According to Randy the know issue of altitude loss after a high speed run is at its heart a frame issue.This seems to me a very worthwhile discussion to have.Once the majority of us understand this, we will be less likely to muck up other threads.Take it away Randy...
Thanks for the feedback
After a few false starts, I can verify that the Zubax GNSS 2 modules work great with the latest firmware for the PixHawk (3.4rc1). I had been using the Pixhack, but it does not support UAVCAN. I switched to a PixHawk, but received a clone which did not use the usual DF13 connectors. I sent it back and bought a genuine 3DR PixHawk. After following the tutorials on Zubax's website and setting the barometer to 20Hz directly on the GNSS 2 module (not through Mission Planner), and setting GND_PRIMARY to '1' in Mission Planner, I got everything up and running. Now with the barometer readings up on the mast along with compass and GPS, the altitude holds MUCH better. Even at high speeds I am seeing a good altitude hold now.
I was looking for the explanation of altitude loss when moving copter horizontally forward (and other directions) - but I never found a simplest explanation. When the craft is moving forward it is leaning towards the front (pitch) i.e. at the constant throttle level it must lose altitude due to a decreased upright thrust. It applies to all movement directions and actually I observed the same behaviour - losing altitude when moving in all directions. So, it suggests that it has nothing to do with
frame, pressure bubble etc but rather with losing a vertical component of thrust due to a pitch/roll change.
it also has a lot to do with your thr_min and thr_mid parameters, set thr_min to the lowest possible setting which will spin your motors, than adjust thr_mid accordingly. I think what it does it increases the resolution of the PWM signal, so more fine throttle control is possible. Also, oneshot125 improves throttle control in stab mode noticeably even on relatively low KV motors - 900kv, this translates in better alt hold performance.
I think it's all of the above. I added a plexi glass cover over my pixhawk and it instantly solved the problem so it's definitely not just aerodynamics but air pressure. After all it uses a barometer that can detect the difference in pressure in under a meter!! So it's a sensitive little bugger...
I have an update on the Gill static pressure port 3D model. I've continued to work on the design and manufacture. At this point I believe that I have a feasible, cost-effective solution. The 3D print house, Shapeways offers relatively high resolution prints at a very competitive price.
Here is a link to an updated pressure port model that can be purchased from Shapeways.
The material is a scintered nylon, and several color options are available.
I'd like to enable an external Baro I've installed on my I2C bus but don't see the parameter GND_PRIMARY in any of the GCS apps (have checked Mission Planner, APM Planner and QGroundControl). Any ideas how I can switch to the external Baro or even a way to see what Altitude output it is providing as a test?
probably best to load up master on the board for this test. so if you're using the mission planner, go to the Install Firmware page and press Ctrl-Q. With this bleeding edge software at least, you'll certainly find the GND_PRIMARY parameter in the full parameter list (or full parameter tree).
Randy, I've been flying 3.4 today on my sawn-off (330mm) Turnigy Talon v2 framed quad (with AUAV-X2) and I must say it flies beautifully. With the external Baro, I get a much smoother lift off in AltHold than with 3.3.3 (with internal baro), the lift off is very decisive on raising above 50% throttle, whereas with 3.3.3 doing the same, the revs actually drop, then raise (like its recovering from some pressure related confusion), then it launches, but its definitely not as steady as with 3.4. Great work! As always with much respect and thanks, Paul
I wonder how is results with Zubax GNSS 2 GPS with HMC5983(temperature resistant) and barometer on itself?