Hi All

As an avid multi-copter enthusiast I need to ask one simple question .......

Why is it that the DJI NAZA and WOOKONG systems work 99% tuned 'straight out of the box' for most combinations of motors, ESC's, frames and combinations - but the APM project needs sooooo much tuning to achieve a similar result?

Am I being very naive here, or just plain practical??


Views: 32190

Reply to This

Replies to This Discussion

Simply put naive. What do you want to do? How much time and money do you want to spend? AR Drones are great out of the box and cheap, yet they are limited. And I destroyed 2 learning how to fly and learning their limitations.

Hi Dale, you can't compare APM with the DJI products, two different "development philosophy".
DJI come from a military technology, have brought their expertise in the field of "consumer" minimizing parameterization, APM by a team of enthusiasts.
APM offers many advantages, the user savvy enough can configure your setup and under certain conditions it fly better than a product DJI, its Achilles heel is the speed of the processor (16 MHz only), with this speed is unthinkable to obtain a stability comparable to pure flight board that install processors 10 / 50 times more powerful.
But the advances in the V3 release that will come out shortly lead APM Copter able to compete with expensive commercial products, it lacks an easy interface in the Planner and the work would be complete imho.
With PX4 and VR Brain (the new powerful board) you can get the same performance of the DJI products, with a good tuning I have shown in my video that this is possible.

You are right about Naza and Wookong  systems working great without need for tuning. They are impressive in position hold and altitude hold, and fly real nice.


But with latest versions of APM Copter firmware it is not necessary to do much tuning if you only want a good flyer. For my preferences I always lower stab P (3.8) and rate P (0,13) a little bit before first flight and it works without further tuning.


I did put my Phantom with Naza in position hold side by side with my TBS Discovery with Arducopter 3.0.0-rc5. I think they behave very similar. Arducopter parameters are default except lowered stab P and rate P. And Phantom has pretuned parameters from factory special for that model.

You can still see Phantom is more stable at position hold. But they are very similar, and APM may improve with tuning of PIDs. Both are great Flight Controllers. Which to choose depend on what you want to do with your multicopter! 

yes but apm is not reliable like dji..there are bugs  ;)

APM 3.0 software is coming along nicely and will improve things greatly. But the commercial counterparts still has some tricks up their sleeves when it comes to trouble free tuning (or no tuning at all in many cases). I think much of it comes from having faster and more dedicated hardware for the job instead of basing it on the Arduino architecture. But APM will catch up on that in the long run with PX4/APM3 hardware.

Here are some examples from my favorite (commercial) autopilot, the ZeroUAV YS-X6. First video is a test flight on a stock DJI 550 hexacopter (8" props) with no tuning what so ever (using factory defaults). The next video is the same YS-X6 moved to a big CF 1000mm octocopter (15" props) using the exact same factory default settings. DJI WKM behaves pretty much the same on the F550, but needs a little bit of tuning to fly the big octos.



Are you sure it is less reliable ? Less reliable in witch conditions ?

I did flew version 3.0 release candidates during 20 hours without any crash on a hexa copter, testing all modes including acro with 90° banking, auto mission and auto landing, at speed as high as 24 m/s, ascension speed as high as 10 m/s, descend speed 16 m/s. This is quite good results for me.

I have another Hexa machine with a camera, 5.3 kg weight, flying arducopter since version 2.0 with an APM 1.4, never crashed. I did manufacture a big Octo, 9 Kg, APM 2.5, no crash. Only changed two PID parameters from defaults. I don't see the point with the tuning complexity myth. More a machine can be tuned very fastly in flight with a 3DR telemetry radio. This is for custom machines. For stock machines, default parameters are just working.

It's open source, there is an issue tracker, a Git repository, hardware price is low, you can contribute and help to find and solve bugs, or help to complete and rework Wiki pages for version 3.0 if you are comfortable with wikis.

As a human i like bugs and i have bugs myself. I like to find them and find solution to kill them, and discuss about that with the Ardupilot community. I did spent quite a lot of time testing v3.0, watching closely for safety issues and hardware problems detection, discussing and writing patches to solve a couple of them. I would not be able to do that with commercial projects and for me Arducopter is great for most tasks.

The Dev Team did work hard and the result will be for sure a 3.0 release with safety as a primary concern and a new first-class nav controller. The enhancements list is long for this version.

A non negligible amount of crashes are caused by hardware issues or user omissions. V3.0 will lock a percentage of those, forbidding arming or switch to auto mode. This will certainly rise the perceived Arducopter reliability.

Version 3.0 rc6 will certainly enter public availability through Mission Planner in a couple days for more generalized testing so that some small adjustments can be done before the final release for a smooth transition. When public 3.0 rc6 will be out, we'll certainly have new interesting DJI / Arducopter comparisons here :)

Well this is just an anecdotal story but recently while doing tests with my HAL Hexa one of the aluminum motor mounts broke. I was able to fly it back to me on 5 engines from about 50 feet up and 150 feet away with enough control to land it up right and not brake anything not even a prop. This was with a the Arducopter APM 2.5. 

My friend who was standing right next to me at the the time, also a hexa pilot using NAZA M was flabbergasted I was able to fly it back,. He said on two occasions he had broken a prop in mid air with his Tarot 680 hexa and both times it crashed almost immediately. 

Of course these are two different rigs with two different pilots and many variables. 

Hi Marco

Thanks for your original advice to post on diydrones.com.

I've learned more practical advice and answers to some seemingly obvious questions on this thread than I have trawling through the many forums of seemingly self-appointed experts for many months now.

Thanks everyone for all for your replies to my question(s) - keeping it simple and factual is a bonus - I'm sure it's also helping others to learn and understand some of the fundamental issues and hurdles faced by the APM project vs the commercial alternatives!

Keep up the postings everyone - knowledge from this thread is proving informative ...... and useful! I find broadcasting excessive knowledge of a subject can sometimes get in the way of what most of us really want to achieve - which is simply 'good reliable flying time'!


Hi John, thanks for the reply

I have an early ZeroUAV YS-X6 (test flown only) - but I can't seen to learn to love it. I guess it doesn't help that my early one needs the router attached to do lots of things (inc Android communication) that a simple USB connection offers with other systems including my APM, NAZA and Wookong-M. It's been for sale on Ebay twice now and at £395 including a new TP-Link router, nobody else seems to love it enough to buy it either?

@titeuf007: mmmm where? :-)
In any system there are bugs, don't think that the DJI products are exempted.
The advantage of APM Copter is in being able to find them through all the logs, impossible with DJI Naza or WKM.
In case of crash is always possible to know whether the problem is, hardware or software, in the second case the bug is found and then fixed.
I "try" to coordinate the group of testers of ArduCopter about a year and I assure you that the problems are greatly decreased in this year, almost non-existent, some bugs on automatic were virtually all found and resolved, the user that can fly in Stabilize always brings home his quad whole.
When I
joined the ArduCopter group remember that every day I was wondering if I could bring my quad intact at home, now I wonder if it crash for software bug.

My last crash was last year, for a big bug found and then fixed after this attempt.
The flights that I have with ArduCopter assets from that event are over 3000, without a single crash or fly-away.
I think "APM Copter V3" is a really milestone release, another world respect the previous version, surely there will still be some thing to fix but is unlikely to happen that someone in an emergency is not able to land the quad own safe.

it wasn t a critical, i love my apm and i am very impressive for the price.


all normal users like me see  you hard work and we can only say you a big THANKS for this great project

Reply to Discussion



Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2018   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service