This article is very outdated. Small brushless quadcopters are readily available now. Such as the Hoverbot Nano [1] mentioned in another commment, or a Eachine Lizard [2], and lots more.
And you can go even smaller, such as a pepperfish [3]
The brushed designs have been completely outclassed by the new breed of micro brushless quads. The Hoverbot Nano is one of the best examples of lightweight brushless design.
Cheers, that looks very interesting. I'm just wondering what are the main advantages of brushless vs brushed motors for quads. Do brushless let you control the speed of the propellers more easily?
Brushed motors are around 75% efficient, brushless are around 90% efficient. This has two effects: obviously, you get more flight time from your battery with a more efficient motor, but that's not the main effect. The big one is that power output is limited by the combination of how much heat you can remove (which is roughly the same, perhaps 20W) and how efficient your motor is. With a 90% motor, your 20W of heat max means you can have 180W of thrust, while on a 75% motor the same 20W of heat only corresponds to 60W of thrust!
Brushless motors need some high-speed switching electronics to control the motor, swapping from one set of coils to the next at just the right frequency. Brushed motors are stupid easy - just connect two wires, and the voltage corresponds to torque (and thus speed). Any hobbyist could build a brushed controller in their sleep, but it takes a lot of dedication and engineering to design a brushless controller. Typically, this complexity is abstracted away from the end user in a prepackaged Electronic Speed Controller, but most micro quads integrate the ESCs on the control boards.
Brushless motors are more expensive than brushed motors. They're arguably less complex, but we've had many years to figure out how to build super cheap brushed motors, and can make them very small and very cheap.
In general, brushless motors are so much superior that you shouldn't even consider brushed motors in a quad (unless you need to hit a $19.99 price point), but do recall the cube-square law: Length squared is area, length cubed is volume. Area corresponds to propeller swept area or wing area and thus lift, while volume times density is equal to the weight. This is why a hummingbird or sparrow can dart around, while a swan requires a running start and a turkey needs to sprint and buffet (yum) the air with its wings just to get airborne for a few seconds and roost in a tree overnight. And it's why a $19.99 plastic micro quad with brushed motors will fly just fine, but to lift a DSLR you'll need brushless motors and carbon fiber.
Just to expand on what folks are saying with some stuff outside the world of quads - Brushless are stupid high power density but the ones without sensors (i.e. most of ones used in quads) have issues when transitioning between forward and reverse. For ground based systems with high inertial loads (or a fancy way to say "if I stuck this thing to a wheel and tried to drive a system around with it" you will need to tweak it to get it working. This is because of the way the startup for a BLDC motor works when using back-emf. Back-emf being caused by the magnets on the stator moving past the non-driven coils requires that the motor be moving to know where it is. The way most cheap BLDCs get around this now is by just giving the motor a series of high current kicks so the stator starts moving and it can know where it is. With low inertia loads (props) this works REALLY well. With higher loads it can result in cogging and having your ESC release the magic smoke.
All that being said - software is getting smarter and this problem is being handled. SimonK has options to make it work. But if you want a ground based system the easiest approach is a brushed motor because for those most part they "just work".
Motor guy here, not into quad copters though. Brushless motors with optimally tuned software can have higher efficiency which translates to longer battery life. The downside is you need 3 half-bridges to drive one. A brush motor spinning in one direction needs only a single low-side FET. The cost in transistors and software is significantly higher for brushless, but the efficiency gain is real. I'm not so sure about any dynamic performance benefit, you can do about the same with a brushed motor if you're willing to drive it with a full H-bridge.
IIRC brushless are just more efficient in general. Since the power to the coils is entirely controlled by a microcontroller you can modulate how far "ahead" the magnetic field is of the motor, resulting in roughly the same performance at all speeds.
This as opposed to a brushed motor where the brush geometry in relation to the rotor sets the magnetic field a fixed distance ahead of the rotation. Resulting in the magnetic field being too far ahead at slow speeds and not ahead far enough at high speed.
In any case practical experience has shown me that a brushless motor is way more impressive than a brushed motor of a similar size and weight.
It's very much like how cars now have electronic ignition control while older cars had mechanical ignition control with a rotating distributor. Because software controls the electrical energy inputs to the device, the timing is infinitely variable, which is generally not true with mechanical systems. Brushed motors also tend to not last as long because the internal rotating switch (the brushes) undergo constant wear. The big downside of brushless motors is that the software needs to know where the motor is in its rotation cycle at all times, but that's typically done with indirect methods (back EMF) in small motors, and with direct methods like encoders in larger ones.
In my experience the brushed motors used for that size either break quickly with crashes, or just stop to work within 4-6 hours (max!) of flight time. The brushless motors are much more solid and less likely to fail. I have broken many of those too, but they last a lot longer.
Even for brushed (Tiny Whoops) you can now get FC which run BetaFlight.
I've got KingKong Tiny 6, and already have all the components to build 2" with Tomoquads Mikro 2 frame and brushless motors.
with the recommended parts, minus the camera/vtx combo ($20), your total is roughly $70. for $60 you could get an eachine qx90 which is almost identical.
or you could even get much, much better quality/control out of a brushless build with the babyhawk, which goes for around $100
if you're new to flying drones, i wouldn't recommend this at all
When seeing the components the first thing that came to my mind was "outdated", afterwards I saw the article is from 2016.
If someone is interested in a more sophisticated approach he might want to check out the RotorX Atom:
https://rotorxracing.com/products/rx122diypr-atom-diy-kit-wi...
In terms of hardware it is practically the same. However Crazyflie has its own development while these cheap controllers can work with any standard open source flight software such as Cleanflight, Betaflight, etc. They are also cheaper than Crazyflie, you can have a drone for half the money.
And you can go even smaller, such as a pepperfish [3]
This is a small brushless quad with 0703 motors with a 1 cell battery. https://i.imgur.com/EdFFfsc.png / http://rotaro.no 54mm in diameter
[1] https://hoverbot.io/ [2] https://www.banggood.com/Eachine-Lizard95-95mm-F3-FPV-Racer-... [3] http://fishpepper.de/projects/pepperf1sh/