe it doesn’t (i can not find any info). Anyway it could be useful to have a second one for reference.
It does if the fault is detected, you can use EKF lane affinity so different lanes use different sensor instances.
1 Like
Reading this i can not fully understand.
I have a can arspd sensor and an I2C one.
Will the switch between them be automatic if a fault gets detected in the primary one? Or should i set something?
Does this works the same with gps and barometers?
I see i can switch manually using an RC switch, but i would like it to happen automatically also, if needed.
The issue is that the system only switches when a probe is flagged unhealthy, and quite often a bad reading (for example caused by rain partially blocking the pitot tube) isn’t enough to trigger that switch. If the airspeed ends up being overestimated, that creates a real stall risk.
Personally, I wrote a Lua script to force the switch from one probe to the other whenever the difference between the two becomes too large (> 3 m/s). The idea is to force—if it isn’t already the case—the use of the probe reporting the lower airspeed. That way, worst case I underestimate the speed, which is safer, rather than overestimating it.
But it’s a bit tricky, because there’s no direct airspeed binding available in Lua scripts, so I had to add the required bindings in my custom-compiled firmware. Unfortunately, I haven’t found a better way to get a truly satisfactory level of redundancy for airspeed probes.
1 Like
Without affinity all lanes use primary sensors for everything but IMU as long as the sensor is reporting healthy. With affinity each lane uses it’s assigned sensor. If sensors used by the lane aren’t consistent lane confidence in its estimate will drop, triggering EKF lane switch. It won’t always be reliable as sensor malfunction may look like valid input ex. clogged Pitot probe appearing as change in wind.
1 Like
If I understand correctly, only one airspeed sensor is used at a time based on health.
Would the development of blended (or averaged) airspeed between multiple sensors be possible?
In situations where it’s not possible to mount an airspeed sensor on the nose of an aircraft (camera in the way, etc.), and if one is mounted on each side of the forward fuselage where asymmetric airflow may be encountered, would blending be a solution?
Generally we expect coordinated flight with low sideslip so assymetric mounting shouldn’t be a serious issue. TBH I would expect one sensor to provide accurate data (within expected errors from probe angle relative to airstream) and the other to change value depending on particular alignment of static and dynamic ports in the flow.
Thanks!
Would slideslip or strong winds aloft have more of an affect on potential airspeed error with a sensor mounted on one side of the aircraft?
I’m a helicopter guy doing my first plane, and only bought one sensor, but still wondering if blending is possible, especially since it looks like up to 6 sensors are supported.
The error will be assymetric with assymetricall mounting. You shouldn’t be getting extreme sideslip, that would cause significant airspeed error, unless the probe is too far back and is shielded with high sideslip.
I’m using double pitot sensor since i started this topic.
My airplane has no moving rudder and i see no difference in airspeed from both sensors. One is digital sensor I2C and the other one is digital CAN (if it matters).
One thing you can do is to connect 2 pitot tubes (left and right) to one sensor only. This is how it is done on some aircrafts too.