Pages Menu
TwitterRss
Categories Menu

Posted by on Jul 20, 2013 in Blog, Video

The Morning After…

…that horrible crash, all night re-builds have been done, a day of testing has already taken place, and AC1 is back in the air, now at version 5.3…

So it’s not the morning, nor is it the day after. It’s sunset, and 2 days later, but that’s not a phrase is it ?

Round 2 – Open Field

Today’s open field should have excellent GPS reception, has only 1 big important building I’m 99% certain I can stay away from, even if my craft goes batshit on me, and for that reason, I am launching far far away from it nearer the other end of the field.

Looks quite close, but I'm miles away from it. Even more true of the Sun...

Looks quite close, but I’m miles away from it. Even more true of the Sun…

Not too close though. There are cows in the field beyond, though I know from experience they really couldn’t care less about the presence of my multirotor. They can barely be bothered to even look up, though of course this is probably partly because I was always at reasonable height when directionally near them…

So – 3 flights, and 1 of them published here for my reference and your amusement…

Talk us through it, JW…

Really ? OK. There are 3 reasons for this narrated flight – or rather I am testing 3 things.

Post-crash analysis

Apart from 1 new motor and 2 new props, I am testing the same setup that crashed so spectacularly in Buriton the other day, the plan being to see if any components will fail, and suggest themselves as the cause of the problem. I have re-mounted the GPS IMU with a new double sided pad, having thoroughly cleaned both surfaces. I will now check its rigidity at the beginning of every flight. Having said that, I notice I don’t check it at the beginning of this flight. It’s a new habit – I’m working on incorporating it into the normal checks. Sometimes it’s not instant.

GPS Lock suspicions

I’m also keeping a very careful eye on how long it takes to find and lock GPS satellites, what GPS mode hover is like, and how much I have to correct it.

I need to establish whether the frankly rather appalling results in my back yard test ground this morning were as a result of rubbish GPS signal in the area and furious wind, or if my GPS components have a problem seeing satellites, holding on to them when it does find them, and of course using that data to stay put when told to.

On the test bench however, my GPS unit outputs data normally and unchangingly for prolonged periods of time, which suggests nothing is wrong with it.

General trust-building / gains etc

As this is now the 3rd test flight session (so maybe 7th flight since rebuild) and I am in a wide open space with no real danger, either from obstructions or difficulty of crash recovery, I can try some slightly more adventurous flying. A little bit more height – maybe up to 150 ft this time, and some more purposeful manoeuvres to start to get the feel of my slightly lowered gains.

Gain Structure

Essentially, I have lowered vertical gain from 160 to 140, and have pitch, roll and yaw gains on my X1 (Aux 5) rotary with specially set travel limits so that maximum is only 20% above the mid setting, and minimum is just 15% below it. This means I can adjust my gains in-flight without worrying that I am altering anything too drastically.

The mid points are set to: Pitch & Roll = 140, Yaw = 125, Vertical = 140, attitude gains to 140 uniformly for pitch and roll. Though that, and Vertical don’t change with X1 – their real-time control is inhibited.

So for this one, I am back to nose-in eyes-on (no choice about that until my new FPV cam arrives !) and my gains have been lowered as Rob at Aeroxcraft suggested, so I can try and get less shaky descents on the gimbal. So I’m testing that too…

Cows - don't care about multirotors. They just don't.

Cows – don’t care about multirotors. They just don’t.

And that’s all there is to say as I say the rest on my narrations…

No gallery today – just a few half decent screen-grabs from the various cams I had going…

Results

Back in the sunset photography business ? Apparently so...

Back in the sunset photography business ? Apparently so…

In 3 flights, nothing went wrong apart from a tiny bit of GPS drift in flight (nothing unusual), and I found a battery that had a dodgy solder joint to a cel. I am beginning not to trust Zippy Flitemax batteries. Sure – they don’t catch fire easily, even in crashes from height (onto grass), but I’ve lost 3 in a year to dodgy cables or soldering.

So, if nothing is wrong with any components (3 6-min flights is usually enough to show up any problems) then, I am left with little option but to blame the GPS for yesterday’s crash. I can’t prove it came loose in flight, and from looking at both on-board and ground-cam footage, I can’t see anything hanging off, but I can think of no other explanation for why a craft that was perfectly under control suddenly ceased to be.

CAA advisory

My flying friend Andrew Palmer brought SafetyNotice2013012┬áto my attention today – so I am sharing it here. Sound familiar ? It did to me. Please read and take action. I have done, and will do, way past 2014…

Interesting – what I have learned from this is that in the event of a GPS fail like the one I experienced, I should switch back to Atti mode (as I did), but then remain there, no matter what the wind is doing, until the craft is back on the ground.

My biggest, and perhaps only mistake in the crash incident was to try going back to GPS mode to try and stabilize it again, something I wouldn’t have tried had it been even slightly in my head that the GPS disc might have fallen off ! But when I next quit to Atti mode, as the flyer began to fly off out of sight, it was beyond saving, and it did not prevent the loss of control as it had done earlier.

Further to the CAA advisory I would add that should this happen to you, do not try and go into any IOC mode either – both course lock and home lock will fail if the GPS disc is loose or off its mount. Seems obvious, and I didn’t try it, but on another day in another universe etc etc I might have done, and I would have been wrong. Likewise – activating Return to Home would be disastrous.

In reality, the only thing that might have been real-time deployable, and could have saved the day was if I had fitted a quick-release parachute system that I was able to activate from a TX switch. The ones I’ve seen don’t require power cut to rotors to function – you have time to manually do that after the chute has deployed and slowed your descent to 2m/sec, which is unlikely to damage anything it collides with, or itself, past some minor scrapes and bumps…

Conclusions…

Well – the tests will continue, and assuming I check that mount every flight, and continue to get consistent, predictable results, then I will begin flying missions again, with an early focus on fields and hills where there aren’t any people or buildings. When trust in my flyer fully returns, after perhaps 50 more flights or so, then it’ll be back to my favourite rivers and lakes, and quarries and ruins :) But until then, the hills and fields round here are pretty good, and I’ll keep bringing you those amazing Hampshire sunsets and cloudscapes.

Now let’s try that 300th hour of airtime again, shall we ? Tests after crashes don’t count.

JW

Never forget why you're doing all this...

I’m don’t stop considering myself lucky to live so close to this much countryside… though if we’re honest this river really isn’t a winner…

Why am I here ? Must be all that open space...

Why am I here ? Must be all that open space…

Long golden shadows...

Long golden shadows…