47 replies to this topic

[Stu M]

Hello OpenPilot,

Like a number of people, I started off by trawling the forums of regular RC sites. I'm not going to say that people were not helpful, just maybe a little dismissive/inexperienced in their suggestions ... then a fellow Quad owner pointed me towards OpenPilot.

As we know, flying UAVs etc can be an expensive hobby. I bought a Gaui 450X (which in fact isn't made by Gaui at all, with the only Gaui parts on the thing being the gyro and the 10" blades). In normal flight, she flies exceptionally well. I have twiddled with all the settings on the radio and she is now beautifully balanced and stable - not too jittery on the gas and a pleasure to pilot. She flies magnificently and I am well chuffed when I get her back in one piece (!). However, there is a serious problem when trying to descend ...;

After flying her for a number of months, I soon found that the mainframe, arms and legs, aren't as strong as some people would have you believe. They may well be made of modern carbon-fibre but the merest prang often means changing something or other.

The next thing was to invest in spare parts so I bought 2 full sets of spare 10" blades, 4 spare motors and an entire mainframe, arms and legs. Except for the electronics, I have a full machine. It would appear that at the end of 7/10 flying sessions, I am to be found bent over a table, fiddling around with Allen keys and swearing profusely when a screw pings out and rolls off the table mat. This is not (and I can say this with all modesty) a reflection on my flying skills.

Let's look at an example. Today, I was hovering 10m up on a windless day (absolute '0'm not a puff of anything). The outside temperature was around 23 degrees, 54% humidity and 1019.1hPa (for anyone interested). I gently reduced the power and she started to descend. Then all of a sudden she starts to 'flip-flop', gently at first but then totally out of control. Adding power does absolutely nothing and the the 'flip-flop' action is far to brisk and violent to allow anyone with pilot reactions to counter-act with pitch or attitude adjustment. She then flips right over and plummets to the ground and shortly afterwards, I am to be found hunched over a table replacing this super-strong carbon-fibre ... you can imagine the rest.

I have tried descending whilst flying forward, backwards, sideways and even in sweeping arcs but the result is still the same; without touching the power, the motor (sound) pitch go up a tad as she fights gravity and compensates. Today, I even saw the forward right blade slow to a standstill for a fraction of a second before starting up again. Odd behaviour for a machine which for 99% of the time can be flown like an extension of your body!

I would really like to get into FPV but I have lost most of my confidence with the 450 as I now have very few spare parts left ... I have heard that getting the 450 to descend is one of it's traits. Maybe there is an OpenPilot boffin who has come across this problem?

Anyone who has managed to get to the end of this message without deriding me completely would care to have a stab at answering?

All the best,

Stu

[ridgebackred]

Stu,
Vertical descent through prop wash is something I try to avoid if at all possible.
If flying in a confined space where vertical descent is required, descending very slowly helps.

However, you have tried that and apparently it did not help.

>>Today, I even saw the forward right blade slow to a standstill for a fraction of a second before starting up again.<<

This may sound like blasphemy to some, but I like flying with undersized props. (Trial and error for each machine.)
Running smaller props at higher speeds has increased stability for me across three different brands of controllers.
By the numbers, I know I am giving up some efficiency. The stability improvements make it a tradeoff I am willing to accept.

It would take a high speed camera and a synchronized real time data logging of the controllers update rate to prove the following pet theory of mine:
Certain oscillation frequencies occur at just the wrong time in the update loop for the controller to see and act upon them.
Changing the prop diameter could be shifting the oscillation frequency into a 'seen' range. Who knows?

[ANV]

If your descend is fast, the thrust of motors can be really small. In a case you have oversized props (at least for this descend rate), the motors might be near stall point and if one of the motors stopped for a moment it is not easy for control board to recover from this - you re likely to get progressive oscillations. The easiest way to verify that is add some dummy weight on your quad. I think ~300 grams is enough.

[peabody124]

 ridgebackred, on 25 September 2011 - 05:09 PM, said:

ay sound like blasphemy to some, but I like flying with undersized props. (Trial and error for each machine.)
Running smaller props at higher speeds has increased stability for me across three different brands of controllers.
By the numbers, I know I am giving up some efficiency. The stability improvements make it a tradeoff I am willing to accept.

I totally agree with this.  I think it's the speeding up of the ESC due to two factors 1) they work like shit at lower rpm's and 2) you have less inertia.  These make a difference - especially for things like decent which is running at relatively low RPMs.

[ridgebackred]

A forum member, Spydmobile has flown multirotors over wildfires.  Extreme vertical wind velocity conditions to say the least.

His empirical observation is that a quad is more controllable than an octo over wildfires. Higher disk loading = better stability is the takeaway observation for me.

Edited by ridgebackred, 26 September 2011 - 12:55 AM.

[SeismicCWave]

>>As we know, flying UAVs etc can be an expensive hobby.<<

Tell me about it.

>> I bought a Gaui 450X (which in fact isn't made by Gaui at all, with the only Gaui parts on the thing being the gyro and the 10" blades).<<

First of all I want to know if it is indeed a Gaui 450. I am not aware of a Gaui 450. I know Gaui makes a 330, 330X and a 550. However Tom Stock (Hoverthings) has taken the Gaui frame and made it larger and call it the VC20-450 frame. Paul Baxter (Witespy has been selling that frame plus hardware he puts together and calling the witespy quad).

The reason I am asking is because I am trying to determine which flight controller you have. All flight controllers behave differently.

>>In normal flight, she flies exceptionally well. I have twiddled with all the settings on the radio and she is now beautifully balanced and stable - not too jittery on the gas and a pleasure to pilot. She flies magnificently and I am well chuffed when I get her back in one piece (!). However, there is a serious problem when trying to descend ...;<<

If you have the Gaui GU344 controller it is definitely not very good in descend. Most multi rotors are not very good at descend. Most vertical take off and landing flying machine with rotor blades are not very good at descent. Ok I don't know about the Harrier but I am sure it is a handful to descend also because that maneuver had killed quite a few pilots.

>>Let's look at an example. Today, I was hovering 10m up on a windless day (absolute '0'm not a puff of anything). The outside temperature was around 23 degrees, 54% humidity and 1019.1hPa (for anyone interested). I gently reduced the power and she started to descend. Then all of a sudden she starts to 'flip-flop', gently at first but then totally out of control. Adding power does absolutely nothing and the the 'flip-flop' action is far to brisk and violent to allow anyone with pilot reactions to counter-act with pitch or attitude adjustment. She then flips right over and plummets to the ground and shortly afterwards, I am to be found hunched over a table replacing this super-strong carbon-fibre ... you can imagine the rest.<<

That can be attributed to a couple of reasons induced by the oscillation during a descent. Most likely your gain is a bit too high. Some controller board will get into an unstable oscillation (I learned that from Peabody) situation and if it gets severe enough the machine will flip.

However the GU344 rarely does that. So you may have a hybrid machine.

>>I have tried descending whilst flying forward, backwards, sideways and even in sweeping arcs but the result is still the same; without touching the power, the motor (sound) pitch go up a tad as she fights gravity and compensates.<<

That's when the flight controller is trying to fight the oscillation.

>>I have heard that getting the 450 to descend is one of it's traits. Maybe there is an OpenPilot boffin who has come across this problem?<<

As I mentioned before the vertically take off and landing machines has that trait. Open Pilot Copter Control board is no different.

>>Anyone who has managed to get to the end of this message without deriding me completely would care to have a stab at answering?<<

Ok my take on this. The Copter Control board is the only flight controller board that if you tune it properly and if you have done your homework and put your machine together properly WILL NOT flip your machine over for no reason. OK I think my KK board has never flipped over unintentionally also except when I had a motor spinning backwards.

So you simply has to accept the descending wobble. When I shoot video I don't expect to use any of the footage during a descend. I always plan my shoot around an ascending period.

This is one of the few videos I have that show a descending wobble on purpose.



The trick is to fan the throttle and come down in stages. Drop the throttle a little too much let the machine fall then before it start to wobble add a little throttle to arrest the fall and smooth things out. I think there is a very short clip in that video where there were a bit of wobble when I was descending a bit too fast right around tree height. Then you can see the video smoothed out when I increase throttle.

The trick is NOT to let the machine gets so wobbly that it gets into an unstable oscillation situation.

Now on certain flight controller that can happen and I don't use those flight controllers very much anymore. I cannot afford to have my camera on the machine and it starts to tumble out of the sky,

Another thing you can do is to always keep a slow forward flight and fly the machine down like an airplane. Don't try to descend vertically.

Of course the bottom line is get an Open Pilot Copter Control board as soon as the next batch is available and learn to tune it properly.

[JohnyD]

>>Of course the bottom line is get an Open Pilot Copter Control board as soon as the next batch is available and learn to tune it properly<<

Best bit of advice you'll get on the "Descent Wobbles" issue!!  The CC has been the best board I've used all around, and when tuned properly with your machine- greatly REDUCES but does not YET eliminate the Dwobbles-- And has never flipped my quad, unless told to do so.

[Stu M]

Firstly, I'll try not to make this sound like an Oscar acceptance speech ... but thank you all for your enthusiastic and intelligent replies. I can see that I am not dealing with 'weekend pilots/designers'.

rigebackred: Yes, I am descending VERY slowly but I will certainly try a set of 8" blades and see what change in stability I can get. An excellent idea but will an set of 8" blades function normally on a 450-size machine?

ANV: I have tried increasing the weight of the machine by between 200 and 600g. Much the same result I'm afraid - flip-floppy oscillations on the way down.

peabody124: You saw my reply to rigdebackred ...

SeismicCWave: I bought the machine from BuzzFlyer in the UK. At the time, the 450 was under the heading: UAV/GAUI. The controller has a BuzzFlyer sticker on it so, like the radios, it was maker X who sold it to BuzzFlyer who then re-branded it. More than likely it's a hybrid. Their website has now changed somewhat as the 450 is now under the heading '450X Quadcopter'.

JohnyD: err ... yes.

____________________

Just to put the cherry on the cake, I went out for a flight late yesterday afternoon. It was superb weather. At no more than 2m from the deck, she shot off into the distance, turned around, flew back, did a pirouette and came in to land.

Flip - tumble - smash. From 2m up. Over she went and, yes, another pair of arms had to be replaced. That leaves me with one good pair of arms left in my (sic) armoury ...

2 pairs of arms replaced in 1 day. This is not good as replacement parts for this machine are getting difficult to source.

So, unless I can figure out how to repair the broken arms (I can stick a few images here if you would kindly keep an eye on this posting), it may be a case of changing the Quad entirely. You will see (subject to when I can get the photos up) that the 4 locating slots that hold the motor plate in position are as thin as angel's hair - hence the limited life span.

Once again, many many thanks for taking the time to reply to my pathetic little problem. I am very pleased to have found a group who are not only passionate about UAVs but eloquent in their language.

I shall put some pictures/videos up soon, please keep an eye open.

Stu

[juz]

I reckon its your GU344 and/or escs. Run some a turnigy plush (equivalent to the gaui ones) escs, they are cheap and you'll have them for future builds/projects. I have never liked the gaui escs. Keep playing/tuning the gu344. You might also want to get a frame that is tough and/or cheap and has good parts support (VC-450 or BQ-2).

The "trick" for me, to achieve a vertical descent is initially release a very large amount of throttle (for a second), to start a quite agressive/fast descent, the quad starts to drop, then add some power, but not enought to arrest the descent. Put together smoothly, I end up with a "smoothish" ride down.  The extra throttle seems to give the gyros/escs/motors extra authority. OR..just dive man.

I don't like to try and descend vertically, its usually not a pretty move with a quad.

this video ; gu344/gaui motors/hobbyking SS 18a escs

Edited by juz, 26 September 2011 - 07:55 AM.

[SeismicCWave]

>>greatly REDUCES but does not YET eliminate the Dwobbles-- And has never flipped my quad, unless told to do so. <<

Hmmmm........Dwobble = death wobbles?

For sometime I actually had a hard time flipping a Copter Control board. I finally learned to crank up the manual stick rate so I can flip.

[SeismicCWave]

>>An excellent idea but will an set of 8" blades function normally on a 450-size machine?<<

It doesn't hurt to try. You may find that the machine will hover at a higher throttle stick position.

>>I have tried increasing the weight of the machine by between 200 and 600g. Much the same result I'm afraid - flip-floppy oscillations on the way down.<<

Increasing weight is never a good idea on any flying machine.

>>SeismicCWave: I bought the machine from BuzzFlyer in the UK. At the time, the 450 was under the heading: UAV/GAUI. The controller has a BuzzFlyer sticker on it so, like the radios, it was maker X who sold it to BuzzFlyer who then re-branded it. More than likely it's a hybrid. Their website has now changed somewhat as the 450 is now under the heading '450X Quadcopter'.<<

In that case some photographs of your machine and setup will definitely help. At least we know what we are dealing with.

>>At no more than 2m from the deck, she shot off into the distance, turned around, flew back, did a pirouette and came in to land.<<

Was that all voluntary or involuntary?

>>Flip - tumble - smash. From 2m up. Over she went and, yes, another pair of arms had to be replaced. That leaves me with one good pair of arms left in my (sic) armoury ...<<

Well sound like you have more issues than simple wobbling quad on descend. Even the Gaui 330X with stock GU344 will not do that.

>>I shall put some pictures/videos up soon, please keep an eye open.<<

Thank you. That will help. We can advice you to ditch certain items and keep certain ones.

[SeismicCWave]

>>I reckon its your GU344 and/or escs.<<

I have a feeling that he may not have a GU344.

>>Keep playing/tuning the gu344.<<

There isn't much to tune on the GU344 except for the gain.

If indeed he is using a GU344, here are some tips I have found that makes it fly better:

1) Do not use hover mode. Always hover and fly in cruise mode.
2) Tune gain to around 50%.
3) After calibrating the speed controllers drop the throttle high end travel to 80%.
4) Drop the aileron and elevator channel travel end point to 80% also.

>>You might also want to get a frame that is tough and/or cheap and has good parts support (VC-450 or BQ-2).<<

I actually like Rusty's frame better for crashing because you can replace the arms. I crashed and broke two arms on my VC-450 frame. I glued the frame together with CA glue for rigidity and cannot replace the broken arms without damaging other parts. So I just threw the frame away.

[juz]

Sure, each to their own mate. But I have crashed the VC-450 many times without damage. I'd put money on Tom's frame being tougher than Rustys (especially the arms), I've crash tested both. Rusty frames of the same size may be lighter though (not sure). Also, you can replace VC-450 arms just as easily as on a Rusty frame, they are available at hoverthings.com. The frame does not need CA. The instructions don't mention CA. I haven't heard of anyone having to use CA. Silly rabbit for using CA. I have many hours of accurate FPV flight with one frame with no CA and its still rock solid. Need to change an arm?? 10 minute job!

BTW I also have two Rusty frames.. MQ-1 and UAP1.. and I like 'em too.

[ANV]

I'm almost out of ideas , but just a thought(paranormal a bit)
One more thing is to verify all electrical joints and connections. Sometimes wires only look like they soldered, but in fact they only touching each other just a little bit. This goes to the next guess: if your frame is not strong enough (ie first mechanical fundamental frequency within motor RPMs) you're likely to get resonance even with well-balanced props/motors and this can make controller to go mad because of vibrations OR bad soldering takes its role.
Another guess... At the descend some of the elements (battery???) "looses" weight a bit and pushes (releases) bad soldering point making it loose connection.
Try to pull wires that soldered togeter (with a noticable force). Don't be afraid to break it - if it soldered good it won't break.

BTW, while motors spin-up do you see if there is something wrong? That is because you pass through the same "bad-throttle" point.

[Stu M]

Once again, thank you for your interest. Let's start with some answers then;

juz: "The "trick" for me, to achieve a vertical descent is initially release a very large amount of throttle (for a second), to start a quite aggressive/fast descent" - not bad. I can always give that a go.

ANVL: "electrical joints and connections" again, will give it a go and see if we have any loose bits 'n bobs.
"while motors spin-up do you see if there is something wrong?" other than the odd blade doesn't turn straight away because I'm not on a level surface, no. Fine.

SeismicCWave:
"I have a feeling that he may not have a GU344." You have a bad feeling. He DOES have a GU344.
I only ONCE tried to fly the Quad in 'Hover' mode. The thing hardly left the deck before flipping over. I have always (apart from that one occasion) flown in 'Cruise' mode.
I will order a set of 8" blades now. You're right, it's worth a go.
"Increasing weight is never a good idea on any flying machine." Oh, righto.
"In that case some photographs of your machine and setup will definitely help. At least we know what we are dealing with." I will see what I can dig up. I stupidly didn't take any photographs yesterday and now I am 130 kms from the machine and won't be going back to it until the 1 week of October.
"Was that all voluntary or involuntary?" Oh, totally voluntary. I don't have any GPS kit on-board and wanted to convince myself that she could fly, fly, fly …! Getting her to land, land, land is why I'm here!
"Well sound like you have more issues than simple wobbling quad on descend. Even the Gaui 330X with stock GU344 will not do that". OK, well let's see what we can do …

Check back soon for some photos …

Stu

[Snagglesworth]

If you have the 344 on your quad do you have the green wire plugged into channel 5? I found with my gaui that it would have a few dodgy moments when the green wire was plugged in. Unplug it, set the micro switch on the top of the 344 to cruise mode and turn the gain dial to about 50%. That sorted my gaui out and turned it into a stable machine. Oh, but I still get slight wobbles on desent but nowhere near enough to flip it.

Edited by Snagglesworth, 26 September 2011 - 01:27 PM.

[Stu M]

Snagglesworth:
"If you have the 344 on your quad do you have the green wire plugged into channel 5?" I'll check but in 2 weeks ...
"set the micro switch on the top of the 344 to cruise mode" It always has been.

I'll collate all the answers and once I'm back in the country, I'll go through everything logically and step by step.

Stu

[Stu M]

I have left the sound up so you can hear the motor pitch change.
I hasten to add that this is NOT due to any input manipulation on my behalf.
As you can see, not a breath of air was present.



And here's one from August of this year. Another reason to change a pair of arms:

Edited by Stu McAlister, 26 September 2011 - 02:16 PM.

[Spydmobile]

Hi Stu,

This has been a problem me since my first flight, I have learned a lot about it since, first, I have recently changed to triblade props, for example, I have quad I was building RD-4
http://lab.westernwi.../model/fleet/14
Running 12 props, and ANY decent would cause wobble, I switched to 9x7 tri blade props, and presto instance dramatic change to stability, my flight cant possibly last as long, but it is more stble, this supports other advice about smaller props, next when you add weight WHERE you add it can amplify the oscillations too, if you oput the weight immediately under CofG it will tend to stabilize but if it is too low under the prop plane, it will cause the pendulum efect and make the oscillations worse. So be wary of just adding weight down low, I have had succes with adding a second identical battery and running it in parallel, by adding a parallel lipo and switching to tri blade props I have made some of my wobbly machine stand rock solid in gisting winds.
Cheers
Franco

[ridgebackred]

My guess from watching the video's:

1) The props are too large for the power rating of the motors or esc's.
   An American muscle car saying: "There is no substitute for cubic inches."   (More horsepower!)

2) See 1)