Dear Friends,
after a lot of work on Electronic and Frame we are happy to present first flight test of VRCopter complete of VR 2 Axis Gimbal .
In this fly we test this Copter configuration :
- VR Copter Frame .
- VR oring anti vibration system.
- VR Copter Landing Gear
- VR Gimbal Frame and Electronic for GoPro Hero 3.
- GoPro 3 Hero Silver
- VR Brain 4.5
- 4 MK3639 motor
- 4 ESC with SimonK Firmware
- 3DR GPS
- 3DR Telemetry KIT
- 4 APC 12 Propeller.
- 10 Amp 3S battery flight time around 20 min without payload until 24 min.
In this test flight I checked the behavior of the drone in different situations both at high altitude I flew about 50 m high and low altitude.
The video that is present in raw format created with the GOPRO 3 silver model you do not notice the presence of high-frequency vibrations, which were removed by introducing a special anti vibrating plate inserted between the frame of the drone and the gimbal . There are only some low-frequency vibrations due to the influence of the wind at high altitude on the mechanics of the gopro gimbal frame we're still working to improve it further.
Overall, we think that this platform can be an interesting alternative to cars like the Phantom of the DJI is that in the previous version thanks to a range of great flight almost 20 minutes.
This drone configuration will be available in DIY KIT before the end of year in the next week will be available more info about it
original blog post : http://www.virtualrobotix.com/profiles/blogs/vr-copterx4-with-vr-2-axis-brushless-gimbal-a-good-entry-level-pl
more info about status of VR Gimbal project : http://www.virtualrobotix.com/group/vr-gimbal-user-group?xg_source=activity
best
Roberto
Comments
Edit: Even poor baseflight has a better "rotateV" than orig multiwii here: http://code.google.com/p/afrodevices/source/browse/trunk/baseflight.... And I wonder why not even that was used for the 32 Bit port....
@R_Lefebvre. The MPU DMP stuff is just a smokebomb in that code. I think it was once used but was even worse.
They use the multiwii imu for that like shown here: http://code.google.com/p/multiwii/source/browse/trunk/MultiWii_shar...
You can comare it with the actual gimbalcode here: http://code.google.com/p/vrgimbal/source/browse/Firmware/VRGimbal/I...
And thats where this 32Bit hardware comes at play. On this gimbal here - there is much more possible than that. I guess the 32Bit gimbal cpu is not used beyond 10%. So that current code is ok to show that it is working and that is a good thing but customers could expect much more, where an alexmos or martinez board cpu tilts out long before. From a recent discussion on other 72Mhz stm based controllers (openpilot etc) there seems to be a problem with the 6th harmonic farting around in the 433Mhz band (telemetry / Uhf etc). You can set the clock multiplier for the cpu in software. So it may be reasonable to alter it +or- 8Mhz (over/under clock it) to keep that freq. band clear. (Reference for my proposal is here: http://fpvlab.com/forums/showthread.php?16153-realFPV-400Q-Crashpro....) Just an idea - it may be different with this stm gimbal - i don't know.
Cheers Kraut Rob
Ok, those numbers sound pretty good then. Which motors and props?
My H8 is a bit heavy at 1500g for just the frame, but like most things I do it's overbuilt. I can stand on it without a problem. :D
Hi R_Lefebvre
so I fly VRCopter in different configuration.
The basic version for FPV is around 1400 gr without battery.
If you add 10 Amps 3S Battery you are around 2000 gr
if you add a GoPro 2180 gr in that configuration i doing a fly test and i fly for 24 minutes ... with a 4S battery I fly for 28 min.
So the maximum AWG is around 2400 gr with x4 configuration . I don't yet check the x8 . In my previous reply i told 1 kg of payload available .
Best
Roberto
So they aren't using DCM? What is the AHRS system?
From what I have observed, the AlexMos AHRS is exceedingly crappy. I've tried to talk to them about it, but it's all smoke and mirrors there.
I'd definitely like to try and help make one of these systems better.
Hi Roberto,
When will your online store be available?
Yes, R_Lefebvre that can be done and that is a working concept, like done in Multiwii. But when you decide to implement the timefactor it should be done correctly in my opinion. With my limited experience with control loops I can say that having some smoothing on the delicate D part, like advised in the "wiki" and like done in arducopter and other projects helped me in my own project. The next thing that should be taken into consideration is the sensitivity of the accelerometer. The "imu" will not take acc values beyond 1,5G (or something like that) and that is ok to get rid of other forces the easy way. On the other hand I think it is not a good idea to set the sensitivity to 2G on the acc, because it will see saturation by vibration etc too soon. Anyway, someone should implement a proper earthframe calculation like done by William Premerlani and some "known to work" control loops. I wonder why all these libraries are in the trunc and not actually used.
I think the VR Brushless Gimbal is an outstanding piece of openhardware with great potential and I would definitely go for it, if I had a copter strong enough to pull off some decent camera :) :). Cheers Kraut Rob
Roberto, are you saying the empty weight of the quad is 1kg? Or what? Surely the All-Up-Weight can't be 1Kg, with the motors, batteries, electronics, etc?
Crash, yes, I see it now.
I agree that technically, the D-term segment should be dividing by Dt rather than multiplying. However, since the time segment is a constant, DT_INT_MS, and since we tune heuristically, it sort of becomes irrelevant. There could be no Dt in the calculation at all, and as long as the processor time is constant, it would still work exactly the same except you would end up with a different heuristically determined D-term. ie:
Let say Dt is 0.01 seconds. And lets say the system is stable with a D-term of 0.1. Well, D-term*Dt = 0.1*0.01=0.001 .
Now, if we fix the D-term calculation, then you would find the heuristically determined D-term would be 0.00001. Because D-term/Dt = 0.00001/0.01 = 0.01. So the calculation would wash out the same.
As I said, you could get rid of the Dt entirely, and just set D-term to 0.01, and have the exact same result.
Again, this is assuming a stable cycle time (and we must assume so, because it appears the Dt being used is a constant), and assuming heuristic tuning where we simply adjust D-term until the gimbal works.
@ Hugues
So the AWG is around 1 kg the drone is compact but is developed using excelent material : Ergal and Carbon Fibler plate.
In X8 configuration we use 8 motors with 12'' propeller up and down the main difference in configuration is that the ratio speed of propeller above is between 7-10 % more of motor on top.
Best
Roberto
-
1
-
2
-
3
-
4
of 4 Next