So I saw the episode of Mythbusters the other day where they built a jetpack/personal flying machine.
1 comment that stuck with me was that cowlings (enclosure for the propellers) were supposed to increase thrust through the propellers. Anybody have experience with this? Would it make sense to add cowlings around the propellers of my hex for increased thrust or would the additional weight offset the gains in the scales these multicopters are created at?
Apparently it has to be quite a tight fit to create suction.
Thanks in advance.
It is hard to answer to your question. There are advantages and disadvantages.
The primary efficiency-related purpose of a duct/shroud/ring around a propeller is to fix a problem created by abnormally high disk loading (trying to get too much thrust out of a prop that is too small).
All wings, whether fixed or rotary, generate thrust by accelerating air. A natural by-product of this is the creation of a low pressure zone above the wing and a high pressure zone below. The pressurized air seeks a way to return to equilibrium, and fortunately, most of it goes in the desired thrust vector direction. However, the air near the tip of the wing finds a path of least resistance by merely wrapping around the tip toward the low-pressure zone, creating a vortex field or "tornado" like effect. These wingtip vortexes can often contain (waste) a lot of energy, and the higher the upper and lower pressure differences, the worse they get.
You've probably seen the little fins sticking up on the wing tips of some larger aircraft - these are there to mitigate this undesirable vortex phenomenon and help the efficiency of the wing. On a rotary wing (propeller), you really have only two options - you can make the blade longer, thereby accelerating more air and producing the same thrust (F=MA) with less pressure differential, or place a solid, fixed barrier near the blade tip in the form of a ring/tube/cowl/duct. Naturally, the dimensions and contours of the barrier are as much a subject of engineering analysis as the design of the blade itself.
It is interesting to note that quite a number of aeronautical techies get wrapped up in the whole duct-design-thing and start evangelizing their use as if they were a panacea of efficiency. In every case, their data compare an already overloaded propeller with the same prop using the magical shroud. Viola! They see tremendous results in fixing a mistake they should not have made in the first place. Having been thoroughly researched by literally hundreds of scientists since the 1930's, the properties of ducted propellers are well understood. Assuredly, if there were such great advantages in using them, we'd almost never see a "naked" prop on any scale of aircraft.
However, there are indeed proper applications for a shroud, the most notable being one of dimensional constraint. If for some reason you must generate lots of thrust from a short prop, such as with an autogyro (gyroplane) where you obviously don't want the prop to collide with the main rotor, then having a duct is a good idea. The Carter Copter and the high-speed-record-setting Sikorsky helicopter both come to mind as appropriate applications of duct technology. If you need a compact form factor for some reason, a duct might be the answer. If you want to reduce the prop strike dangers, it's hard to ignore the potential safety benefits of some sort of barrier, efficiency considerations aside (the shrouded tail rotors on some helicopters comes to mind - they are a significant ground ops hazard).
On a small-scale quad, there is no point in applying a duct for efficiency - there are too many other areas of design attention which would yield greater results. The collision ring bumper is a good idea, especially if you plan on flying around people or pets. But don't expect any great efficiency gains.
Brad, so just to summarize, and confirm my thoughts on this matter, is this statement correct?
If you want to increase the efficiency of a give propeller, you can do so by one of two ways. You can shroud it, which reduces tip vortices, or you can simply make it longer. And the longer propeller will always outperform the shrouded propeller. Therefore you should always use a longer propeller unless you have a physical restriction which prevents you from doing so.
So for example, if you are building a model jet aircraft, obviously you are limited to a diameter which fits within the fuselage. But if you took the exact same airframe, and mounted the motor at the front or back, with an large external unshrouded propeller, it would be more efficient.
Nice summary, looking forward to Brads reply!
@Cliff: Yeah, I tend to get rather verbose at times. Mea culpa. ;-)
@R_Lefebvre: Yes, you are correct. For efficiency sake, you should always make the propeller (rotor) blade longer unless you can't for some reason*. Your metaphorical model aircraft with the larger, unshrouded prop would unquestionably be more efficient.
Basic actuator disk theory gives us the Newtonian momentum-based ideal power calculation. FM is ideal power/induced power, and ideal power for a given thrust is always inversely proportional to disk area, which is a function of the square of the radius. A little more radius goes a long way. In other words, a high FM is laudable but irrelevant if your disk loading is unreasonable to start with.
* Another issue is when the propeller tip speeds start to become greater than about mach 0.7 or so, forcing a whole other set complications known collectively as "compressibility effects." I hope we never see drones zipping around at several hundred MPH for all sorts of reasons, so we can safely ignore such issues.
I think there is a direct correlation to aircraft wingspans. For all the same reasons, a longer wing is more efficient than a shorter wing. The problem is there are often other constraints which prevent increasing the wingspan to infinity. Manoeverability is one when considering an aircraft such as a fighter plane. ie: it's slower to roll a long wingspan than it is a short one.
Another is simple storage space. Airliners need to fit in the airports.
Then physical effects. It gets harder and harder to stretch a wingspan and retain the strength/weight ratio needed to prevent them folding up. Simple cantilever loading problem.
For an electric multicopter with fixed-pitch, control response time is proportional to the rotational inertia of the blades, which also (undesirably) increases with the square of the radius. There are always compromises when you're designing anything; the key is to make informed choices.
Thanks Brad, interesting and as usual very informative.
Oh right, there's that too.
So one could make a case for using smaller, but shrouded propeller, with the idea of making it more responsive.
Though, I think we'd also have to consider in this case, no just the inertia of the propellers getting smaller, but then you also have to spin them faster. I'm not sure which would be more responsive at the end of the day, off the top of my head.
Anyway, just pointing out the design considerations.
Thanks Brad. This thread (and others on which you have commented) have distilled out some key information. (I continue to comb the DIY Drones blog for distilled information about tradeoffs related to: physical component placement on a small quadrotor, HMI alternatives for "on the ground" devices, and adaptable pre-flight (at home, in field) checklists. Would also be nice to discover something like a systems engineering perspective on quadrotor systems.)
I read both, and thanks! Can't say I followed most of the math (been a few years...), but there were some quotable text chunks in both. For example, from the ...833.pdf: "Helicopters are intrinsically unstable. They exhibit oscillating modes that are slow enough to be controlled by humans in full-scale vehicles, but these are much faster in small UAVs." No kidding. I consider myself to be the most unreliable component of my evolving quadrotor system. (If my simulator experience is an indication, I would go broke if I got into flying model helicopters.) Both papers have influenced how I plan to prepare and do tethered testing... Thanks again.