DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Why is the COG (center of gravity) way off on the Mavic?!

manu.weyers

Member
Joined
Jul 1, 2017
Messages
7
Reactions
0
Age
58
If you lift the Mavic with the two diagonal opposite motors (on the propellers hub), the Mavic will lean heavily forward ("nose heavy").

This "mechanical" design imbalance is strange ... , and as a consequence, provokes a problem that can be seen when you decent the Mavic vertically down at max speed (full throttle low), ... the controller/motors can't keep the Mavic level and a noisy wobbling appears very clearly!

The only reason i see they made this compromise, is the folding design restrictions and the free field of view for the camera (no propellers)

Sad imo...




correct_COG.jpg
 
Wobbling in the downwash has nothing to do with the CG, which is largely irrelevant anyway. Lower your gains a bit and it will stop doing it. They're a bit high out of the box IMO.
 
Consider this:
Yawning on a quadcopter (4 motors) is done on increasing 2 diagonal motors rpm and decreasing the 2 other ones.
In difficult situations as rapid descents, to keep a horizontal level, it has to compensate for the cog that is way off + the yaw-heading + descent=low rpm of the motors.
These 3 parameters can counteract each other, to come to a situation that it is technically not able anymore to keep the copter level!

youtu.be/SXr6ohqgnPA?t=166
 
Last edited:
I'd imagine to make forward flight and acceleration the path/direction of least resistance (outside of the motors off, falling straight down scenario)
 
If you lift the Mavic with the two diagonal opposite motors (on the propellers hub), the Mavic will lean heavily forward ("nose heavy").

This "mechanical" design imbalance is strange ... , and as a consequence, provokes a problem that can be seen when you decent the Mavic vertically down at max speed (full throttle low), ... the controller/motors can't keep the Mavic level and a noisy wobbling appears very clearly!

The only reason i see they made this compromise, is the folding design restrictions and the free field of view for the camera (no propellers)

Sad imo...




correct_COG.jpg


I did this: hang by the rear rotor hubs, one at a time. The vertical line (hanging string) indicates the CoG. Found the mass CoG to be within a few mm of the geometric CoG - at least in the plan view.

The vertical CoG (IIRC) was much harder to determine but seemed close to the middle of the battery. (I'd have to dig up the post but I'm feeling quite lazy at the moment).

What is more surprising, however, is because of the folding requirement, the rear rotor disk and front rotor disk are not at all parallel and this results in a massive difference in power loading fore and aft. This is probably the largest factor limiting max speed other than the very high "top load" on the body during high speed cruise flight (due to nose down attitude).

The vertical descent wobbling you are seeing is possibly due to ring-vortex-state with each rotor losing lift briefly until the attitude changes resulting in corrective RPM changes that briefly break the RVS. This goes on and on probably cycling from one rotor to the next. And I've seen it in the P4P as well during "aggressive" vertical descents.

If you must descend like that, I've found that hard "rudder" (yaw) at the same time gives you a nice, fast and stable corkscrew descent. (I've done this more with the P4P than the MP, however).

Edit: clarified locating the CoG.
 
Last edited:
If you lift the Mavic with the two diagonal opposite motors (on the propellers hub), the Mavic will lean heavily forward ("nose heavy").

This "mechanical" design imbalance is strange ... , and as a consequence, provokes a problem that can be seen when you decent the Mavic vertically down at max speed (full throttle low), ... the controller/motors can't keep the Mavic level and a noisy wobbling appears very clearly!

The only reason i see they made this compromise, is the folding design restrictions and the free field of view for the camera (no propellers)

Sad imo...




correct_COG.jpg
Whilst I like the techy questions and discussions I need to know what your point is!
Are you asking a question or just making an observational point looking for reaction ? ;-)
 
IMG_2722.JPG IMG_2723.JPG Yes, I saw this "angeled" rotor disks too... imo it is by design...

Would be nice to find that post where you were talking about the cog/co-mass/co-trust!

Side & front view of rotor disks

I am talking about "center of diagonal motor trust" is way off "cog" ...
Symmetry is important on quadcopters so that in "in not so normal" conditions, the motors (trust) can handle the flight controller orders!!!
 
Whilst I like the techy questions and discussions I need to know what your point is!
Are you asking a question or just making an observational point looking for reaction ? ;-)

It is a question, Why this assemetry?
 
This was hashed to death many months ago in another thread.

This photo shows you the disks relative to each other, a whopping 10.2° difference fore and aft resulting in an even whoppier 61% difference in thrust between the fore and after rotors in a stable hover.

upload_2017-8-8_14-25-58.png
 
  • Like
Reactions: Gotcha
Most likely because it allowed for the other design aspects / features you pointed out in your OP (folding/camera clearance) without having a significant impact on the flight characteristics (or one that is compensated by other design decisions as pointed out above).
The Mavic by default descends at a max of 3m/s. Some have pushed it to 10m/s and it still can fly itself. You're not in any way getting in a scenario such as

In difficult situations as rapid descents, to keep a horizontal level, it has to compensate for the cog that is way off + the yaw-heading + descent=low rpm of the motors.
These 3 parameters can counteract each other, to come to a situation that it is technically not able anymore to keep the copter level!

Really there is nothing wrong at all with the flight characteristics of the Mavic frame (even beyond design parameters), just maybe gains a bit too high (I use 90% on pitch, 80% on roll).

DJI are known not to fine tune their gains for smooth moves, because... well, their gimbal systems are so good they can afford pretty average airframe moves, they'll be compensated and not appear in the "product" (imagery) which is what matters. As a result they tend to rather tune for efficient trajectory control even if that's a bit jerky to avod noobs crashing into things because response was too slow to compensate for their mistakes.
 
Last edited:
  • Like
Reactions: Lars Olsson
In my experience is having a "mechanical stable design" + "well written software driven stabilization" winning pair...

What i am trying to say here, is that i see 4 identical motors, and seeing all sorts of asymmetric design points, there must be situations that a combination of factors will led to a motor that will not be able to respond (under-powered) to the controller set-point, ...

MP is mainly designed to make smooth forward translations? Is that the reason for the asymmetry of the angled propellers and the off-COtrust?
 
In my experience is having a "mechanical stable design" + "well written software driven stabilization" winning pair...

What i am trying to say here, is that i see 4 identical motors, and seeing all sorts of asymmetric design points, there must be situations that a combination of factors will led to a motor that will not be able to respond (under-powered) to the controller set-point, ...

MP is mainly designed to make smooth forward translations? Is that the reason for the asymmetry of the angled propellers and the off-COtrust?

The design requirement was probably written as follows:

"Build us a 1) very compactly folding drone, that 2) retails for $999 in the US market, which 3) carries a 4K camera and will 4) kill* the Karma in the marketplace."​

The asymmetry is a consequence of the folding design.

You are correct that the fore/aft asymmetry limits the drone in at least one fashion: because this asymmetry results in most forward flight thrust coming from the rear rotors, then the maximum speed is limited by that as those rotors will hit max RPM before the front rotors. (Explains occasional overspeed RPM warnings that I have seen on occasion). I assume this also defines the maximum pitch angle.

Endurance and speed are also limited by the horrendous drag resulting from the nose down attitude causing air to hit the large upper fuselage area.

As to stability I haven't see a sign of any instability in any conditions including high wind and turbulence except the vertical descent oscillation you mentioned and which I address in my first reply above.

*While it could be argued that the Karma accomplished its death all by itself, the MP was going to do it serious damage. This also assumes that DJI had wind of GoPro's project.
 
Last edited:
This was hashed to death many months ago in another thread.

This photo shows you the disks relative to each other, a whopping 10.2° difference fore and aft resulting in an even whoppier 61% difference in thrust between the fore and after rotors in a stable hover.

View attachment 19321

If you could find that post again I will be happy to read it!
 
It is a question, Why this assemetry?
Cheers, must say I like the visual balance of the mavic with the rotors, don't know why, just do, it seems purposeful!

I would like to think the designers must have some awareness/knowledge of aerodynamics and how it all works and isn't just because it folds up better that way!
 
The wobbling while decending fast is known as "Vortex". The props are going through their own downwash as it decends. All quads/drones will do this decending vertically straight downward fast. The "simple solution" is to fly forward while decending,making the props clear their own downwash.Racing drones are really bad about this,as there is no stabilization in acro. Just fly forward while decending( it dosen't take a lot of forward stick),and see how smoothly it will come down. If you must decend straight down,do it slowly.
 
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
132,257
Messages
1,571,444
Members
160,996
Latest member
sklumph