Espritmodel.com Telemetry Radio
Reply
Thread Tools
Old Sep 03, 2012, 12:12 AM
Vulcan skyhook multirotor
Andre6553's Avatar
South Africa, GP, Pretoria
Joined Feb 2011
665 Posts
Quote:
Originally Posted by maximal View Post
Andre, what's your point?

Why are you trolling other FC threads (at least the ones I'm reading) with the YS videos?

AQ position hold videos are more impressive anyway!

-Max
maybe i spend so much money the last year on other fc at last i got something that can work for me.

Believe AQ is just as good but i like the plug and fly option.
Andre6553 is offline Find More Posts by Andre6553
Reply With Quote
Sign up now
to remove ads between posts
Old Sep 03, 2012, 01:19 AM
Registered User
Joined Jul 2007
76 Posts
at least I like the YS-X6 heavy wind test. Have you tried the Wookong-M or the Naza befor you bought the YS-X6 Andre
roxyboby is offline Find More Posts by roxyboby
Reply With Quote
Old Sep 03, 2012, 02:23 AM
Vulcan skyhook multirotor
Andre6553's Avatar
South Africa, GP, Pretoria
Joined Feb 2011
665 Posts
Quote:
Originally Posted by roxyboby View Post
at least I like the YS-X6 heavy wind test. Have you tried the Wookong-M or the Naza befor you bought the YS-X6 Andre
No not have any dji systems
Ysx6 has more fetures thats why i go with them
Andre6553 is offline Find More Posts by Andre6553
Reply With Quote
Old Sep 03, 2012, 03:54 AM
Registered User
Joined Dec 2005
50 Posts
Quote:
Originally Posted by kinderkram View Post
What about a Hexa with only 5 motors?

http://www.youtube.com/watch?v=7-QGMlwbi2E
Uhhh, that's a dangerous video! Now people will think that an X6 type multicopter can fly with 5 motors. This is physically not possible except a few special cases! And this video is a special case: Absolutly overpowered with no wind and only small pitch and roll inputs. In realistic cases the hex would go crazy. The problem with hexcopters and 5 motors is that every pitch and roll command will result in a yaw movement and every yaw movement will result in a pitch and roll movement. That's physics and independant of the attitude controller! So with quads in "normal" configurations with e.g. a camera as payload the will get uncontrolable in at least one axis.

So why does DJI claim that their X6 flys with 5 motors? According to their manual they seem to rely on their care free mode. The hexcopter will spin around the yaw axis quickly and with the care free mode it might be possible to land it safely. I wouldn't call this "controlled attitude" because yaw is not controlled anymore. But I guess that's what you call good advertisement ;-)

Conclusion: If you want a hazzle free redundant system: Take an octo!

But besides this: The Autoquad seems to be a very good project! Good job!

Regards
Jast
jast is offline Find More Posts by jast
Reply With Quote
Old Sep 03, 2012, 04:00 AM
Registered User
hugy's Avatar
Joined Apr 2010
586 Posts
Jast, if have to disagree here with you.

I had a 5.5kg hexa copter with dji wookong lose a prop midflight. The copter was almost as stable as with all 6 props on. Only a slight drift to the side that was missing the prop was noticable. I could just steer it back towards me and have a safe landing. No spinning or somthing like that.
hugy is offline Find More Posts by hugy
Reply With Quote
Old Sep 03, 2012, 04:29 AM
Registered User
Joined Dec 2005
50 Posts
Quote:
Originally Posted by hugy View Post
Jast, if have to disagree here with you.

I had a 5.5kg hexa copter with dji wookong lose a prop midflight. The copter was almost as stable as with all 6 props on. Only a slight drift to the side that was missing the prop was noticable. I could just steer it back towards me and have a safe landing. No spinning or somthing like that.
Then you were very lucky! If you have some slight asymetric in your system it might work like this. But if you would loose a prop with the other turning direction it would be even worse! There is no question if this is the case or not. It's physics and if you take a sheet of empty paper and think about the moments you get in an emergency case with five rotors, you will understand that this is not a stable situation. That's what DJI describes also in their manual.
jast is offline Find More Posts by jast
Reply With Quote
Old Sep 03, 2012, 04:50 AM
pbw
arachnida
pbw's Avatar
Belgium, Flanders, Antwerp
Joined Aug 2008
214 Posts
OT: Are you the "German Asctec Jast"? Long time no see if you are!! The autoquad will probably be my next FC, and it all started with a Silverlit X-UFO and a Funpilot that I never sold because of nostalgia.
pbw is online now Find More Posts by pbw
Reply With Quote
Old Sep 03, 2012, 05:37 AM
OS Sponsor
kinderkram's Avatar
Germany, NRW, Dusseldorf
Joined Jun 2010
5,199 Posts
Quote:
Originally Posted by jast View Post
Then you were very lucky! If you have some slight asymetric in your system it might work like this. But if you would loose a prop with the other turning direction it would be even worse! There is no question if this is the case or not. It's physics and if you take a sheet of empty paper and think about the moments you get in an emergency case with five rotors, you will understand that this is not a stable situation. That's what DJI describes also in their manual.
It was never meant to be for flying around a Hexa with only 5 props.
But you have a much better chance to get it down to the ground without crashing it in case of an emergency.
For a more realistic test you need to be able to switch off a motor in flight.

We still haven't tested L1ac in depth - what you have seen so far is just a proof of concept. It's still at a very early stage and will not be part of the beta test.
kinderkram is offline Find More Posts by kinderkram
RCG Plus Member
Reply With Quote
Old Sep 03, 2012, 05:59 AM
Friend is a 4 letter word...
JussiH's Avatar
3rd stone from the sun
Joined May 2007
2,670 Posts
The controller in Menno´s vids is not running the stock PID controller, but instead it uses an adaptive L1 controller, which requires complete mathematical profiling of craft geometry/mass, motors, props and power distribution. Basically its a "selftuning" controller that can adapt almost instantaneously to almost any changes in flight conditions. It is what many would consider a "holy grail" of multirotor control

But there is a catch: The adaptive controller is not (and will never be) part of the public open source branch, since parts of it is based on closed source technology that was licensed from other parties. So if (when) its made available to the public, it will involve a license fee. And its still in an early stage of development and testing.
JussiH is offline Find More Posts by JussiH
RCG Plus Member
Old Sep 03, 2012, 07:47 AM
Registered User
Joined Dec 2005
50 Posts
That sounds interesting! We did implement MRAC-Controllers and were able to estimate output-mapping and controller gains within a few 100ms during flight. However we didn't see many advantages in the adaptive controllers. There are only a few things which can happen (like a motor problem). They can be detected (unless you have PPM controllers) and then there is a perfect static solution to still be able to fly. I'd rather use the remaining calculation time for better data fusion or other number crunching tasks ;-) On the other side we just need to adapt our controllers to 5-6 fixed systems :-)
jast is offline Find More Posts by jast
Reply With Quote
Old Sep 03, 2012, 08:54 AM
Registered User
hugy's Avatar
Joined Apr 2010
586 Posts
how long should the step 3 cal mag be running. Mine is running for about 3 hours right now. I don't see the mae dropping anymore but it still won't give the define values.
hugy is offline Find More Posts by hugy
Reply With Quote
Old Sep 03, 2012, 09:08 AM
Registered User
Joined Jan 2012
395 Posts
Question about Vector nav

I noticed in the specs one of the possibilities is the VectorNav 100. How is that incorporated into the working of the board? Thanks rkenosha
rkenosha is offline Find More Posts by rkenosha
Reply With Quote
Old Sep 03, 2012, 09:15 AM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,368 Posts
Quote:
Originally Posted by hugy View Post
how long should the step 3 cal mag be running. Mine is running for about 3 hours right now. I don't see the mae dropping anymore but it still won't give the define values.
Each loop of 'sim3 --mag --incl' should produce a bunch of defines like this

Code:
0.025442 0.000000 0.031998 0.001304 0.003797  = 0.062540
[HLoops:  1654

#define IMU_MAG_BIAS_X		-0.038319396049	//  0.000000983824 +0.000000157994
#define IMU_MAG_BIAS_Y		-0.030053607254	//  0.000000999264 -0.000000030221
#define IMU_MAG_BIAS_Z		+0.056593579621	//  0.000000919453 -0.000000107440
#define IMU_MAG_SCAL_X		+0.211107452186	//  0.000000969515 +0.000000011254
#define IMU_MAG_SCAL_Y		+0.216149258375	//  0.000000925903 +0.000000015821
#define IMU_MAG_SCAL_Z		+0.200516989523	//  0.000000961659 -0.000000046187
#define IMU_MAG_ALGN_XY		-0.013353562863	//  0.000099559034 -0.000000511795
#define IMU_MAG_ALGN_XZ		-0.022929032868	//  0.000099837836 -0.000000011327
#define IMU_MAG_ALGN_YX		-0.040559377560	//  0.000099012393 +0.000000016370
#define IMU_MAG_ALGN_YZ		-0.004882450596	//  0.000099939887 -0.000000127055
#define IMU_MAG_ALGN_ZX		-0.062296579265	//  0.000099838568 -0.000000103042
#define IMU_MAG_ALGN_ZY		+0.006072738691	//  0.000099978917 +0.000000114612
#define IMU_MAG_INCL		-67.229172176667	//  0.000000095673 +0.000001522129
Did you seed the param file with your local IMU_MAG_INCL before running this step of the calcs?
GoFaster is offline Find More Posts by GoFaster
Last edited by GoFaster; Sep 03, 2012 at 09:26 AM. Reason: mixed up cal and sim3
Reply With Quote
Old Sep 03, 2012, 09:20 AM
Registered User
hugy's Avatar
Joined Apr 2010
586 Posts
Sorry I'm talking about mag cal step 3

This is the current output, still running.

Quote:
Mag Loops: 6327000
Mag 0th order Variance:
+0.000000000 +0.000000000 +0.000000000 +0.000000000011
+0.000000000 +0.000000000 +0.000000000 +0.000000000010
+0.000000000 +0.000000000 +0.000000000 +0.000000000016

Mag 0th order Params:
+0.201171073 -0.001614243 +0.035154598 -0.004724476879
-0.054425402 +0.212606214 -0.070884724 +0.003878642634
-0.065971517 +0.021485315 +0.202039551 -0.015272749344


Mag 1st order Variance:
BIAS SCALE
+0.000000000 +0.000000000
+0.000000000 +0.000000000
+0.000000001 +0.000000000

Mag 1st order Params:
BIAS SCALE
+0.000055074 -0.000699286
+0.000070687 -0.000876613
+0.000268699 -0.000903047


Mag 2nd order Variance:
BIAS SCALE
+0.000000000 +0.000000000
+0.000000000 +0.000000000
+0.000000000 +0.000000000

Mag 2nd order Params:
BIAS SCALE
-0.000000405 +0.000000154
-0.000000266 -0.000000027
+0.000001600 -0.000001164


Mag 3rd order Variance:
BIAS SCALE
+0.000000000 +0.000000000
+0.000000000 +0.000000000
+0.000000000 +0.000000000

Mag 3rd order Params:
BIAS SCALE
-0.000000006 +0.000000029
-0.000000042 +0.000000081
-0.000000154 +0.000000233


Mag Inclination Variance:
+0.000000092

Mag Inclination Parameter:
-67.129997253


Static value Variance # 1:
+0.000000000 +0.000000000 +0.000000000

Static value Params # 1:
+1.023851836 +0.533637760 +0.799673354

Static value Variance # 2:
+0.000000000 +0.000000000 +0.000000000

Static value Params # 2:
-1.126655012 -1.010447570 +0.495051731

Static value Variance # 3:
+0.000000000 +0.000000000 +0.000000000

Static value Params # 3:
-0.409670215 +0.142293587 +1.823041931

MAE: 0.103502
AVG: 1.98423
lowMAE = 0.063042, n[0] = 0.000033
hugy is offline Find More Posts by hugy
Reply With Quote
Old Sep 03, 2012, 09:20 AM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,368 Posts
Quote:
Originally Posted by rkenosha View Post
I noticed in the specs one of the possibilities is the VectorNav 100. How is that incorporated into the working of the board? Thanks rkenosha
My understanding is one of the features of the VN100 is that it is pre-calibrated, which means that one would not have to run through all/most of the current calibration and calculation process. I say 'most' because I think you would still have to account for the hard and soft iron effects of the particular aircraft config.
Maybe you would only have to perform the dynamic calibration.

Perhaps someone who knows for sure can comment.
GoFaster is offline Find More Posts by GoFaster
Last edited by GoFaster; Sep 04, 2012 at 09:02 AM.
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
New Product Autoquad team proudly presents: Autoquad ESC32 JussiH Multirotor Electronics 272 Jan 01, 2014 01:32 PM
Discussion Naza dji sh gps - south hemisphere public beta testers results! Velociraptor73 Multirotor Talk 26 Nov 20, 2012 01:45 AM
Discussion How Bloomberg's Soft Drink Ban Will Backfire on NYC Public Health MtnGoat Life, The Universe, and Politics 1 Jun 26, 2012 11:19 PM
Discussion The AutoQuad! [ video ] kapteinkuk Multirotor Talk 5 Jun 10, 2012 11:19 AM
Download G-3 Public Beta (.426) is out. Includes Airplane Importer! L0stS0ul Simulators 37 Mar 03, 2006 05:28 PM