Looking for way to kill power to high current loads

[Mattjd]

Hi, working on a quadcopter project at school. We have four motors right now, am thinking of going to six and their controllers need to be supplied with 25 volts and are said to supply a continuous 40 amps

the controller in question is

http://www.helipal.com/hobbywing-flyfun-40a-opto-esc.html

So the total supplied current is going be pretty high.

The competition has a requirement of a kill switch. We're thinking of hooking up a relay to each controller and having 1 signal to kill all of them at once. I've been looking at automotive relays, and am looking for opinions.

[Benta]

A "Kill Switch" for this kind of thing is not done with a relay. It normally needs to be a mechanical way to cut off power. On race cars, it can be a short cable with banana plugs on the outside that anyone (eg, resque personel) can pull.
Something controlled over the RC equipment I wouldn't call a kill switch.

[Mattjd]

 Understood, however I'm trying to explore a way to rc power to the quad copter bc in past we had to manually unplug the battery while rotors are still running and it's quite dangerous

[Benta]

The idea of a kill switch is that it needs to work when everything else fails.
This means it can't be RC controlled, because if you have massive interference it won't work.
Infrared, perhaps?
The switch itself need to be preloaded, so a weak impulse or a crash will release it.

Lots to think about 

[Marco]

Maybe they don't trust their software not to lock up?

The motors are brushless though, it's not like a shorted MOSFET will make them run continuously. Just disconnect the ESC's control wires, easier than trying to disconnect the power.

[Benta]

Marco, it's a kill switch.
Nobody cares about the technology behind, it's there to allow a layman make the thing inoperable.
That you know it's a brushless motor doesn't help, OK?

[Marco]

OK?

Heh, internet tough guy You might be having fun brain storming how to reliably get the kill switch signal in there, but I'd suggest not trying to command me what do discuss. It's antagonistic and futile, I only obey mods who do that.

The laymen pressing the kill switch just wants the rotors to stop, whether internally disconnecting a couple of signal wires accomplishes that or trying to block a couple 100 Amperes is irrelevant to him.

[Mattjd]

As I mentioned, putting the killswitch aside, does cutting the power to the fan controllers using automotive relays seem feasible? We'd like a way to turn on and off the copter through rc.

[BradC]

As I mentioned, putting the killswitch aside, does cutting the power to the fan controllers using automotive relays seem feasible? We'd like a way to turn on and off the copter through rc.

Feasible but heavy.

[Marco]

A starter solenoid weighs well over a lb and takes amps of coil current to close. They aren't designed to run continuously, so no, it's not an option. You need a 300A+ MOSFET based switch if you really want to kill that power in a reasonable way.

Just disconnecting the ESC control wires is much simpler though and the ESC continuing to drive the motor without control input is an unlikely failure mode.

[Mechatrommer]

The idea of a kill switch is that it needs to work when everything else fails.
This means it can't be RC controlled, because if you have massive interference it won't work.

hence activate "in-circuit" kill switch when there is no transmission detected for sometime. and then put another short cable on the out side in case anybody can manage to reach it while the copter is tumbling upside down with full power motor. this relays are the lightest we can get we can parallel 5 of them for each ESC, but i dont like this mechanical idea in a portable and lightweight prioritized item like a quadcopter, it can take up many space, draw more current through their coils everytime. i would say the kill switch should be the mosfets in the ESC, make them as robust as they can, double the capacity etc and make it right in the copter's FW. a short cable on the outside as mechanical kill switch only for the organizer's entertainment.

[Marco]

i would say the kill switch should be the mosfets in the ESC, make them as robust as they can

Even if they fail short the motor will stop, so their robustness isn't really relevant to their ability to kill the motor.

[Mattjd]

im looking into power mosfets to act as a switch, but am having trouble thinking of how to actually press the button.

[Benta]

Heh, internet tough guy You might be having fun brain storming how to reliably get the kill switch signal in there, but I'd suggest not trying to command me what do discuss. It's antagonistic and futile, I only obey mods who do that.

Stay cool, it's nicer for everyone.

[nctnico]

Put a fuse in series with the battery (if it ain't there already then it is a good addition) and use a thyristor to simply short the power (and thus blow the fuse). Thyristors are very rugged devices and this way you can be sure the entire device is powered down.

[Teledog]

Perhaps this unit..and have it switch a HD contact relay?

https://hobbyking.com/en_us/turnigy-receiver-controlled-switch-1.html

[Belrmar]

Put a fuse in series with the battery (if it ain't there already then it is a good addition) and use a thyristor to simply short the power (and thus blow the fuse). Thyristors are very rugged devices and this way you can be sure the entire device is powered down.

Well having to replace a fuse each ime you land is going to get a bit annoying and expensive, i would try to check if the ESC's have some kind of power down pin that you can feed from the receiver... just brainstorming. If we dont want to get heavy i would try to go to less of a kill switch to more or less a forced shutdown

[nctnico]

Put a fuse in series with the battery (if it ain't there already then it is a good addition) and use a thyristor to simply short the power (and thus blow the fuse). Thyristors are very rugged devices and this way you can be sure the entire device is powered down.

Well having to replace a fuse each ime you land is going to get a bit annoying and expensive, i would try to check if the ESC's have some kind of power down pin that you can feed from the receiver... just brainstorming. If we dont want to get heavy i would try to go to less of a kill switch to more or less a forced shutdown

I'd expect a kill switch to be an emergency device and not a regular shutdown (for that you can use a couple of fat MOSFETs).

[Phoenix]

The competition has a requirement of a kill switch.

Is there any specific rule wording around what a "kill switch" is required to do? It's not a matter of what you or anyone else unfamiliar with the specifics of the competition think it should do, it's a matter of complying with the rules - and proving/arguing how your implementation meets it.

(I've been involved with electric racing car design).

[Mattjd]

The competition has a requirement of a kill switch.

Is there any specific rule wording around what a "kill switch" is required to do? It's not a matter of what you or anyone else unfamiliar with the specifics of the competition think it should do, it's a matter of complying with the rules - and proving/arguing how your implementation meets it.

(I've been involved with electric racing car design).

right, so from what I've gathered a kill switch is not even needed.

So what I've been working on in my free time is a switch to turn power on and off to the motors.

A draw up of my plan is here

I plan on using a power mosfet being driven by a gate driver, the input to the gate driver will be from a DAC, taking in a 5.5v (currently unknown frequency) PWM signal from the quadcopter receiver. This will allow us to turn the quadcopter on and off using the transceiver.

My parts list are

Power MOSFET

https://www.infineon.com/dgdl/irfh8202pbf.pdf?fileId=5546d462533600a40153561f7c351f0c

Gate Driver

https://www.diodes.com/assets/Datasheets/ZXGD3005E6.pdf

http://www.ti.com/lit/ds/symlink/dac081s101.pdf

I believe my design is sound/valid, I just need to find the values of Rsource/Rsink from the gate driver and determine if I need a gate-source resistor too. I believe we will be getting the signals for ~SYNC and SCLK from a raspberry pi that we are also using in another system (not entirely sure though, one of the more knowledgeable quadcopter members still needs to confirm).

In the end, I'll have 1 DAC and gate driver, driving 4 power mosfets (ideally, of course 4x everything is still doable).

I'd love any and all criticisms.

[xani]

The idea of a kill switch is that it needs to work when everything else fails.
This means it can't be RC controlled, because if you have massive interference it won't work.

It sure can, just not in usual way.

Just make it react to two inputs. One is "kill switch signal". Other is "RF is down for more than ~2 seconds".

Then if you lose RF for whatever reason, it will shut itself down,  it ensures that there will be no case where engines have power but there is nothing to control them.

Of course, then any interference will cause your quad to drop from the sky...

[Rerouter]

For most quadcopters, this would be called the failsafe state, e.g. loss of communication, or other weirdness, most recievers can be programmed for a failsafe state, so you could use an output to trip your kill switch and disconnect power from the motors, wire this in series with a physical kill switch of some kind on the quad (probably on the belly, my experience is its the safest place to grab a hovering quad)

[Kalvin]

Losing RF-signal and communication link due to interference or some reason should deactivate the power supply. Yes, you will be losing the copter as it falls down and may hit someone. Solenoid driven contactor ("relay") would be a simple solution. Pressing a kill-switch or losing the communication-based watchdog will de-energize the contactor and the power supply will be turned off. In order to energize the contactor one has to press a start button so the circuit may not energize itself after it has been shut down.

[Ian.M]

Why would you do that for the loss of signal failsafe state?   If you kill the power with the craft out of ground effect, its certain damage and at any significant altitude or velocity, it will probably be destroyed.

You should *NEVER* be flying the craft close enough to anyone (except the operator team), that unexpectedly cutting power so it drops out of the sky is the least worst choice to minimise the risk of injury.   I'm not saying it shouldn't have the capability to disarm/shutdown on command in flight mode, just that it should be explicitly commanded as pilot judgement is required.

If the LOS in flight is due to interference,  the intelligent thing to do would be to initially auto-stabilise and adjust throttles to closely approximate hover power.   If its got an altimeter enter hover mode,  If its got a GPS (and its got lock), enter position hold mode.  If you've got enough sensors to support auto-land, sustained LOS should probably trigger that with an option to trigger RTB mode for operating over water or other hostile terrain.

OTOH if it is landed, LOS should cut the power + disarm.

The problem with killing power is how to do it without making the craft much much larger and heavier, so making it more dangerous.   If your ESCs have an accessible reset pin on their MCUs, and the ESC firmware hasn't disabled the reset pin, grounding all the reset pins will shut down all the ESCs, even if they've got a firmware bug.    Unfortunately I don't think that's common.       

Many ESCs will cut power on prolonged loss of throttle signal, but that's relying on firmware so its not a true killswitch.

Probably the best option is to route the main power through a pluggable link, with fairly low retention force and an accessible loop of cable sticking up, and have a 4' stick with a strong hook on the end to yank it out.   There is some risk while arming it, but that can be mitigated if the landing gear has loops the stick can be passed through - put the stick right through the gear and stand on it and the craft isn't going anywhere while you plug the link in, then withdraw the stick.
 

[Rerouter]

Ian.M It comes down to your local regulations somewhat, for Australia if the device lacks a return to home function, or position hold. it its intended to fall out of the sky on signal loss rather than continue in whatever direction its flying in, the assumption being that a small drone with stopped propellers is generally less dangerous than one flying forward at what could be maximum speed.

For the DJI crowd, gps is baked in, and the hold or return options are met, for the racing group, a GPS reciever is weight that is generally ditched, with the controlled drop being preferred. not a landing, but a drop, if a propeller is partially damaged the drone can shoot off in 1 direction while trying to reach level without ever quite reaching it. leaving a motor kill / stall the only reliably safe option

And finally, racing drones due to there light weight mostly bounce on grassy areas undamaged, even on concrete you would rarely break a frame, maybe bend a motor and break some props, but you carry spares. I can say my own crash counter is in excess of 150+ on a 1.3kg drone some from almost 60m up, and have yet to break a frame of the core electronics, many props damaged, but still quickly repairable, the only crash i struggled to walk away from was top speed into a light pole (~135kmph). that bent a motor, trashed a camera, and broke a number of screws.