What's in a nano-drone?

[VinzC]

Hi all.

I've just gotten back from London where I bought a nano drone at Hamley's and even after a couple of attempts I still cannot seem to successfully fly it. My clumsiness wants to point at the fact there doesn't seem to be any stabilizer on this thing — something that prevents it from oscillating up/down, at least. In fact you have to trim it first before flying it to make it "fly still" if I remember what I read on the manual. After one hour all I have done is lose a blade, the damn thing is unstable, oscillates up/down, flies in any direction and I can't control it reliably. Not that it is impossible, I've seen the steward gently land it on his hand as if he had been doing that all his life! Needless to say it hurts the ego a friggin' lot!



So after wondering a lot, now I'm puzzled as to what's in that beast. The fact that it needs to be trimmed — the 4 trimming buttons are on the remote: L(eft), R(ight), B(ackward), F(orward) — suggests that there's no micro-controller inside. I haven't torn it down yet but I could see a few SOT-23 parts, resistors, capacitors and that's about it. Inside the shell, a plastic skull that's holding the PCB (yes, the PCB, red, is the flying skeleton, clever) there is a (minuscule) LiPo battery (as per the manual), no idea what capacity though. There are 4 sub-miniature motors (~4mm diameter, ~10mm height but don't take it for granted) with blades directly attached to the shaft.

I tend to believe only the remote is accounting for the drift in the commands sent to the flying object, which must be only comprised of a 2.4GHz receiver and a few analog or PWM components, to make the circuitry downright simple.

Has anyone opened such a drone? Does my deduction make sense?

[Mechatrommer]

with the so many china rc toys i've dissasembled, your deduction is almost correct, just a receiver who spit out native PWM language to the mosfet driver. you dont expect that small thing has full fledged 3 axis gyros with mcu processing L1 adaptive control algorithm in it.

[Audioguru]

Most little drone ads say it has 3 gyros but also it has manual trims. Then is it stabilized by gyro (probably not) or by manual (probably)?
I can fly my little drone pretty well because I also fly radio controlled model airplanes and I automatically adapt to backwards directions when it is coming (turn left for it to go to my right side) compared to when it is going away (left is left and right is right). It is easy to adjust the trim a couple of times as the battery runs down.
My drone can do flips and rolls but I haven't tried them yet.

A tiny helicopter I have is charged by the usb on my computer. I forgot and left it plugged in when I turned off my computer and the next day the helicopter battery was dead and would never work again. I could not find a battery small and lightweight enough so now with a bigger and heavier battery it flies differently like a pig. Ever see a pig fly? Something like how a cow flies. But the little helicopter has enough power to carry the extra weight but it swings back and forth and takes a while to stop. 

[Skashkash]

You would be surprised what's in these things.

The  mini/nano quads that I've played with all  had

32 bit processor
2.4 ghz transceiver
3 axis gyro
3 axis accelerometer.
Plus the fets to drive the motors.

And the one I bought a couple of weeks ago also had a camera. (more of a mini quad).

Not bad for 30 USD, delivered next day from Amazon. Crazy how much tech is in these cheap toys.

[VinzC]

Ever see a pig fly?

Like that one?

You would be surprised what's in these things.

The  mini/nano quads that I've played with all  had

32 bit processor
2.4 ghz transceiver
3 axis gyro
3 axis accelerometer.
Plus the fets to drive the motors.

I expected at least some form of stabilizing devices like gyros since (though I might not have mentioned) the drone perfectly lines horizontally, which suggests it has something to sync its rotors perfectly. Now I'm puzzled as to why not make the CPU, since it has one, also control the vertical and plane stability using the accelerometer. At least it would spare the hassle of having to re-align the toy after every crash — and having to trim at all. I'd also doubt there'd be any crash since they result of the thing going haywire for any reason: controlling mistakes or drift...

Now, if that's the case... maybe one can hack it and improve vertical and plane stability? Sounds reasonable?

BTW I couldn't seem to find any schematics on the internet. But I didn't search thoroughly, I admit.

And the one I bought a couple of weeks ago also had a camera. (more of a mini quad).

Ah, right! There were also nano-drones equipped with a HD camera. But the price was double, i.e. £60. Still not bad indeed for a toy that cheap.

[CJay]

Friend of mine runs or ran a website devoted to RC helicopters (bear with me) and he maintains that flying micro helicopters and quads in a 'small' room will cause enough turbulence after a few minutes to make further flight difficult or even impossible for a period of time.

You need a space with plenty of height and lots of volume, the micro quads etc. are no good for outdoor use either as even on a day that seems still to a human, there is probably enough air movement to upset them so stick to indoor flight if you aren't already.

Trimming them can be a bit of an artform, I found I could get my first micro heli 'close' by increasing lift so it just 'skates' and then trimming for stationary, I.E. not turning or sliding in any direction, the second one I could not trim at all, it came with two sets of rotor blades, after half an hour I'd wrecked both and it became a teardown.

[Fungus]

I've seen the steward gently land it on his hand as if he had been doing that all his life!

Maybe he has... 

[mikeselectricstuff]

I think the trimming is mostly about dealing with some fixed offsets that vary from unit to unit, like sensors ( some of this happens when you start it, they  say to do it on a level sufrace), and differences between motors (magnet strength) /props etc.
They all have MCUs and MEMS sensors - they'd be completely unflyable without.
They do take some practice - the main problem with flying indoors is lack of space to make corrections before  it hits something.

[FrankBuss]

This looks exactly like the Revell Nano Quad Copter:



Some Google articles says that it is "skill level 4", with skill level 5 the most difficult, so might need some practice to fly it.

[Skashkash]

 My toy quad does re-calibrates every time it powers up. There is a simple procedure to re-calibrate from a flat surface.  The axis trim adjustments on the remote control do seem to be just some fixed offsets.

The gyro and accel have noise, and there is no altimeter. So you'll never get perfectly rock steady hover or motion. The lack of a real altimeter really hurts.   You,  the pilot are always in the loop correcting.
 
That said, you'd never get these things to fly three inches without crashing if all  that stabilization was not running.

 They can take a little time to figure out flying. But it's really not that difficult. The very cheap ones might be harder to fly or perhaps you just got a factory dud.

All that said, these things are really a lot of fun. And arguably practical when used with the camera to take pictures of the roof gutters to see how bad they are clogged (pretty badly).   

Some pics of controller with reverse engineer info https://github.com/lgeek/h107_rev