Hi All, Just pulled this image in from SUAS News (Thank you Gary) to illustrate a few really good concepts that it incorporates and talk about better Multicopter design a bit.
This copter has a lot of things right and is really a study in design excellence.
1. Most obvious is a sliding ball full enclosure camera Gimbal, very pro and really something we should be striving for.
2. The motors are on the bottom under the arms: More aerodynamically efficient with no prop wash interference and very little intake interference. Noticeably increased flight times and greater lift capability and much quieter.
3. Large slow high efficiency props and pancake motors: These are way more efficient than faster motors with smaller diameter more conventional props.
4. Carbon fiber aerodynamic frame arms: Again more aerodynamically efficient, light and strong.
5. Fully enclosed framework and a simple spring loaded landing gear that provides as little interference with the prop wash as possible.
Basically this quad incorporates the best design features I have seen so far and definitely provides food for thought.
I thought it might be worth soliciting comments and other thoughts about optimal multicopter design techniques.
Best Regards,
Gary
Comments
Virtual robotics have a great example of aerodynamics and design
http://www.virtualrobotix.it/index.php/en/prod/bod
I respect everybody scientific results,my experiance is only from flying with props top,and down ,and top-down together and measuring time I was flying...but I must tell you I cant imagine situation where you have 20% lost just bcs. of props above arms or maybe I understood wrong bcs I do agree with 1-2%..
10-20% would be all lost you have bcs. of coaxial setup...but that is completely different story..
As I sad before and someone else too,all this discussion is not worth of calculate.
If you are worried about aerodynamics you better try to find way to make big central part and camera with gimbal more aerodynamically efficient
Arnie,
As you say, it varies greatly depending on exactly what your setup is. On an aerodynamically efficient setup the power consumption for equivalent thrust would vary by a small amount, perhaps 1-2%. An inefficient setup would yield much larger variations, I would guess the order of 10-20%.
Hi Chris,
I understand that the answer depends entirely on the setup (frame, motor, prop sizes and distances) but can you give an estimate of how much more power is consumed with props above frame compared to the same setup with props below the frame? Just to have an idea of how much loss of power we are talking about?
Thanks,
Arnie
The air above the prop is static at a distance, but the frames are often built very close to the propellor disk meaning that the inflow speed at the frame is almost that at the propeller disk. So the effect on power consumption and thrust is still noticeable under hover conditions, more so than one might think intuitively. I have results of static tests somewhere, I'll try and dig them up...
Either way, the conclusion is that a bottom mount is more efficient in terms of its aerodynamics.
Hi Chris,
That is entirely consistent with data posted by others as well.
The situation for a multicopter is that it normally lives in a state of hover.
So the area above the prop is essentially a static air mass whereas the one below it is very much a moving dynamic air mass.
Essentially the interference with the intake air is considerably less lossy than the interference with the dynamic output air.
This is true for hovering vehicles.
But a dynamically moving airplane generally at least reduces the advantage because the intake air is now also a dynamic moving air mass. (at least when it is flying) that is also dependent on good aerodynamics and minimization of turbulence ahead of the prop.
One thing you can't get past with a strut, frame arm or beam directly under the prop though is that it has a direct, significant and calculable interference with lift that is at least equivalent to how its effective flat plate area directly interferes with the thrust of the prop (it is blowing against its own lift to that degree) and it is a dead loss.
Turbulence can make it worse.
In a stationary hovering craft in calm air, with the prop under the frame arm, especially if it is some distance down, interference with the static air mass can be vanishingly small (and harder than blazes to calculate.)
Best Regards,
Gary
Gary, I've done significant testing on prop blockage both experimentally and computationally. I tested fuselage interference on fixed wing platforms, but it's essentially the same issue as frame interference on a multirotor. My results showed that there is quite a difference in prop efficiency between tractor and pusher configurations (equivalent to top and bottom mounts). There are two primary effects reasons for this:
- The first effect is that which you mentioned, the additional drag of the frame which results in a net decrease in effective thrust. For a bottom mounted configuration this effect is minimised since the flow over the frame will be slower (lower dynamic pressure), less turbulent and have less rotation. This can be summarised as the effect of the prop flow on the frame.
- The second effect is the converse; the effect of the frame on the flow through the propeller disk. More specifically, the presence of the frame slows the local flow giving each section of the propeller a relatively larger local angle of attack. This results in more thrust, more torque, and therefore more power consumption; in different configurations this can either increase or decrease the net efficiency. I found that for pusher configurations (bottom mount) the efficiency can actually be increased to the order of a couple of percent.
These results were found both in my theoretical formulation and my wind tunnel tests. My fuselage tests are slightly different to a multirotor frame, but I'd expect similar results. If your'e interested in a more in depth explanation I can send a copy of my thesis on this topic.
Nicely written Gary...
Gary, as always thank you for your precious technical input on this things, it really does help a lot. I've started to work on precisly engineered assembly of armadilo and this drag coefficient wiki page you've share has really enlighten me what you mean by using just the right size for the job. I'm a designer, i understand a bit of everything but i love to get more into detail when opertunity is presented.
much abliged
With arm above motors/props interference is small but also hard to compute because of variety of forces involved.
But with arm below, the arms interference is relatively straightforward and is at a minimum the direct interference of the arm with the prop wash because the propellers wash is actually directly impinging on the arm itself with a direct force counter to it's lift.
A 1" square rod will place 1" of flat plate interference along the entire length of the rod that is under the prop wash.
That is the equivalent of that percentage of the prop circle that is occupied by the frame arm directly interfering with that percentage of produced thrust (Instead of being applied to net thrust it is actually thrusting against the copter itself in the wrong direction so is subtracted from net thrust rather than added to it.
This is not insubstantial and generally will vary from 3% to 7% or more depending on arm size versus prop diameter. For a non tapered arm the effect decreases for larger prop diameter assuming same arm size.
A round arm with the same outer diameter as one side of a square rod has about half the flat plate area of an equivalent square arm straight on.
Generated turbulence is also added to the loss in efficiency and the net loss can get up over 10%.
With the arm over the prop, total aerodynamic losses are unlikely to exceed 1 or 2% at most unless it is a very wide arm or very close to the prop.
What I have said above is generally true for hovering multicopters, losses for an overarm design would increase somewhat during rapid upward movement because you are now interfering with a dynamic airstream.
Please note Dragans bottom diagram is actually increasing the diameter of the arm to larger than the side of square arm so equivalent flat plate area would be increased.
Minimum flat plate area is actually achieved with an aerodynamic wing like profile () in fact a raindrop profile is generally highly efficient for a given cross section because it is a naturally forming aerodynamic shape.
Equivalent flat plate area can be ultimately reduced to less than 10% of the cross section.
The numbers I have used are not cast in stone, but are at least reasonable ballpark figures applicable to our multicopters in hovering and low speed flight.
They are likely made worse in all instances by high speed or acrobatic flight.
Although in continuous forward flight some extra lift may be generated which can actually increase efficiency a bit at some speeds.
Basically I would use round CF or fiberglass arms where possible of a size appropriate for your copter.
Generally a prop over configuration with round arms permits other design optimizations and has an inherent practicality that can compensate for the increase in efficiency made with a arm over prop configuration.
For additional understanding of flat plate area and drag coefficient, Wikipedia makes it clear:
http://en.wikipedia.org/wiki/Drag_coefficient especially note aerodynamic cross section!
-
1
-
2
-
3
-
4
-
5
of 23 Next