HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Aug 09, 2012, 03:11 PM
Registered User
United States, CA, Hayward
Joined Apr 2006
690 Posts
Quote:
Originally Posted by stanleyy View Post
guys, I don't want to disappoint you, but the heart of the whole mutlicopter flying magic is hidden in the proper configuration of control loops and gain settings
that is the case with almost every other system, but not with the ecilop (and probably the zenmuse). you can get stable footage even with a mistuned FC. if you understand what it does you know why. the ecilop uses the laws of physics to its advantage unlike the others which are fighting the laws of physics.

of course the better tuned you are the more enjoyable it will be to fly.
chris24g is offline Find More Posts by chris24g
Reply With Quote
Sign up now
to remove ads between posts
Old Aug 09, 2012, 03:59 PM
Registered User
Dorset. UK
Joined Mar 2010
107 Posts
Quote:
Originally Posted by stanleyy View Post
guys, I don't want to disappoint you, but the heart of the whole mutlicopter flying magic is hidden in the proper configuration of control loops and gain settings
And those settings are usually tuned for specific weight of the aircraft, gyro positions, servo types, battery weight.
You have to prepare yourself for a few (or many) iterations until the whole thing gets into working state.
I did that in the past on my older frames and it always took some time to get it right.
Anyway - camera gimbal settings should not cause any flips (unless it oscillates like crazy, but such misconfiguration you can easily solve before take-off).
If it really wants to flip, I would first check motor rotations and propeller types (CW/CCW), this is in most cases the reason of such behaviour.
Wow, how to win friends and influence people...............not.
Who mentioned camera gimball settings ?
This is far from my first quad and has already been flown with a basic kk fc without problems.
I bought the Naza after seeing Jelico's impressive videos, but have since learnt that the Naza does have some funny habits.
I'm just after info for using the Naza with this size and weight of quad.
revogeoff is online now Find More Posts by revogeoff
Reply With Quote
Old Aug 09, 2012, 05:03 PM
Registered User
Slovenská republika, Bratislavský kraj, Bratislava - Podunajské Biskupice
Joined Feb 2011
821 Posts
Hey revogeoff,
my apologies, I obviously misunderstood your post

I don't know what you mean by "funny habits" - I didn't observe any issues with the default values of Naza. If the device has a tendency to flip, then there must be something really wrong..

My friend has been flying his hexa frame controlled by Naza with many different battery setups and weights (starting from 2 kg up to 3.5 kg) with the same PID settings!
No flips or other disasters - there was maybe a little bit of wobbling but in any case he could always fly it nicely.

Naza is in my opinion very stable device, especially when I compare it to my previous arduino-based controllers, running mutliWii or Aeroquad...
stanleyy is offline Find More Posts by stanleyy
Reply With Quote
Old Aug 09, 2012, 05:30 PM
Registered User
Slovenská republika, Bratislavský kraj, Bratislava - Podunajské Biskupice
Joined Feb 2011
821 Posts
Quote:
Originally Posted by chris24g View Post
that is the case with almost every other system, but not with the ecilop (and probably the zenmuse). you can get stable footage even with a mistuned FC. if you understand what it does you know why. the ecilop uses the laws of physics to its advantage unlike the others which are fighting the laws of physics.

of course the better tuned you are the more enjoyable it will be to fly.
We can't compare Ecilop and Zenmuse - these are two completely different systems, although they can both give great results. The idea of self-balancing camera gimbal in Ecilop is great and I admire Aleksey's work - but still the system needs properly tuned gyro/servo setups, and roll/pitch signals from FC that can compensate the tilt of the aircraft during flight.
You can also use different springs, your servo arms can be different which means you need different gains to compensate the same movement of gimbal etc..

The self-stabilization can also be influenced by different aerodynamics of battery and camera, which can be a real problem in agressive flight or in strong wind.

So there are quite many factors that can and will influence the quality of video. It is not just about the frame... It needs a lot of tuning and time, I am sure Jelico spent a lot of time on his frame to get such nice results...
stanleyy is offline Find More Posts by stanleyy
Reply With Quote
Old Aug 09, 2012, 07:04 PM
Registered User
United States, CA, Hayward
Joined Apr 2006
690 Posts
Quote:
Originally Posted by stanleyy View Post
The self-stabilization can also be influenced by different aerodynamics of battery and camera, which can be a real problem in agressive flight or in strong wind.
wind is not as much of an issue with the ecilop as people assume.

gyro setup is minimal as well. they are just in rate mode. you are speaking theoretically and i am speaking from 1st hand experience.

basically your assumptions are way off base.

here is video i shot days after i got the ecilop in the air. there is some post stabilization but the video was very smooth out of the camera and my wookong was still in need of tuning

Patriots Jet Team (0 min 19 sec)
chris24g is offline Find More Posts by chris24g
Last edited by chris24g; Aug 09, 2012 at 07:10 PM.
Reply With Quote
Old Aug 10, 2012, 01:55 AM
Registered User
Slovenská republika, Bratislavský kraj, Bratislava - Podunajské Biskupice
Joined Feb 2011
821 Posts
Quote:
Originally Posted by chris24g View Post
wind is not as much of an issue with the ecilop as people assume.

gyro setup is minimal as well. they are just in rate mode. you are speaking theoretically and i am speaking from 1st hand experience.

basically your assumptions are way off base.
so you are trying to tell us that if the gyro is "just in rate mode", it is not important how you set it?
I am also tweaking my ecilop so I do have first hand experience as well.
Btw. nice video
stanleyy is offline Find More Posts by stanleyy
Reply With Quote
Old Aug 10, 2012, 02:00 AM
Registered User
Dorset. UK
Joined Mar 2010
107 Posts
Quote:
Originally Posted by stanleyy View Post
Hey revogeoff,
my apologies, I obviously misunderstood your post

I don't know what you mean by "funny habits" - I didn't observe any issues with the default values of Naza. If the device has a tendency to flip, then there must be something really wrong..

My friend has been flying his hexa frame controlled by Naza with many different battery setups and weights (starting from 2 kg up to 3.5 kg) with the same PID settings!
No flips or other disasters - there was maybe a little bit of wobbling but in any case he could always fly it nicely.

Naza is in my opinion very stable device, especially when I compare it to my previous arduino-based controllers, running mutliWii or Aeroquad...
Let me explain "funny habits" . The first time I test flew the Ecilop with Naza there was some slightly gusty wind but only up to about 10 mph. The Naza flew in this but was wandering around the sky, it was fighting the wind well though and not trying to flip. Second test flight, zero wind Naza ties to flip the quad 10 secs after take off. Tried again in manual mode and no flipping. There are lots of stories of inconsistancy with the Naza on various threads, May have to stick it on a basic quad for more relaxed testing !!
revogeoff is online now Find More Posts by revogeoff
Reply With Quote
Old Aug 10, 2012, 02:07 AM
Registered User
Slovenská republika, Bratislavský kraj, Bratislava - Podunajské Biskupice
Joined Feb 2011
821 Posts
Quote:
Originally Posted by revogeoff View Post
Let me explain "funny habits" . The first time I test flew the Ecilop with Naza there was some slightly gusty wind but only up to about 10 mph. The Naza flew in this but was wandering around the sky, it was fighting the wind well though and not trying to flip. Second test flight, zero wind Naza ties to flip the quad 10 secs after take off. Tried again in manual mode and no flipping. There are lots of stories of inconsistancy with the Naza on various threads, May have to stick it on a basic quad for more relaxed testing !!
uh, I never got such issues, maybe they have really some quality problems in the factory or something.. Or it is really some sofware problem.
Are you flying with or without GPS module? Because malfunctioning magneto sensor can screw up many things...
stanleyy is offline Find More Posts by stanleyy
Reply With Quote
Old Aug 10, 2012, 03:02 AM
Registered User
Australia, SA, Renmark West
Joined Sep 2011
399 Posts
Quote:
Originally Posted by revogeoff View Post
Would Jelico or anyone else running the Naza on their Ecilop care to share their Naza gain settings in atti mode. Mine wants to flip shortly after take off, not funny.
I don't think it's a gain issue but I don't think you are imagining the problem either. The NAZA is a great FC but there is one little trick to it that I've I've noticed. It obviously resets what "up" is in the first half minute or so after it's switched on and then must sample and average the gravity vector while it's in flight. That can lead to some behaviour that looks mysterious until you think about how it's setting the horizontal plane that it's trying to maintain with the motor deck.

If you start with it on rough ground, not level, on takeoff it will try to maintain the attitude it was in. If that attitude has a tilt on it, it will start to move sideways and if it's only coming off the ground slowly, it may catch its skid on the ground and even "trip" over it. I'd had this a few times on my lawn (not actually going over but scaring me into a rapid pull back on the throttle). In response, I'd started taking off by giving it a lot of throttle and "popping up" to a metre or so. It was when I saw the machine's subsequent behaviour in attitude mode - it went off sideways until I corrected it - that I realised what was happening. The answer is to make sure the aircraft is level when you turn it on and give it a good half minute to get its verticals set before takeoff.

The other occasion I've seen a problem was after 5 minutes hovering on GPS lock in a strong wind. The aircraft was naturally tilted into the wind and after a while the NAZA decided that must be level so a subsequent quick landing and takeoff gave the same sideways movement as it left the ground.

It's just something you need to be aware of.
RedSky is offline Find More Posts by RedSky
Reply With Quote
Old Aug 10, 2012, 03:37 AM
Registered User
Slovenská republika, Bratislavský kraj, Bratislava - Podunajské Biskupice
Joined Feb 2011
821 Posts
RedSky, what you just wrote is a serious information, thank you!
I think we should try to contact DJI and ask for clarifications...
I don't like the idea that Naza is doing accelerometer re-calibration during flight,
I can't imagine how they could do it correctly!
stanleyy is offline Find More Posts by stanleyy
Reply With Quote
Old Aug 10, 2012, 08:02 AM
Bert'sEyeViews of NYC
nycewheels's Avatar
United States, NY, New York
Joined May 2011
167 Posts
Quote:
Originally Posted by RedSky View Post
I don't think it's a gain issue but I don't think you are imagining the problem either. The NAZA is a great FC but there is one little trick to it that I've I've noticed. It obviously resets what "up" is in the first half minute or so after it's switched on and then must sample and average the gravity vector while it's in flight. That can lead to some behaviour that looks mysterious until you think about how it's setting the horizontal plane that it's trying to maintain with the motor deck.

If you start with it on rough ground, not level, on takeoff it will try to maintain the attitude it was in. If that attitude has a tilt on it, it will start to move sideways and if it's only coming off the ground slowly, it may catch its skid on the ground and even "trip" over it. I'd had this a few times on my lawn (not actually going over but scaring me into a rapid pull back on the throttle). In response, I'd started taking off by giving it a lot of throttle and "popping up" to a metre or so. It was when I saw the machine's subsequent behaviour in attitude mode - it went off sideways until I corrected it - that I realised what was happening. The answer is to make sure the aircraft is level when you turn it on and give it a good half minute to get its verticals set before takeoff.

The other occasion I've seen a problem was after 5 minutes hovering on GPS lock in a strong wind. The aircraft was naturally tilted into the wind and after a while the NAZA decided that must be level so a subsequent quick landing and takeoff gave the same sideways movement as it left the ground.

It's just something you need to be aware of.
Well said. I noticed very similar behavior on my Gaui-Ins, took about 1 minute to lock in Satellites and get the gyro stabilized. If you take off right after start-up, its pretty wobbly but gets better if you keep it steady for a bit. I think those algorithms try to adjust based on some average behavior?
Just uploaded another video from a few days ago. Ok but still can't get the roll right, it does not always go back to center. Is there any way to have the Wookong M roll function (I think F2) assigned to a channel for trim so I can manually center it?
Flying Ecilop over Hell Gate Bridge (4 min 3 sec)
nycewheels is offline Find More Posts by nycewheels
Last edited by nycewheels; Aug 10, 2012 at 08:08 AM.
Reply With Quote
Old Aug 11, 2012, 08:05 AM
Registered User
Joined Jan 2011
1,006 Posts
Well well well. My ecilop decided to get airborne for 10 minutes. Problem solved. It was a configuration issue. Now is flying.

I'm having a lot of jello, maybe because I filmed in 1/50 s speed. I will try in A mode (and fix it to something like 8).
Anyway it's working. Now need to make it film like the Thom one...
nicodh is offline Find More Posts by nicodh
RCG Plus Member
Old Aug 11, 2012, 08:14 AM
Registered User
Australia, SA, Renmark West
Joined Sep 2011
399 Posts
Quote:
Originally Posted by nycewheels View Post
Well said. I noticed very similar behavior on my Gaui-Ins, took about 1 minute to lock in Satellites and get the gyro stabilized. If you take off right after start-up, its pretty wobbly but gets better if you keep it steady for a bit. I think those algorithms try to adjust based on some average behavior?
Just uploaded another video from a few days ago. Ok but still can't get the roll right, it does not always go back to center. Is there any way to have the Wookong M roll function (I think F2) assigned to a channel for trim so I can manually center it?
https://www.youtube.com/watch?v=4wUZ87Q0Qfg
That's a really spectacular video - great shooting.

I too looked for a way of controlling roll on a WK-M stabilised gimbal. You can't feed any gimbal commands into the WK-M apart from pitch, all I came up with for roll was mixing a RC channel with the WK-M's roll output, using a cheap V-tail mixer. I still don't know if that would work as the V-tail mixer in my parts box appears to have failed.

What I ended up doing was buying a CRIUS-SE multiwii board and using that as the gimbal controller. That has the facility for inputting pitch and roll commands for the gimbal. It works well but I still have a problem in that my very classy Spartan Quark gyros have -50 exponential as default in rate mode. That's fine for what they're designed for - control of a helicopter tail but not what you want with a gimbal. The exponential is programmable but I'm waiting on a simplex adapter before I can change it. At the moment the gimbal (predictably, in view of the exponential) underreacts to small deflections giving it a rather nervous look in tutbulent conditions..
RedSky is offline Find More Posts by RedSky
Reply With Quote
Old Aug 11, 2012, 08:44 AM
Registered User
Australia, SA, Renmark West
Joined Sep 2011
399 Posts
Quote:
Originally Posted by stanleyy View Post
RedSky, what you just wrote is a serious information, thank you!
I think we should try to contact DJI and ask for clarifications...
I don't like the idea that Naza is doing accelerometer re-calibration during flight,
I can't imagine how they could do it correctly!
It's not a fault, IMO, it's just the way it works.

If you think about it, the accelerometers give 3 values, x,y and z. z is the axis vertical to the FC board and x and y are forwards/backwards and sideways.L/R

Now when the system is initialised, it sets x and y as zero, z will be a value probably rather less than half the maximum and is the gravity vector. If you now start flying around there will be varying values of each of x y and z as the aircraft banks tilts and accelerates. If the sticks are then centred the FC will endeavour to get x and y zero again. However, gyros drift so the FC must periodically reset the vectors. It's my guess (and I'm sure DJI won't be telling us exactly how they do it - it's commercially valuable information) that the NAZA integrates the values of x y and z over a period of a few minutes so that the average x and the average y equal zero.

Any FC that doesn't have any optical method of finding the horizon will have to do this. The period it integrates over will depend on how free of drift the gyros are, a longer period being OK for very stable gyros, which would explain why an expensive FC like a WK-M, with its temperature compensated gyros, is less prone to errors - it can use a long integrating period.
RedSky is offline Find More Posts by RedSky
Reply With Quote
Old Aug 11, 2012, 08:48 AM
Registered User
Joined Aug 2008
227 Posts
Quote:
Originally Posted by nycewheels View Post
Well said. I noticed very similar behavior on my Gaui-Ins, took about 1 minute to lock in Satellites and get the gyro stabilized. If you take off right after start-up, its pretty wobbly but gets better if you keep it steady for a bit. I think those algorithms try to adjust based on some average behavior?
Just uploaded another video from a few days ago. Ok but still can't get the roll right, it does not always go back to center. Is there any way to have the Wookong M roll function (I think F2) assigned to a channel for trim so I can manually center it?
https://www.youtube.com/watch?v=4wUZ87Q0Qfg
How do you fly FPV ? through the NEX5 ? or FPV fix camera ?

If you find a solution correcting the roll center at start, tell us
thomasamoth is offline Find More Posts by thomasamoth
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Five easy piece f-14 tomcat bobdaeronort Foamies (Scratchbuilt) 80 Oct 25, 2012 05:43 PM
Discussion Easy Glider VS Easy Glider Pro (what's different?) MkFn Electric Sailplanes 7 Nov 17, 2011 07:42 PM
Discussion Easi Glider film boca55 Serbia 0 Apr 20, 2008 04:46 PM
Discussion Spring Easi* in NZ harrys one Aerial Photography 8 Nov 14, 2007 01:22 PM