It is true that as we progress things can get scary- no one wants to see thousands of dollars in copter and camera plunge to the ground. There has been talk of a parachute system, but I think it does need to be automatic. Things happen very quickly and it can be very hard to rely on manual triggering - I can just see a stunned operator hit the parachute button right after it smashes into the cement. How do we make a fail safe system?
My dream would be to use an independant little pic cpu (run from battery) that has inputs for prop sensors (easy to mount optical sensors on arm that measure actual blade spin) and the ppm sig into the ESC. Now we know what the ACM is requesting in blade speed versus actual spin.If the actual spin becomes way out of whack with what was requested we can pop a parachute.
This would cover everything from bad motors, blades, system, etc. and be an independant check on things without the need for a split second decision.
Replies
Hi John,
last november I lost my Quad, it just went up and away.... then I decided to make a parachute system (wich is at half stage).
I did some thinking on RECOVERY & LOCATION, that I'd like to share:
To implement a 100% bulletproof Recovery & Location procedure, is almost impossible, but we can try!
We will be using some if not all the following elements:
For Recovery:
- GPS
- Barometer
- Parachute with deployment system servo activated
(mine is 58" from http://www.apogeerockets.com/parachutes.asp)
you can check the size you need at http://www.aeroconsystems.com/tips/descent_rate.htm
For Location:
- GPS Tracker (GSM/GPRS) (Xexun TK102-2) (http://www.xexun.com/ProductDetail.aspx?id=102)
or
- Loc8tor (http://www.loc8tor.es/Store/Loc8torModelPlanes.aspx)
and/or
- RF Beacon (http://fmtv.us/rf_beacon.html)
(http://www.rcgroups.com/forums/showthread.php?t=1274173)
and/or
- Sparkfun GPRS module (if compatible with APM)(http://www.sparkfun.com/products/9607)
The recovery of a MultiCopter (or any other flying thing) has the following main stages:
STAGE 0 - Initial values
STAGE 1 - Getting the Aircraft "safely" in one piece to the ground.
If your aircraft crashes from high and gets smashed, chances are high
that any device installed in it, will break, will get disconnected as
batery tend to fly away, or will just stop working due to impact. So
if we can get the aircraft "gently" in one piece to the ground, much
better.
STAGE 2 - Locate the Aircraft
Once the aircraft is safely on the ground, if we do not know where it
landed, we have to locate it.
EMERGENCY SITUATIONS:
- Radio Signal Loss
- Loss of orientation and/or control, due to pilot or mechanical failure
- Very Low Battery
- Hardware failure
STAGE 0 - Getting initial values
Upon statring the APM we must:
- Get Origin Home GPS Position
- Get Origin Home Altitude
We need these values to be able to :
- return to home
- keep minimum safety height
- deploy parachute if above minimum parachute deployment height
STAGE 1 - Getting the Aircraft "safely" to the ground:
1.1 - Aircraft Recovery System not using the APM (Parachute)
Using a Parachute to get the aircraft, safely in one piece, to the ground upon
radio signal loss.
- Use a 9ch RX
- Connect servo to 9th channel, this servo wil activate parachute deployment
- Setup FailSafe options on Radio Transmitter, in my case (F9c)
- F/S ch9 to 100% (in order to activate parachute servo)
- F/S ch4 Rudder to Full Left
- F/S ch3 Throttle to 0%
- We should deploy the parachute only if above minimum parachute deployment height
Setting up the FailSafe this way, when losing radio signal, the parachute will
deploy and motors will be disarmed as servos will simulate the disarm stick
position Bottom/Left
This system can be activated automatically upon radio signal loss, or in case
of loss of orientation and/or control, be activated by switching off the
radio, wich will in turn activate the FailSafe.
1.2 - Aircraft Recovery System using the APM (Parachute)
1.2.1 - Upon Radio Signal Loss
Using the APM to detect radio signal Loss, then upon radio signal loss:
- Use chX to activate parachute deployment (servo)
- We should deploy the parachute only if above minimum parachute deployment height
- Disarm motors
The parachute will deploy and motors will be disarmed.
1.2.2 - Upon Battery Very Low Voltage
Using the APM to detect very low battery voltage, then:
- Use chX to activate parachute deployment (servo)
- Disarm motors
The parachute will deploy and motors will be disarmed.
1.3 - Land using FailSafe Values
Upon radio signal, land in slow descend using a pre-set FailSafe value for
the throttle. This value will make the Aircraft descend slowly
1.4 - Return To Home (RTH with APM)
1.4.1 - Upon Radio Signal Loss
Using the APM to detect radio signal Loss, then initiate RTH procedure, based on:
- Getting Home position when connecting the APM, this must be independent
from the GPS hold and other flying modes
- Climb at least 40m to avoid any obstacles on the way home
- Return to origin GPS position.
- Once we reach the origin position, recheck radio signal, if back, then
deactivate RTH, otherwise continue.
- If having Barometer, descend, Land and cut-off motors
- If NOT having Barometer
- keep height and activate the parachute recovery procedure (1.2.1)
or
- Use FailSafe Throttle value to descend slowly
Through out this whole procedure, the Battery Very Low Voltage procedure,
must be checked upon, in case we lose battery while in RTH.
1.3.2 - Upon manual activation using rx chX
We also should be able to activate the RTH manually, in case of loss of
control/orientation. This could be achieved using chX of the APM.
STAGE 2 - Locate the Aircraft
2.1 - Using a GPS Tracker (GSM/GPRS)
Once the aircraft has landed, we just need to make phonecall to the GPS
tracker, to receive it's GPS position and a link to google maps, to easely
locate it.
2.2 - Using an RF Beacon and a 2-way radio receiver
Once the aircraft has landed, we must take a 2-way radio receiver, to
start locating the Tx, by "listenig" to the bips it transmits, the
stronger the bips, he closer we are.
if this can help us to get a good safety system, the better..
Dani
Parachute is simple, for the mechanical part.
Electronic part is a bit harder, as you don't only need to release the parachute, you need as well to stop motors in failsafe condition to be fully efficient.
So there are a few problems to solve :
1) how the failsafe condition is detected (motor spinning is not enough)
2) full detection reliability and no false trigering
3) wich circuit will be responsible for this. there is no doubt it can't be the main AVR.
4) how to cut motors, this should be done at battery level to be sure. not easy because would need a big static or mechanical switch...
Simpler should be better for a safety system. I would use the receiver failsafe for a simple design. So that the parachute is released when transmitter is powered off.
Using a two positions button on the transmitter could be usefull as well for manual trigering.
Great idea. Cheap 3 axis acceleromiter chips have a free fall detector and dedicated output. Use this on a separate power supply in addition to ACM output and crash detection methods, like going past 90 deg or more a rate beyond a maximum. If the cause is a failed motor, the others should stop. That and the independent RPM sensor should cover about any failure.
Having watched my $2k MK plunge into the ground earlier today I like the idea, But! There are lots of failures far more common where the motors are fine.
airframe attitude or rotaitional speed might work i guess..
Regards
D
I've used a parachute with my 3kg quad, it's saved my bacon on several occations. The idea of a complex auto deployment system that may deploy when it should not, or not deploy when it should would worry me. I have a single switch on my tx operating a release servo on the quad. Believe me, you find the switch okay when you need to!
The parachute is from the rocketry world 48" diameter retained by an elastic band around the rolled up chute, the servo releases the band. because of the nature of the parachute fabric it tends to spring open when released and is deployed within a 10ft drop.
Or you release using the radio failsafe, just stamp on your transmitter to release the parachute and shut the motors down! (joke).
I like the idea of an independent system "watchdog" like this. Vibrational noise monitoring could also be another area of investigations, where an imbalance caused by a rotor strike could be picked up. What about redundancy in the control system itself? Has anyone here given thought to "clustering" the Arduino platform for fault tolerance?