Hi guys,
I got permission from my customer to share this story.
The quad is a DJI Phantom, RTF in 10 minutes when you get your box delivered to you. yes no kidding, it's that simple.
Of course this kind of product brings a lot of new beginner customers (as well as experienced one looking for a small form factor and easy of flying).
So here is the story of this customer 4-5 flight, in a super wide open area, safe, really safe for experimentation.
This is the video of the "trigger failsafe to save my $$$" event.
Why I post this?
I really think we should go in that direction for the APM failsafe out of the box. Granted we have more options but complexity and flexibility is not a reason to not have a basic default behavior even if not set in mission planner (for these beginners not reading instructions!)
(below it is the text version of what was going on his mind.)
Enjoy!
Dany
I didn't know what had gone wrong. I was baffled and confused when it
all went wrong so suddenly. It was only while studying the video later
that I figured it out.
I normally fly in GPS mode. That's pretty much idiot-proof as long as
the Phantom has a good satellite-lock. But I had switched to Atti mode
to experiment with improving video smoothness.
You can tell from the video it's in Atti mode. I flew past my wife at
high speed, then levelled off and swung it around to point back. If it
had been in GPS mode, it should have stopped by itself and hovered in
that spot. But in Atti mode it kept right on sailing further away,
still at fairly high speed.
The problem was that it was such a bright sunny day, with bright snow
and black trees in the background. As it drifted even further away from
me I wasn't sure and couldn't see anymore which way it was facing. So I
switched to Home-Lock and pulled back on the stick, thinking that should
bring it straight back to me regardless of which which direction it was
aimed at the time. That was the plan anyway.
I didn't realize that Home-Lock doesn't work in Atti-mode. I'd never
tried that combination before.
So, when I pulled full back on the stick, the Phantom responded (quite
properly) by backing up even more quickly away from me. And that REALLY
surprised me, because I wasn't expecting that at all.
At that point I should still have been able to figure out what was
happening and recover. But I was so stunned by this unexpected
reaction, I just didn't know what to do. I looked down at the
transmitter to confirm it was actually in Home-Lock mode. But when I
looked back up I couldn't even see the Phantom anymore.
Last I saw it, it was heading for the trees across the lake. So my only
thought was to add throttle to ensure it clears the trees.
It was much too far away from me. Did it just go out of transmitter
range? Maybe. But fail-safe should have kicked in automatically then.
Did it lose satellite-lock? If that's the case, then I'm really in
trouble because I can't even see it anymore to control it manually and
it might already be headed on its way back to China right now.
I was undecided about turning off the transmitter to force a fail-safe
return to home. If it has lost satellite-lock, then turning off the
transmitter will kill any slim chance at all of my regaining control.
But I didn't know what else to do.
Turns out, as you can see from the video, I had a solid transmitter
connection the whole time. I should have simply switched to GPS mode.
Then Home-Lock would have worked perfectly.
But in panic, I switched off the transmitter, and instantly fail-safe
kicked in and saved my butt. But, it had already been flying for ~6
minutes, and I was worried it was going to run out of battery time
before I could get it back to the ground.
Comments
Dany, we did a major overhaul and re-write of the error detection and the failsafe responses in 2.9.
Testing and verifying every possible scenario and response involved hundreds of test flights by dozens of developers and testers prior to releasing the code. Now those tests are being validated by many more hundreds of users and thousands of flights. That has resulted in a few false positives being generated and we have made adjustments to reject them in the future. To date we have had no incorrect responses.
One of the key features of the system is that if it is enabled by default then it has to make good choices when the user does not have it setup correctly and quite simply enabling the failsafe scheme by default for 2.9 would not have been prudent until those few thousand flight tests by those few hundred users were made.
When we are satisfied with enough error free flights we'll enable the failsafe scheme by default in one of the next releases.
Dany, that's an interesting idea. We already have external LEDs, and one of them is already dedicated to GPS reporting. It blinks if it's not locked, and solid if it's locked. But I could make it do something else if it has a lock but not a lot of satellites.
If people want to come up with a scheme, I will see what I can do. Best thing to do would be to make an "issue" on GIT hub and have it assigned to me?
The scheme would have to be simple. Some kind of flash pattern if it's locked, but not enough satellites.
He just needs to set it up for fpv so he can tell which way its going, lol.
That's great Randy, let's make this better out of the box (hehe after the firmware update of course)
I think there is just too many people not really doing their homework before flying and better defaults are needed.
We also need to work on a better way to communicate visually (no ground station involved) with the pilot for the different errors in flight. Losing GPS satellites should be more visible, (Naza-M does red blinking for less than 6 satellites and will even not allow you to engage GPS mode if not good lock)
Not sure the LED blinking / flashing is the most efficient way to signal things. Having to count the number of blinks and remembering what they mean is not optimal in a panic moment. Green, Yellow, Red and blinks combination really helps the Naza/Phantom pilot. I am sure we can even make this better. Who want to start thinking about it? :)
Things like that goes a long way, call it making it a bit more idiot proof. Let's face it, with the prices down on parts we dont need to put $3K on a quad anymore to get something flying $500 is all you need (not the best parts but you get your APM and all)
The target customers are more the average person just looking to quickly fly. Well of course we still need to target the true DIYers we can't afford to not make it safe and simple "out of the box".
Again, not ranting, just trying to push in the direction I see the hobby evolving.
I think we haven't made the throttle failsafe on by default because it just wasn't ready until now. I think we just need another release or two to be sure that it's rock solid. I think we'd also like to make limits on by default once they've had a make-over (which should come in 2.9.2) plus a bit of time for it to settle in...
as per my post: (my ask... make it not required! )
Why is that strike-through?
Crasher, I give credit where credit is due. The DJI MultiRotor systems are pretty solid, and for the most part "just work". Their frames are terrible, and the FOD was a debacle because they don't do datalogging. But I think the controllers are generally really good. If you actually looked at the positive things I say rather than just the negative, you'll see I constantly use DJI as the benchmark for ease-of-use, and performance of the hardware. They got the built-in vibration damping nailed, and sensibly include the GPS and Mag on a remote mounted puck. This is largely why the system performs well for most people. They make it easier to get it right.
Most of my ill-will towards them is because of their helicopter offerings. They are... frankly I think it's sickening the price structure they have set up. $11,300 for a helicopter system that can do the same thing the WK-M can do for $1000. The WK-H at $800 has less functionality than the Naza-M. And the Naza-H is a pretty bad, and was obviously rushed to market to get ahead of the large number of FBL controller makers who are coming out with GPS modules soon.
The entire helicopter line up is set up, very cynically I think, with the Ace One system at the very top end of the range, and they dumb-down everything below it to try and justify the insane price for the Ace One. But when you look at what they're offering for comparable prices on the multi-rotor stuff it... well frankly it's not fair for helicopter users. So yeah, I'm a little jaded.
I am with you Gary on this one! At least default options should be set... but advance users should still be able to change it at will.
This is from the current wiki documentation on Failsafe for APM 2.9.x
The strike through part is the scary part as I know the APM works great on failsafe (most of the time, when it does have good GPS lock of home and so on...) Minimal set of things should be left for the user to know...
Throttle and Battery Failsafe
As of version 2.9, ArduCopter has two failsafe, the throttle failsafe and the battery failsafe. Both are disabled by default.
Throttle Failsafe
If enabled and set-up correctly the throttle failsafe will trigger if:
When a throttle failsafe is triggered one of the following will happen:
If the failsafe clears (throttle climbs back above 975) the copter will remain in it's current flight mode. It will not automatically return to the flight mode that was active before the failsafe was triggered. This means that if, for example, you are flying in stabilize mode when the failsafe is triggered, and this causes the flight mode to change to RTL or LAND, if you wish to re-take control of the copter you will need to change your flight mode switch to another position and then back again to Stabilize.
I think its should be assumed people never read the manual. When it comes to switch selections less is always more. Choice is a bad thing!
Excellent Video From the video it got a long way out I liked the hoover while you dug the hole to land .Thanks for
sharing other people will learn from your experience Have a Great Day!
-
1
-
2
of 2 Next