Pixhawk the new flight controller from 3DR
- Thread starter Gary Mortimer
- Start date
R_Lefebvre
Arducopter Developer
Should work totally fine.
You guys have seen my videos right? Here's one of my latest, done with SRH. The gimbal isn't tuned perfectly, but the heli flight is very smooth as I hope you can see. The gimbal was operated by some guy in my club I just handed him the controller he'd never done it before.
This is from a medium-lift octo, single man operation on a 2-axis gimbal. Yaw tuning wasn't perfect.
And here's my latest work. Not quite the subject of this forum but it shows the full capability of the system:
I just got a contract to install a Pixhawk in a very large gas powered helicopter, so I will be busy with that the next few months and this video work will be taking a back seat.
You guys have seen my videos right? Here's one of my latest, done with SRH. The gimbal isn't tuned perfectly, but the heli flight is very smooth as I hope you can see. The gimbal was operated by some guy in my club I just handed him the controller he'd never done it before.
This is from a medium-lift octo, single man operation on a 2-axis gimbal. Yaw tuning wasn't perfect.
And here's my latest work. Not quite the subject of this forum but it shows the full capability of the system:
I just got a contract to install a Pixhawk in a very large gas powered helicopter, so I will be busy with that the next few months and this video work will be taking a back seat.
Last edited by a moderator:
Old Man
Active Member
Mr. Lefebvre,
After reading all of the preceding related to large heavy lift birds and Pixhawk, I have a couple of questions relative to using Pixhark in a heavy lift bird that likely has a lot going against it. I have an A2 I'm working with now that is it was human I would have killed it and put it out of my misery already. I also have a large frame quad with 28" props and 170 kv motors that generate their own set of issues.
Reading about the PID set up necessary to obtain a good lift off that could encounter issues in hover, and vice versa with obtaining good hover response but suffer at lift off, could Pixhawk work in current format if center C/G was moved to as low a point on the airframe as possible? The current plan has the aircraft using 10k ma 6s batteries but the voltage may be too low to obtain effective motor response with the large props. That destabilizes in and of itself. If the C/G was lowered might that not increase a natural gyroscopic effect that would increase the effectiveness of the controller? Re-locating batteries onto gear legs or constructing mounting points well below deck height would move the majority of non frame weight to a location less destabilizing. My thoughts anyway.
I have a need to find another controller and the domestic foundation behind Pixhawk is quite compelling but it would have to have a solid chance of functioning reliability. In your opinion, could it work? My next question applies to controller voltage input. What is the max input Pixhawk can safely handle? I may be moving up to a 7s or 8s battery to obtain better motor performance and propeller response. If Pixhawk cannot handle that what would be the best way to power the controller?
Thanks for your consideration.
After reading all of the preceding related to large heavy lift birds and Pixhawk, I have a couple of questions relative to using Pixhark in a heavy lift bird that likely has a lot going against it. I have an A2 I'm working with now that is it was human I would have killed it and put it out of my misery already. I also have a large frame quad with 28" props and 170 kv motors that generate their own set of issues.
Reading about the PID set up necessary to obtain a good lift off that could encounter issues in hover, and vice versa with obtaining good hover response but suffer at lift off, could Pixhawk work in current format if center C/G was moved to as low a point on the airframe as possible? The current plan has the aircraft using 10k ma 6s batteries but the voltage may be too low to obtain effective motor response with the large props. That destabilizes in and of itself. If the C/G was lowered might that not increase a natural gyroscopic effect that would increase the effectiveness of the controller? Re-locating batteries onto gear legs or constructing mounting points well below deck height would move the majority of non frame weight to a location less destabilizing. My thoughts anyway.
I have a need to find another controller and the domestic foundation behind Pixhawk is quite compelling but it would have to have a solid chance of functioning reliability. In your opinion, could it work? My next question applies to controller voltage input. What is the max input Pixhawk can safely handle? I may be moving up to a 7s or 8s battery to obtain better motor performance and propeller response. If Pixhawk cannot handle that what would be the best way to power the controller?
Thanks for your consideration.
Last edited by a moderator:
Av8Chuck
Member
I don't know anything about the Pixihawk but I'm curious what are you trying to lift with 28" props and 170KV motors?
Av8Chuck
Member
But what do you call a multi rotor moving and storage company?
I've seen a couple A2 crashes with large MR's that were over propped. They took off OK but started to become very unstable, kind of looked like they got "flushed" then crashed. These were really expensive MR's and the owners were not happy to say the least so I didn't stick around to ask what happened. I don't know what they hovered at, I'm guessing below 50%, but it appears that the A2 likes to be underpropped, although neither had any payload on them with batteries on top so maybe it was just a case where the CG was too high.
I've seen a couple A2 crashes with large MR's that were over propped. They took off OK but started to become very unstable, kind of looked like they got "flushed" then crashed. These were really expensive MR's and the owners were not happy to say the least so I didn't stick around to ask what happened. I don't know what they hovered at, I'm guessing below 50%, but it appears that the A2 likes to be underpropped, although neither had any payload on them with batteries on top so maybe it was just a case where the CG was too high.
R_Lefebvre
Arducopter Developer
I'm not following what you're talking about with PID setup for lift-off vs. hovering? Is this something with other flight controllers? I've never heard of this issue with ours.
About the CG: Some people think that having a low CG helps with natural stability, since the weight will "hang" underneath the lifting vector. This was actually disproved long, long ago when they were inventing rockets. It's called "The Rocket Pendulum Falacy". http://en.wikipedia.org/wiki/Pendulum_rocket_fallacy
The problem with this idea is that people think the weight is a pendulum, like in a grandfather clock. However, the reason a rocket and a quadcopter do not benefit from a low CG is because the thrusters always operate in line with the mass displacement. There is absolutely no natural stability. And for sure "gyroscopic effect" is not relevant at all, since neither the mass or the frame are spinning.
The best way to design a frame to give it the best chance of being stable, is to have the weight concentrated as close to the center of the machine as possible, and directly in line with the propellers. This creates the lowest angular momentum, and so will require the least amount of stabilizing corrective force from the motors.
I've heard that if you can't have the mass exactly centered, the next best place to put it is actually *above* the frame. But I've never seen any good explanation or proof of this.
Could Pixhawk fly your machine? Well, anything is possible. The inertia of those large props definitely hurts. I haven't tuned anything like that, but I'm sure I could get it. Worst case, it would require some changes to the code.
The Pixhawk requires a 5V input. There is no internal step-down regulator, you must use an external one. One nice thing, is that the Pixhawk actually has two power inputs, and can hot-swap between them, so you can have a redundant power supply. So you will need an external regulator of some sort. This is actually not a big deal at all. Any good quality BEC should be able to do the job, possibly with a little extra filtering to get rid of any spikes. One other option for a machine like you have is a secondary battery system. This is what I do on most of my helicopters using the system. The Pixhawk and Servos are powered by a 2S battery. The servos are direct to battery, and I use a very high performance linear regulator to supply the Pixhawk. Lots of people in the hobby community think that linear regulators are bad, but this is only because they don't know what they're talking about. Yes, they have been widely mis-used in the hobby industry, but that does not mean they are bad. They are the perfect choice for powering a flight controller. I use a custom made Micrel 29300 based system. It's extremely reliable, very low noise compared to a switching regulator, and has exceptional low-voltage performance. I would not use a linear regulator for a servo load, that would be a mistake. But that is why I'm not doing that, the regulator only runs the controller, it has no problems supplying 1W of power to the controller.
The other nice thing about the dual-battery setup, is you power up the flight controller, and do whatever you need to do, program a mission, etc, in complete safety from having any chance of an accidental motor start. The only downside, is that it's an extra battery to manage. This shouldn't be a big deal however, as it doesn't take much battery to power a flight controller for a whole hour.
However, as I said, you could also just use a good quality HV BEC to supply power from the main batteries.
Old Man
Active Member
Thanks for the informative reply! The linear regulator info was quite interesting.
Leaving the large copter out of this part of discussion, I also fly one of the discontinued 3DR quads with Ardupilot. Thinking about your comments on natural stability took me to something I did in error the other day with that craft. While relatively high in altitude, estimated 200'. I cut throttle input completely while in a fast descent. Almost instantly the quad flipped inverted with no propeller thrust to influence attitude. From there in continued to fall in a very stable inverted attitude until full throttle was applied, whence the copter, thankfully, righted itself. It is not set up for aerobatics... In this case the majority of the mass at the center plate is under the plate due to battery location. Thinking back, it's interesting the weight went to the "high" side while the propeller/drag side went to the "bottom".
Leaving the large copter out of this part of discussion, I also fly one of the discontinued 3DR quads with Ardupilot. Thinking about your comments on natural stability took me to something I did in error the other day with that craft. While relatively high in altitude, estimated 200'. I cut throttle input completely while in a fast descent. Almost instantly the quad flipped inverted with no propeller thrust to influence attitude. From there in continued to fall in a very stable inverted attitude until full throttle was applied, whence the copter, thankfully, righted itself. It is not set up for aerobatics... In this case the majority of the mass at the center plate is under the plate due to battery location. Thinking back, it's interesting the weight went to the "high" side while the propeller/drag side went to the "bottom".
R_Lefebvre
Arducopter Developer
Yes, I've also had this exact same thing happen to me. I had an Octocopter, with batteries strapped to the landing gear, which lost power at altitude and freefell. It flipped right over and came straight down like an arrow. I think it's more about aerodynamics, but I couldn't even hazard a guess as to what is actually going on.
Gary Mortimer
Member
I see your note about keeping everything in line. The best multi I have flown to date is the Falcon 8 from Ascending technologies, looks a bit dated now and is not easy to transport but it works. I am very surprised it is not copied more. Its very easy to see which way its facing as well.
Last edited by a moderator:
R_Lefebvre
Arducopter Developer
My H8 design which worked very well was "inspired" by the Falcon. Trick was avoiding their patent, but I was able to.
Why did you find it was hard to transport, I thought it folded up?
Why did you find it was hard to transport, I thought it folded up?
The manual tells you to do things and they don't work biggest pain in the rear. To update the firmware you have to plug it in only after it tells you to unplug it then plug it back in. I mean come on the manual tells you to leave it pluged in then cycle it (won't work)
Biggest pain in the rear. Then the settings to tweak can be more of a pain in the rear to set.
Old Man
Active Member
Do you have Mission Planner loaded on your computer? If you do it should be easy to bring up Mission Planner, connect your FC/AP via USB, select a baud rate or select "auto", click "Connect" at the top right corner of Mission Planner, and go to the Help menu. Bear in mind I'm flying APM but all the updates I get include those for Pixhawk so I'm presuming it is using the same GUI. There is a "Check for Updates" line item in the Help menu. It auto loads once you click it. Just don't down load any Beta versions.
Gary Mortimer
Member
You should have been trying to fly a quad or aircraft autonomously 10 years ago it was much harder still. Are you using the right wiki's?? http://copter.ardupilot.com/
dazzab
Member
It certainly can be for some. But as others have said, it is a very advanced piece of technology. It's certainly not plug and play as it gives you complete control over the functionality. The price you pay for that is having to read the documentation carefully and probably reading through forums and wikis if you get stuck on something. I think the software engineers involved have done a brilliant job creating software for us ordinary humans. Have you seen Mission Planner, Driodplanner etc, not to mention the firmware of course. Pretty amazing stuff all around. Oh yeah, it's free as well. Shared freely by individuals who sole goal is to just make cool things. Have a look at the qualifications of the engineers making all that come together. Very impressive.
Having said that, I think Ardupilot/Arducopter and associated hardware are just fantastic when you get everything in place and it falls together. But there are times when something is wrong and it seems like no matter what you do it just gets worse and worse. In those situations I agree that it's hard to justify the time and effort to work through it. Just today I've been working with a quad with an APM FC that until now has been just fantastic. After sitting for a couple of months I did all the upgrades and now I just can't get stable flight out of it. Crazy huh? I've spent more time than I have to continue on with it. So I'm going to bite the bullet and put a SuperX in it as I did on a hexa once. I'll loose all the functionality of Arducopter and all the upcoming cool stuff but I just need to fly and fly reliably which I know the SuperX will do for me.
So there you have it. Different flight controllers for different people and different needs. Things don't always go as expected. These are very complex machines that we are playing with. Such is life.
Old Man
Active Member
Something to consider. When one has an FC that seems to be doing everything much the way things should be, before updating software take and save a few screen shots of the previous settings so you have something to fall back on if the upgrade doesn't function as advertised. I try never to leave something behind that I can't easily get back to should it become necessary. If possible, burn the previous set up onto an SD card via the transmitter as well.
R_Lefebvre
Arducopter Developer
This is good advice. I always save a trail of parameter files that I can go back to. You don't even need to take a photo. Go to the Full Parameter List, and just save the parameter file. Then you can name it something like "Droidworx Hexa, AC3.1.5 V1.1.param" Or something like that. If you update to 3.2, and don't like it for some reason, you can back-date it
I just spent 3-4 frustrating days trying to get Arducopter to work on a heli for a competition, and couldn't get it to work. I tried everything with the parameters. Back-dating software, etc. Turns out it was mechanical, the yaw linkage was not working! But my brain was stuck in this box thinking that Arducopter was the cause. Sometimes this stuff can be frustrating, but it's not always the software's fault.
Old Man
Active Member
Thanks!
I've been involved with unmanned stuff for quite some time now and every aircraft component I deal with that's flight critical has its own param file. A leeson learned long ago was always to save the original param files for the avionics in case some new component failed to work as it should. The part you removed might have still been serviceable so at least you had something to go back to if needed.
You're trouble shooting efforts reminded me of some of my own experiences. We often look at the most difficult to deal with component/area first instead of troubleshooting from the easy working towards the hard. Done that too many times
I've been involved with unmanned stuff for quite some time now and every aircraft component I deal with that's flight critical has its own param file. A leeson learned long ago was always to save the original param files for the avionics in case some new component failed to work as it should. The part you removed might have still been serviceable so at least you had something to go back to if needed.
You're trouble shooting efforts reminded me of some of my own experiences. We often look at the most difficult to deal with component/area first instead of troubleshooting from the easy working towards the hard. Done that too many times