2
\$\begingroup\$

Clarification: I'm looking for an electrical component to use in a device that will sense large physical impacts. When an impact event occurs, all it has to do is pull one line high for around a second. There is no requirement it pull the line back down.

I do not want this device tied to outputs from my quadcopter in any way. It needs to be standalone. I understand there are beacons with RF signals one can use for the application I describe, but that's not what I want. I just want an isolated impact sensor. :-)

I like the reed switch idea below, but was hoping for something a little more off-the-shelf. I'm going to give this a few more days before I pick one of the provided solutions.

Original Question


I would like to know when my quadcopter impacts the ground (crashes). I am going to wire up a buzzer or other noisemaker to make it easier to find the crash site.

I've thought about using an accelerometer and I've thought about using a piezo sensor, and I don't know which is better or if there is an even better option out there.

Requirements:

  • Pull one pin high for around a second after a crash.
  • Be able to survive a small crash and be useful next time around.
  • Operate around 5V
  • Be small (yeah, I know this is relative. Let's say smaller than 2 inches by 2 inches)
  • Be lightweight (yeah, this relative too. Let's say "not heavy.")
  • Less then $10 USD

So the question is: what type of impact sensor best fills my needs? I do have questions about what device to use to make a 99db sound, but that will wait for a different question. At the moment I just want a sensor that I can wire up independent of the quadcopter and have it know when a crash has occurred.

Please point me in the right direction.

\$\endgroup\$
3
  • \$\begingroup\$ You seem to be confusing "sensor" and "annunciator"/sounder. Sensor says "Oops! Bother!" Sounder says "here I am". Piezo sounder is "robust enough" unless UAV is totalled but a magnetic sounder (earpiece) MAY be more so in vv high impacts. Piezos are fine for most such use and very compact. Bare piezo element needs AC drive but you can buy units with inbuilt oscillator and DC drive . | These need DC only whereas ... \$\endgroup\$
    – Russell McMahon
    Jan 8, 2014 at 19:07
  • \$\begingroup\$ ... these are driven with AC and These are speakers rather than sounders . For "I have crashed" sensor I'd imagine the uAV has signals, voltages etc that fairly clearly show that it has 'stopped flying'. An accelerometer would do the job but seems superfluous. \$\endgroup\$
    – Russell McMahon
    Jan 8, 2014 at 19:08
  • \$\begingroup\$ @RussellMcMahon, Thank you for the info on sound making devices. I will look over the links you provided. I'm trying to find a sensor external to the quad that can tell when there is an impact. \$\endgroup\$
    – kmort
    Jan 9, 2014 at 0:06

4 Answers 4

3
\$\begingroup\$

One simple solution might be to watch an output from your receiver. For example if there is no throttle signal because you have landed the copter, then activate your sounder. You might not need an accelerometer if you can reuse a normal signal for the task. If the copter is not powered for flight, it's either on the ground or soon will be.

After seeing at the reed switch idea below, I got another idea that is a combination of mine and the reed. Place the reed switch on the motor power lead, there is likely to be enough magnetic force to actuate the reed switch. Use that to activate your sounder. Worse case you would wrap a turn of the power lead on the reed. I used this trick on an old car with a bad ignition relay, the reed wrapped on the starter lead would close to bypass the distributor voltage drop resistor. Work like a hose.. lol.

\$\endgroup\$
1
  • \$\begingroup\$ Good, and totally workable, I'm just trying to create something external to the quadcopter. A standalone device that I could use on other vehicles as well. \$\endgroup\$
    – kmort
    Jan 9, 2014 at 0:09
0
\$\begingroup\$

Alright, I have a good one. Use a normally open reed switch, your preferred noise maker, a bolt, and a magnet.

Mount the bolt under a non-magnetic insulator with the reed switch near by. Adjust the bolt's distance from the insulator until the magnet sticks, but not too strongly.

When the UAV crashes the magnet will dislodge, closing the reed switch and activating your noise maker.

Several minor enhancements can be made. For instance, attaching a spring to the magnet to pull it away from reattaching to the bolt or another magnet nearby to catch it. Or setting magnets on multiple axes to ensure one will be dislodged.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
5
  • \$\begingroup\$ Creative idea... Not bad in theory. But I would worry about how well it would work in the real world: either the magnet not dislodging on crash/landing or coming off prematurely while flying (doing flips or something) \$\endgroup\$
    – Adam Head
    Jan 8, 2014 at 20:51
  • \$\begingroup\$ @AdamHead The same concerns could apply to any detection system which has a threshold. The required force to remove the magnet should be very constant. One just needs to find the right threshold. That's what the adjustment bolt is for :) \$\endgroup\$
    – Samuel
    Jan 8, 2014 at 20:55
  • \$\begingroup\$ Looks pretty cool. I think it would have to be done on each axis, and in both directions. Might get a bit heavy. But looks cool. \$\endgroup\$
    – kmort
    Jan 9, 2014 at 0:19
  • \$\begingroup\$ @kmort A single magnet covers two axes (it will move easily in any direction along the surface it's on), so the most you should need is two. \$\endgroup\$
    – Samuel
    Jan 9, 2014 at 0:29
  • \$\begingroup\$ "Contact Light !" B.A. [1st man to talk on the Moon] \$\endgroup\$
    – Russell McMahon
    Jan 9, 2014 at 8:39
0
\$\begingroup\$

Since your goal is to find your drone once you have lost it, I would go in the autonomous beacon direction. Relying on the crashed drone battery seems a bit overconfident. Some of those key locator beacons have a pretty impressive range and would fulfill all your requirements, except for price. For instance:

www.loc8tor.com

www.sticknfind.com

\$\endgroup\$
1
  • \$\begingroup\$ Thanks, but I'm really looking to create a standalone sensor package that gets me that last 200 meters to the bird. I fly VFR and can see the general area it goes down in, but when it goes down in tall weeds or small trees, it is difficult to find right away. A noisemaker would fix this "last 200 meter" problem. \$\endgroup\$
    – kmort
    Jan 9, 2014 at 17:36
-1
\$\begingroup\$

So I've decided to go with an accelerometer after all. I know the question was looking for other answers, but this turned out to be the easiest. The one I got has free-fall detection built into it, so a simple microcontroller should be all I need to talk I2C to it. Thanks for all your suggestions.

\$\endgroup\$
6
  • 1
    \$\begingroup\$ A crash might be better characterized by a sudden spike of acceleration beyond the capability of the rotors, and/or an absolute steady state in a non-vertical orientation. It will not necessarily involve "free fall" unless you cut the power at altitude. Conversely, some legitimate flight maneuvers do involve free fall, terminating short of the ground at a softer/longer peak acceleration achieved by resumption of rotor lift, rather than a ground impact. \$\endgroup\$ Jan 31, 2014 at 4:20
  • \$\begingroup\$ @ChrisStratton I concur. There will be two ways to trigger it. 1) too long of a freefall (as defined by user code on the microcontroller, or even disabled if I want), and 2) big spikes followed by no movement for a while. \$\endgroup\$
    – kmort
    Jan 31, 2014 at 13:19
  • 1
    \$\begingroup\$ And then there's crashing into a flexible tree that sways in the wind... \$\endgroup\$ Jan 31, 2014 at 15:43
  • \$\begingroup\$ @ChrisStratton I'm very hopeful I'll be able to tweak it such that after a big impact, little or no movement will be detectable. I haven't prototyped yet, but will soon. Thank you for your suggestions. It's always helpful to have more cases to consider. :-) \$\endgroup\$
    – kmort
    Jan 31, 2014 at 16:32
  • \$\begingroup\$ Things would be so much simpler if you were willing to use the radio link output, current in the motor leads, or something like that as an input instead of tricky crash detection rules. Using the other systems (shutdown) as an input doesn't mean you have to use them for power or necessarily even make any electrical connection at all. \$\endgroup\$ Jan 31, 2014 at 16:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.