Sorry for the wordy post. I had a crash the other day because I got too low and hit a tree. I was in alt hold. I was looking at my tablet / telemetry concentrating on it’s orientation and didn’t notice until too late it had dropped from 60 meters to 30. I’m not positive why that happened could be I moved the throttle while moving the yaw. But that brings up my point that, is it just me or is it hard to maintain an altitude even in alt hold? I know on a dji the left stick spring loads to the center and it will maintain that altitude until you make a change. I have the notches in the throttle and generally I don’t have a problem doing the yaw without moving the throttle BUT I do find it irritating I’m never quit sure if I have my stick in the right place i.e. is it slightly gaining or losing altitude? Slightly losing is BAD! There seems to be about a %30 ‘dead band’ in the center where the throttle makes no change when in alt hold. So to gain altitude one must push the stick past say %60 then to stop it from continuing to climb you must bring it down to around %30. Then I usually just try to put it about in the middle. I would MUCH prefer if the quad would simply keep the thing at the altitude it was when alt hold was engaged!!! My brother has a dji phantom and flies the thing 2000 meters away with complete confidence it’s altitude will stay the same and this is without even the advantage of telemetry. Is there something I can do to have similar confidence in alt hold with the pixhawk ??
Am I right to assume that you write about a Copter?
Without ANY information it is pretty much impossible to make any determinations!
Please post a logfile… Many things can influence alt hold.
A 450 quad. Sorry no log file. I was interested if the %30 or so dead band [ in alt hold ] so to speak of the left stick is normal or is there some setting I am unaware of? I would like to make that smaller so it required far less travel to effect altitude in alt hold AND know with certainly that the alt with not change so long as the stick is in the center. Obviously that would require proper set up with each craft but I’m not seeing anyway for that to be consistent despite my best efforts.
Also are there any thoughts by the 3gr guys on the alt hold mode being made to work like the phantom with a spring loaded throttle stick? Or maybe I should call it the altitude stick. On the dii this is basically just an altitude change stick as apposed to the pihawk set up where it’s throttle and altitude both which is where I have encountered inconsistencies albiet from pilot error no doubt but inconsistencies none the less.
I’m not trying to compare dii and the pixhawk. The pixhawk is miles ahead of the dii I think. But… I do think they have done some darn nice things with that platform and having the craft ‘stick’ where you put it regarding altitude in the same way loiter does horizonaly would be nice. I’d like to know I can push the stick forward and know it’s not going to drop 10 meters or more if I don’t have my left stick in the EXACT right spot. Seems like alt hold should me just that. Although it could be the user set up isn’t as good as it should be. 
A couple other observations. My quad is heavy. I’m lucky to get 10 minutes with the camera. I can eventually get the quad to maintain it’s altitude and confirm this with the telem but it take a 2-3 minutes to be sure. That’s %20-%30 of the flight time. Hey I’m good right I figured that out all by myself! 
Another thing… if it’s up there 50-60 meters it’s almost impossible to discern if it’s gaining or losing a little altitude simply by eye. I think this is what bit me the other day. Granted I have telemetry but seems one should be able to know his or her craft is ‘locked’ at that altitude without having to rely on that. In fact many people fly without that all all.
I moved this to Copter/3.1 because it’s software-related
Altitude loss can be caused by many things. Z-vibrations are one thing that can throw the altitude controller off. There also have been reports of changing altitude in fast forward flight due to aerodynamic effects.
That’s why I asked a log file.
The deadband is in the software and it’s intended - exactly for that reason, so that you don’t have take care about precisely centering the stick. I personally don’t know if it can be changed but I’m sure some developer will ring in here.
As for changing the sticks - you can use whatever transmitter you please with the Pixhawk and I believe, many if not most transmitters can be modified so that the throttle axis is also spring-loaded. For most quality-transmitters, stick gimbals are available as spares, so just buy one with 2 springs and you should be fine.
The function of the throttle stick in APM:Copter depends on the mode. In Stabilize it’s a throttle stick and in Loiter/alt hold, it controls the vertical movement of the copter. IIRC, since 3.01, you can take off in other modes than stabilize, so you actually don’t need to use this mode at all if you don’t want to. Of course, it would be wise to be able to control the Copter in Stabilize, because that way, you can always rescue it if one of the assisted modes should go bad, e.g. through bad GPS reception.
Thanks for the explanation.
I did auto tune and and then did several flights keeping close attention to the stick centered and watching the altitude on my telemetry in alt hold… and it does in fact keep the same alt with the stick centered. Much better than I thought. My crash into a tree was probably pilot error.
Flying around at high speed it would descend on pitch forward and ascend on pitch back 4 meters or so. Is this about the limitations of the baro?
I have also found that it can be tricky to hold an altitude even in Alt Hold, particularly when I’m flying FPV as I don’t have as much judgement of height, and it is very easy to move the throttle stick such that the craft is commanded to rise or fall.
Yes, a sprung-center Tx is the solution to this. Unfortunately, the programmers can’t do anything about that. 
I actually built myself a custom transmitter, and one of the features I implemented was the sprung-center throttle stick. I love it too.
Yes, you can rise-fall 4m when flying around due to Bernoulli effect on the barometer. Not much we can do about it at this point unfortunately.
[quote=“Rob_Lefebvre”] I actually built myself a custom transmitter, and one of the features I implemented was the sprung-center throttle stick. I love it too.
[/quote]
So does this mean you simply don’t ever invoke stabilize mode and always start in alt hold? Do you know if there is a way to make that ‘dead band’ smaller?? I wonder what would happen if I did radio calibration but only moved the throttle sticks 1/2 way each way… hmmm might have to try that.
One thing I don’t like about that dead band is your never sure what your stick input is doing because your not sure if your past the ends of the dead band. If there was an extremely small dead band and a spring to center TX stick you would know your input was effecting the altitude because you could feel yourself moving the stick against the spring pressure. This would give a secondary stimulus to what you perceive is happening not just visual which is often very hard if not impossible to discern at altitude!
Please look at your transmit manual, you can find a function called “lock the throttle”
You can set the center of throttle stick as the lock value.
When you want to keep in a level and do not want to change, you can lock the throttle in the middle with a switch.
Also, I’m using the Voice speak out the altitude when it changed more than 5 meters in Andropilot on my phone.
[quote=“govsux”][quote=“Rob_Lefebvre”] I actually built myself a custom transmitter, and one of the features I implemented was the sprung-center throttle stick. I love it too.
[/quote]
So does this mean you simply don’t ever invoke stabilize mode and always start in alt hold? Do you know if there is a way to make that ‘dead band’ smaller?? I wonder what would happen if I did radio calibration but only moved the throttle sticks 1/2 way each way… hmmm might have to try that.
One thing I don’t like about that dead band is your never sure what your stick input is doing because your not sure if your past the ends of the dead band. If there was an extremely small dead band and a spring to center TX stick you would know your input was effecting the altitude because you could feel yourself moving the stick against the spring pressure. This would give a secondary stimulus to what you perceive is happening not just visual which is often very hard if not impossible to discern at altitude![/quote]
Actually no. I typically arm in Stab mode, and hold the throttle stick down. Then I carefully bring it up off the ground, then switch to Alt_Hold. Then I remain in AH, and rarely use Stabilize. But I do typically take off in AH.
I supposed with more time and experience, I would takeoff in AH, but it was relatively new feature that was even possible to do that. I crashed that machine before I got a chance to get used to that mode.
I agree about never knowing where you are in the deadband. Everything was better with a sprung-center, and I was also planning on shrinking the deadband for the reasons you state. So you’re bang on.