Pixhawk Feedback?

[Hellsbells]

I been in the multicopter hobby for about an year now and I have used the Naze Lite, Naza-M and the SuperX. I had two flyaways with the Naza and the SuperX would crash out of control after making sharp turns. I am looking to move on to the Pixhawk. Anybody have any feedback on their experiences with the Pixhawk?

 

[Holger]

Pixhawk is good, but requires far more involvement that DJI to setup and tune.

 

[Old Man]

Plus 1 to the above. It's a very solid flight controller with extreme capability but one has to take some time to read through all the WIKI info and learn how it works. Outrageous functionality and there are some super easy to use flight trimming tools but they are not instinctive. If you ever have a fly away with a Pixhawk it will be because you didn't set up one of the numerous failsafes available. I fly both the APM 2.6 and Pixhawk software and I have not been at all disappointed. Matter of fact, a friend just dropped a Pixhawk onto a 700 quad after removing an A2. The Pixhawk default settings flew very close to perfectly and the auto tune feature did the rest well enough to permit immediate aerial work.

It takes a little work but you get more than you pay for. A lot more.

 

[jfro]

Pixhawk is good, but requires far more involvement that DJI to setup and tune.

I will also attest to that.

I have Xaircraft SuperX on low KV quad, (started on x8), and Naza V2 which has been on small quad, large quad, and now my x8 w/ low kv motors. No problems with them other than I didn't like the SuperX on the X8. No fly aways or crazy loss of control issues on either.

I have a Pixhawk on my test (550 ) quad now. It takes a lot of time to learn. I've had 1 small crash with it while flying low and in Loiter (GPS) mode and it was because it loses altitude in GPS mode more so than my other FC's.

This past week, I tried a quick grid flight on the Pixhawk (my first autonomous) and it flew the 3 minute pattern as I programed it to. It came back to the start point (RTL) and came down about 1/2 way. It paused and I switched out of auto mode to stabilize (I think) and within 2-3 seconds, the copter shut down and came down with no power. I have sent the logs to 3D Robotics to see if they can tell me what happened.

I'm not writing about the failure, but just that in my research before buying the Pixhawk, I sort of expected a few more crashes or episodes of "what happened" because of the increased features and general nature of open source software. This past one happened on a first try on the autonomous flight but also after recently using 2 switches for 6 flight modes which is harder to keep track of then just 1 3way switch for flight modes.

I'd say go for it, but expect a fair amount of time learning it and I'd suspect more crashes than you have had to date with your other FC's. The one thing I don't like about the Pixhawk is the way they connect the wires. In my recent crash, my quad weathered the crash with just 1 broken leg a very slightly bent arm which was easily fixed. However, my GPS puck pulled off the arm and 1 cable pulled out of the GPS puck and the other pulled the wiring socket off the circuit board of the Pixhawk. I now have a Pixhawk with no gps connection.

 

[Old Man]

What GPS are you using? The latest Ublox with compass is extremely accurate. Their older GPS was adequate and held altitude quite well but was slow to acquire at times. Ublox took care of that issue. I wonder, had you performed all the accelerometer calibrations? Was it possible the auto land mode was active for low battery but programmed with a descent altitude stop? Just checking.

 

[jfro]

I don't know on GPS. Bought everything from 3dr Jan 2014. Says 3dr on it. I was out of state, so re-calibrated everything.

Would autoland cause a complete brownout? Battery was fully charged and I usually get 10-12 minutes on flights. It was only a 3 minute flight although I did have a new camera on this which I didn't weigh. Battery is now showing 3.89 v per cell.

I started a whoops, what happened thread here. This isn't the place for my issue.
http://www.multirotorforums.com/sho...-Pixhawk-logs-after-crash&p=176246#post176246

 

[Hellsbells]

Why would I expect to more crashes with the Pixhawk than with the Naza or SuperX? I gave up on DJI after having two out of control flights and gave up on SuperX after they told me I was turning too quickly for a hexacopter.

 

[jfro]

Why would I expect to more crashes with the Pixhawk than with the Naza or SuperX? I gave up on DJI after having two out of control flights and gave up on SuperX after they told me I was turning too quickly for a hexacopter.

That was just my opinion. Kind of what I deduced after researching the Pixhawk before buying. I fly by myself and don't know anyone who flies multirotors so everything I know is from internet reading,a few private messages, & 1 1/2 years of flying experience.

I have built 5 MR's and in my 2nd year of flying. I mostly slow fly for video capture and fly line of sight. No fpv and high speed running around.

My understanding is that because 1. it is open source, expect bugs and debugging time, but lots of features. 2. there are a ton of settings that effect things, so there is a learning curve.

I personally have spent way more time setting up the Pixhawk then I did the Naza or SuperX. I endure it because there are so many more features to the software I'd like to use. I have had times when I wanted to chuck it as I couldn't get it to fly as smooth as the Naza. Same frame, motors, esc's etc..... It took 3 times trying to get auto tune to work and that was after a frame mod to tighten it up.

Anyway, mine is just one opinion and it's not worth that much as I'm pretty much a beginner with the Pixhawk. Keep gathering opinions or just get it and start learning. Plenty of peeps enjoying their Pixhawks....

 

[dazzab]

If you want to be on the bleeding edge, enjoy tinkering and have the time to learn all the idiosyncrasies of the Pixhawk then IMHO you will love it. It really is an amazing project. IMHO it's working much better for use in planes than in MRs though.

At one point I had APMs/Pixhawks in six copters and one plane so I have a lot of experience with them. I can't tell you how many crashes I had in that time. Thanks to the extensive logging and the help available on forums you can typically figure out why those crashes happen and it's rarely the fault of the flight controller but rather having a setting incorrect or just not understanding how it works. It's actually rare to ever see a bug cause a crash from what I could tell. And they publish absolutely everything so there's no trying to avoid responsibility either. These are stand up guys who are very very smart making a tremendous contribution.

But when I started to use it with a large copter things really didn't go well. I thought it would be a good opportunity to feed valuable feedback in to the project but one of the developers, who wasn't even involved, jumped in and had a pretty nasty go at me. So I decided to stop spending any more time on the project and just use my backup Wookong on the large copter. Interestingly enough, the copter that developer said was not air worthy and that I was essentially not skilled enough to fly (he even said I'd probably crash it in to a crowd of people), now has over 50 flights on it with no issues at all. So I swapped out the FC on my other two main copters as well and interestingly enough, ongoing issues with stability and reliability disappeared on them as well. I put SuperX controllers on them and now I spend my time flying and filming. I rarely put any time/energy in to the flight controllers.

Now gimbals! Don't get me started on those. There's a really good way to waste a year or two.

As always, your mileage may vary. But my advice for now is that for heavy copters or large investments in cameras/gimbals it just doesn't make good sense to use a flight controller that is still in development.

 

[Av8Chuck]

I haven't had any experience with the team of developers for this controller but I tried using the Open Pilot CC board which was open source. Loved the controller, learned to hate the developers. I think its just kind of the nature of open source, good in concept with some exceptional developers but there seems to be far more mediocre developers than really good ones and its those that get really defensive about the development.

 

[Hellsbells]

Thanks for all the feedback. I ended up going back to the Naza-M. Although there is a change of a flyaway with the Naza it seems that there is just as much of a chance with the Pixhawk or any other FC. I had the SuperX but it kept going out of control after making hard turns. Xaircraft support was okay until they could not figure out the issue and then stopped responding to emails. I had thought of purchasing the next Skyview but I decided to wait 6 months to see how it works out for other people since it is new.

 

[Av8Chuck]

Thanks for all the feedback. I ended up going back to the Naza-M. Although there is a change of a flyaway with the Naza it seems that there is just as much of a chance with the Pixhawk or any other FC. I had the SuperX but it kept going out of control after making hard turns. Xaircraft support was okay until they could not figure out the issue and then stopped responding to emails. I had thought of purchasing the next Skyview but I decided to wait 6 months to see how it works out for other people since it is new.

How hard did you have to bank before the SuperX lost control? Was this a problem in all flight modes?

The whole product support seems to be a cultural issue. Like a lot of support from China if you can get them to engage in a problem they focus on solving a problem quite well, but I get the sense that many Chinese interpret being direct as confrontation and once they perceive that they either don't know how to deal with it or simply choose to ignore it. At least with XAircraft its not difficult, at least at first, to get them to engage with the problem. I never got anyone at DJI to acknowledge that there is any problem.

I'm not defending XAircraft, I'm just trying to figure out why product support in general from China is not as good as it could be. I mean if your going to sell a product internationally how expensive and difficult can it be to hire a copywriter to translate your manuals so that they make sense to the people in the countries where your selling them? That might go a long way in easing the support burden.

Whats a Skyview?

 

[Hellsbells]

I didn't bank that hard as I have turned quicker with the Naza-M with no problem. I was dealing with a Xaircraft in Australia but it went nowhere then I was dealing with Xaircraft from China and they went quiet after no being able to figure out the issue and also trying to place blame on my multicopter but it had the same issue in both of my multicopters.

Here is a link to the Skyview:

http://xaircraftamerica.com/products/ofm-skyview-pre-order

 

[R_Lefebvre]

What's going on with the Xaircraft and turning too fast? They must have a pretty weak AHRS system.

With Arducopter, I'm flying my helis at 130 km/h, pulling 6G acceleration, and up to 450 deg/sec flips and rolls, 3D flight, inverted, etc. No problem. And it's the same exact AHRS system the multirotors use. So you won't have any of those sorts of problems.

 

[Hellsbells]

Not sure what is going on with Xaircraft but I won't buy any of their products again. I ended up selling the Stella gimbal and sending back the OSD that I had purchased two weeks ago. I destroyed one Xaircraft Frame and and Aeroxcraft landing gear due to the SuperX failure. Xaircraft told me to return it to where I bought it which was CNC Helicopter but he refused(his customer service is horrible) to take it back so then Xaircraft said to send it China on my dime. I got tired of the run around and gave up. China customer service is all the same.

 

[tahoelight]

Not sure what is going on with Xaircraft but I won't buy any of their products again. I ended up selling the Stella gimbal and sending back the OSD that I had purchased two weeks ago. I destroyed one Xaircraft Frame and and Aeroxcraft landing gear due to the SuperX failure. Xaircraft told me to return it to where I bought it which was CNC Helicopter but he refused(his customer service is horrible) to take it back so then Xaircraft said to send it China on my dime. I got tired of the run around and gave up. China customer service is all the same.

I am dealing with this same issue at Xaircraft unfortunately. I refused to send my SuperX to China ($60) since I purchased it in the US, but they are not offering any other option for replacement. I have not been able to get it to connect to the computer in months. It still flies fine and actually one of my favorite controllers. (Yes, Hoverfly is still my most reliable flight controller w/o GPS).

I may be trying one of the Pixhawks soon so will chime in when I do, but haven't pulled the trigger. I'd love to here others experience, but sounds like a typical open source situation.

 

[Av8Chuck]

Did you get your SuperX from XAircraftAmerica.com? I know the margins on all f this stuff are thin but they should stand behind this and then the manufacturer should make their distributor whole after it gets sorted out. That certainly isn't unique to XAircraft, it would be nice if Chinese manufacturers and, in this case American resellers could work together better to resolve issues.

I guess that's the difference with professional products sold B2B than hobby grade sold B2C. Its too bad HoverFly couldn't make a go of it. It seemed like they were doing all the right things to help develop a professional AP market.

After all the crap from Open Pilot I have a hard time imagining an open source project making it in a commercial application.

 

[R_Lefebvre]

After all the crap from Open Pilot I have a hard time imagining an open source project making it in a commercial application.

Arducopter is being used in a lot of commercial applications. In fact, just last night I had a long meeting with a company who is currently flying a very expensive military grade quadcopter ($100,000 per system), but is considering using Arducopter. The reason is they can't get the features they need in a closed source system.

 

[Old Man]

Nice thing about an open source, if you're good with code (I'm not) you can look and make changes as you deem necessary. I work with a high end closed source AP, which in and of itself is not a bad thing because the source accepts responsibility for errors/deficiencies and makes fast corrections. Closed source AP's for multirotors are being provided by firms that appear to have no concern for those they are selling to so corrections, if and when they are released, are slow in dissemination and less than transparent in their intent. From my perspective, for our applications, an open sourced product is the only means we have of verifying that what we have been provided is appropriate for the intended/desired tasks. We only need more people with the correct skill sets performing product code review.

 

[Av8Chuck]

From all that I've read [which admittedly isn't that much] the Pixhawk seems like a good solution, I'm just not that technical so I tend to shy away from this sort of thing. As I mentioned I didn't have a great experience with Open Pilot. But I'm interested in giving it a go.

I don't suppose there's anyone flying one in the Thousand Oaks, CA area?