Attendees (unique): 12
UTC0705
Specialization of the MS56XX drivers. Motivated by cleaning up some std_moves.
Now each one has its own named driver.
Andrew: Is it still possible to swap out a sensor model and still have it working, like before?
Peter: AP_Baro probes to instantiate multiple sensors and keeps the first one it manages to. But it’s not a complete coverage of all the possibilities probably.
A: Doesn’t that mean that all 56XX will still be attempted?
P: No, because this new specific baro line overrides the ms56xx_type default type.
A: This isn’t a no-binary-change type of PR.
P: No, it’s not meant to be.
- And old hwdefs no longer compile, so that’s a good thing.
Merged!
UTC0728
A: In RealFlight the gyro rate is a couple of thousand (we interpolate the higher rates), but it doesn’t respect SIM_RATE_HZ.
- You should be able to differentiate that RF is running with EKF_TYPE=10. It’s handy when you expect the EKF to become unhappy regularly.
UTC0733
Merged!
UTC0734
Merged!
UTC0739
Andy: I was trying to increase the simulation speed, but it was hardcoded.
Merged!
UTC0740
A: The warnings were too verbose when we were testing last weekend.
Andy: I added a new parameter to throttle the message frequency.
A: There’s no real reason to limit these fence options to only those specific frames. It excludes ROVER, for example.
- It would be helpful if we would emit a float value of “distance to fence” on a MAVLink message. A STATUSTEXT message is harder to automate alerts on.
UTC0755
A 4-digit PR number!
A: In rovers with reversible throttle, you are supposed to arm by leaving the throttle stick centered and then pushing rudder left? If yes, then there’s only once motion to arm and also disarm, which can then happen while driving.
A: I think it’s missing the throttle check now, when checking for an arm/disarm.
- Trying to arm or disarm should be a combination of two motions, throttle down and turning right.
- Just turning left while flying and getting a “failed to disarm” message would be very scary.
P: No, I think getting an “arming failure” when the throttle isn’t zero makes a lot of sense.
A: We’ll have to ask Randy if he’s OK with the behaviour as it is with this PR.
- We need to decide what is the correct behaviour for attempting an arm and disarm.
P: We need to separate the arm and disarm cases. For sure we don’t want to try to disarm if the throttle is not zero.
UTC0840
Merged!
P: What about a subclass of Location that carries ONLY absolute altitude?
A: It would cost us a bit of flash. But most imporantly let’s not change master from 4.6 too much in the short term. Let the new altitude changes settle for a bit and then re-work it a bit later.