How to Make a Miniquad Fast
Over the last few months there's been a good deal of discussion about how to make a fast miniquad. It's been interesting to watch and I've mostly just been observing. I've seen a lot of interesting discussion. A few weeks ago I posted some thoughts on RCGroups, and I thought it was worth it to post them here as well.
Everyone so far seems to have a tiny piece of the speed picture, but I haven't yet seen anyone grasp at the larger picture. The primary problem is that essentially everyone is approaching this issue as if it were a only a one or two variable problem, but in miniquad reality we have a continuum along at least 6 axes. The most simplistic way to begin to approach it is to look at energy vs RPM (Pitch Speed) vs Thrust vs AUW as well as two components that impact those values Prop size* and Motor. There is an interesting relationship here between these components. As you go up in prop size the amount of energy that it takes to sustain a given RPMs increases drastically, but the amount of energy that it takes to generate a certain amount of thrust reduces drastically. As you go down in prop size the amount of energy that it takes to sustain a given RPM decreases, but the amount of energy it takes to generate a certain amount of thrust increases.* (See footnote!) Of course to introduce another layer of complexity, motors (stator size and KV, and design paremeters such as magnet choice, airgap, stator laminations, etc) impact that for each prop, but motor size is always a balance between KV, weight and efficiency for equivalent power, so it really is more of an intersection point between AUW and prop size and RPMs than a unique factor on it's own. Then you add in the all up weight of the quad, and you get a series of lines that all intersect. The trick here is to find the point where all those lines meet for your AUW, which will optimize the speed for that AUW. Then the additional trick is to find the point across all AUW, and all prop sizes where the energy consumption for thrust and RPM balance is ideal.
There are some fine points of assessing the thrust vs AUW vs parts of these cintinua. If thrust is too high, the quad ends up not making the best use of the capacity of the motors to generate RPMs, if thrust is too low it ends up not being able to reach a high enough forward angle to actually use the RPMs that are being generated. Based on a lot of experience and many builds, the ideal thrust to weight ratio appears to be between 8:1 and 9:1 (via the old fashioned method of assessing thrust to weight, taking static thrust and divide by craft weight). If there is a higher thrust to weight ratio, then the goal would be to switch to a lighter props that acheive higher pitch speed and generate less thrust. Those ratios will shift around some depending on the pilot, but the principle is the same. Of course the handling of the quad also plays a role in the propeller choice, so you can see how quickly this gets complicated.
The reason why 5" has traditionally been so fast is because it sits in a spot in these continua where there is the most flex room. Going to 6" very quickly you loose ground in terms of RPM, so you can easily have a heavier quad that handles well because thrust is easy to achieve, but it will be slow because RPMs are harder to achieve without increasing the weight to compensate for the increased energy requirements. If you go down to 4" and 3" you end up with the opposite problem, where RPMs are easy to achieve, so you can get pretty fast, but then thrust is hard to achieve, which means the range of AUW is very limited, and you have to go super light before you can actually take advantage of the RPMs. Fortunately current draw in general is much less down there, so weight is easier to shave, but it still takes a lot more planning and careful thought on what motor prop and battery will give you the best speed. Even then if you're looking for top linear speed, a small prop size quad is going to have a very very hard time keeping up with a 5". There is just no way to get the ratio between the elements as ideal as a 5" rig without moving the weight into an un-obtainable region of that continuum.
There are two additional pieces that matter when it comes to speed on a track vs simple linear speed that need to be taken in to account, and those are cruising speed and the question of momentum and acceleration. The discussion of the percentage of downward thrust vs horizontal thrust as a factor of throttle percent and weight that Joshua Bardwell and UAVFutures gave deals with the cruising speed factor, and the relationship between camera angle and the balance of those factors that I mentioned earlier in terms of how it impacts piloting, but they both entirely ignored the second factor, which is momentum and acceleration, which is the point that Bob Roogi made in his video, and where the ultra-light quads really have the advantage.
It is also worth noting that what we mean by "fast" is often not clearly defined. This is a big part of the reason why the discussion has been so varied. Acceleration can often feel like speed to a pilot, and may actually result in shorter track times on many tracks even though linear speed is actually slower. It is entirely possible for a quad to feel viscously fast because it can accelerate to speed so quickly, but in reality have a slower top speed than a different quad that "feels" slower due to slower acceleration.
So everyone has a piece of the picture, but no one has it all, and how all of these factors work together to change the feel and response, as well as speed of the quad. In the end it boils down to a simple practical decision of what feels best to the pilot who is racing (or freestyling!), and it is definitely possible for a quad in the right hands to beat another quad in not the right hands regardless of where they sit on these continua.
*I'm drastically simplifying "prop size" to include a huge number of additional variables, including solidity, drag profile, number of blades, moment of inertia, thrust and torque coefficients, etc, etc, all of which impact the amount of energy it takes to generate thrust and reach certain RPMs. That's a rabbit hole you can go down a long ways, but all that can be simplified down to seeing how the prop relates to the two variables of thrust and RPMs. Understanding the rest of those complex interactions is not necessary for an understanding of how a prop fits into the continua that I am discussing here.
Thank you for the great information in this article and for the tremendous amount of data you compile on motors and ESCs. I listened to your interview with Joshua Bardwell and would have loved for it to have lasted two or three times as long, giving you more time to answer in depth.
I do have a question about the props you use for testing. HQ and DAL have so many variations now it's hard to know which you're actually using. With V1s, V2's, V1S, etc. as well as two different types of Cyclones. Do you have specific info on which you are using? For example you might list a HQ 5x4x3. Is there more than one HQ prop that fits the description?
Another thing you mentioned that really caught my attention was light vs heavy props. I had wrongly assumed that the pitch was the biggest factor in determining load with physical weight being the second. This is obviously not always the case. I had quit flying Cyclone 5045s on my RS2205S-2600kv motors because I was worried about overloading the motors and ESCs but then I heard you describe them as a lighter prop several times. Do you have any kind of ranking by load on 5" props? My main quad had an AUW of 590g with RS2205S-2600kv motors and I'm building a second that should be almost the same weight with Armattan OOmph 2205-2550kv motors--are Cyclones a safe bet with both of these? Both have 30A BLHeli_S ESCs.
Thanks again for the extensive knowledge that you freely share!
Thanks for the article. Curious if you know of any work that's been done to assess the aerodynamics of a quadcopter and what effect different frames and prop sizes might have on it.