I've tried everything on here but it still wobbles and flips on takeoff.

Hi all,

So this is the first time I've sought help on here since I got my first APM (2.5 latest firmware).  I've managed to make it this far with all the good info here and in the wiki.  Here's my kit, followed by my problem.

Stock Gaui 330X-S

Spektrum 2.4Ghz Rx and DX6i Tx

So, I had the above taking off and hovering no problem with the stock ESC's (see here for make up of parts)

The ESC's that come with the quad are reputed as being of low quality so I decided to replace them with something a bit better to see if I could make it fly a bit smoother.  These replacement ESC's come widely recommended, so I got some.

Cut a long story short, I cut all the connectors off my old ESC's, soldered them all to the new ESC's and rebuilt my quad with the new ESC's.  Eventually figured out how to calibrate them (which I did manually) and I'm confident this has been done correctly (as I did it several times before locating the instructions and doing it properly).

My problem is, quad wobbles and flips immediately after leaving the ground.  I've ran through all the troubleshooting steps in the wiki numerous times yet I'm still experiencing the same problem.  The only thing I can't get working in the test is the motor CLI test:

Have you run the Motors command in the CLI Setup to ensure that all the motors are hooked up right, turning the right way and the pusher/puller props are on the right motors and oriented the right way?

The motor CLI test simply does not work for me, it just doesn't spin the motors as per instructions.  I have no idea why and I promise you I have quadruple checked that I have set everything up correctly.

  • All sensors tested and working
  • Radio Calibrated - all end points on 100%
  • Quad levelled
  • Props and motor in correct orientation, I can see which way they're going when they start/stop spinning from full speed
  • APM orientation fine (using 'x')

Here's the weirdest thing.  If I give the quad even just a little throttle then take my hands completely off the Tx, it'll sit there on the ground, all props spinning but they will very slowly speed up to the point where the quad will take off on it's own if I don't drop the throttle.

Another test I did was to hold the quad while giving it a little throttle.  It seems to work exactly how it should when I do this in that it will correctly compensate if I tilt it in a given direction.  If I only lightly held it, it tended to start wobbling.

I've tried everything that seems to be listed on the forums/wiki but am at a loss.  Could it be that I need to do some tuning of PID's before I can even get it to take off properly?  I didn't need to touch a

thing with the stock ESC's to get it hovering ok.

Please see my quick video of it increasing the throttle by itself and the flipping.

Flip Vid

Sorry about the "War and Peace" entry.  Any suggestions/questions are more than welcome.


Views: 7026

Reply to This

Replies to This Discussion

I don't know your radio, and it's really hard to see from the picture, but it looks to me like you have channel 1 going to input 3, channel 2 going to input1, Channel 3 going to input 2. 

With a mode 2 radio, I thought it was pretty straight one to one input. Depends on your radio though.

Hi, I'm using a Spektrum DX6i (Mode 1 - I'm from down under).

I'm getting all the right reactions from the radio and quad and I've triple checked my connections.  I've also tested it by holding it in my hand.


My problem is, quad wobbles and flips immediately after leaving the ground.



Lower pid's....


This is common with high pid's. Drop them till it stops. Should fix it.







Taking a stab in the dark here - This is exactly what my quad does when the battery gets too low. The ESCs are designed to shut off at some (programmable) voltage to prevent over draining your LiPo. In airplanes it's a safety feature but in quads, because one of the ESCs will inevitably shut off before the others, it causes flips.

It's possible you're seeing this manifest in your new ESCs and not the old ones because they have different cut off voltages. If you read the ESC's calibration procedure you can usually set how sensitive to low voltage they are. I keep mine as insensitive as possible.

It makes since that this will happen shortly after takeoff. As you increase the current drawn from the battery the voltage will drop. Depending on your battery it can be anywhere from 0.2 to 3 volts. It may be informative to hook a multi-meter up to your battery and see what the voltage does as you spin up. If it drops much (more than about half a volt immediately following spinning up), it's probably time for a new LiPo.

An easy way to test this is to plug a motor and ESC directly into your receiver. Spin it up for a bit and see if it cuts out on you. Remember, because it's one motor drawing current and not four it may take a bit longer.

I don't think this is a problem with proportional gains being too high. The flip was incredibly sudden. Flips caused by dynamic instabilities are usually oscillatory, not purely divergent. That is, it will wiggle and the wiggles will keep getting larger and larger until it eventually flips, not flip instantly.

Best of luck, I know it's a real pain right now but once you get things working it'll all be worth it


Just watched the video a dozen more times. I'm pretty sure you can disregard what I wrote above. I hadn't seen the high frequency oscillations before. Definitely drop your proportional gains, as Ken suggested. Try dropping them in acro first, then stable mode.  Maybe by as much as a factor of two. Try balancing your propellers as well. You'd be amazed what high frequency shakes can cause an autopilot to do.

@ Nate, reprogram your esc's for nicad's. That ends the shut off problem and will allow you to land with out a crash. Just pay attention to flight time to prevent lipo damage for overdraw.

@Ben, ESC is not your issue. Drop them as Nate says. Your copter is over correcting, hence the violent wobble your seeing. BTW, don't fly in your house, you have no idea how danger's it is. The copter is not to be trusted even when working correctly. Its a Ginsu knife on steroids, just ask around the forums and people will tell you their horror stories.


Hi Team,

Thank you for all your suggestions today, it's been a great help.  First of all, I set up my cables as you suggested Dan (can't believe I didn't think of this!).

I have been playing around with the PID's heaps.  Last night I was having some success but then all of a sudden with the same settings (or so I thought) it started going nuts and flipping again.

Then today, I started having success again.  Then same thing, randomly went nuts on a test flight.  However this time, the penny dropped with what I was doing.  Instead of typing in 0.090 for my rate roll P, I was typing 0.900 (for example).  Now I've realised my mistake, I won't do it again and I double check the settings always.

This WAS NOT the original problem though, just a side one that I was causing.

So anyway, I've just had a really successful test flight (inside - sorry Kenneth, I live in the city and it's a long walk to somewhere I can test properly). :)

My rate roll P is at 0.050 which is significantly different from the default of 0.145 but it's hovering extremely smoothly and seems to be reacting to my controls OK.  I'm aiming a more for smooth FPV/AP type application so this is good.

I have one question though.  I noticed on the wiki it mentions you should use a lower PID for bigger motors and vise versa.  My motors are Scorpion 1050kV (with the above mentioned ESCs), I'm not sure whether this is big or small.  Does that PID seem like a strange one even though I am getting the result I want? (I haven't completely zeroed in yet but it'll be between 0.050-0.060).  Also, I haven't changed the stabilize roll P at all yet.

I am still going to play with the PID's to fine tune and learn a bit more but I've solved my immediate issues for now it would seem.

Also, I've just hooked the 3DR radio up to it.  They are so good and easy to use.  They just work!

Thanks for much for all your help, it is much appreciated.  I live to fight another day in this hobby!

You should be changing those numbers to be the same on roll and pitch. They are locked together by default. Not moving them together is where you got the odd behavior.

You may need to start over as you move both because the number you have now will not be the same when tuning them as one.

No the pid does not seems strange at all and you will want a less hyper copter for FPV/AP any way.

Ignore the wiki on motor size, use what works for you as that's just a ruff guide line and I mean Ruff on motor size.

I am glad you are closing in on a working copter. I will be away this week end but I am sure other will help.


The Wiki says theres some debate about leaving all of a linear ESC power lines hooked up.....But those ESC's don't fall into that category. You may want to spring $10 on a UBEC to power your copter. I ran all 4 of mine hooked up for a bit and didn't get the wobble so maybe there is some noise in your system. A quality BEC will do much better then those ESC's your using to power the APM.  Oh, and I cut 3 of my wires before realizing I could just have unclipped them....DOH!

But if you want to improve your ESC's.....

I would consider Ordering  that 19.99 Programmer

Seems to have allot of people flying much smoother.

as well as the avr so you can hook it up to your computer and flash those ESC's with simonk firmware. lots of People are reporting that will give you the best ESC for a quad. Although personally when buying an ESC I would either get the 5.99 non programmable version of those for 5.99 (and flash them with the simonk) or pay $10 for the blueseries..AKA mystery ESC's. of course people are even flashing those..... but that flash will get them setup for multicopter.  Almost every ESC out there was built with planes in mind.

And that will flash most ESC's, So far Ive heard the plush, the SS, and Mystery are all good to go. Some trouble with the red brick though.

. Oh yah, you can use it to program your turnigy 9x as well!

I'll stand on the other side of the fence from Will. no offence..... My secret is to keep changing things until it works.

I used 4 of these http:// http://www.hobbypartz.com/07e-flyfun-30a.html

Never flashed em, left all 4 power wires hooked up. My theory is that I don't want all the 5v load coming from one esc and heating it up. Also, they only supply 2amps, which is alot, but I got video tx, telemetry radio, lights and so on. I'd rather not burn up an esc by overloading the 5v line.

They take the cool programming card so no funky stick work to set them up.

My apm2 is on sorbuthane on a carbon deck, and ALL the wiring is below that deck. I did because I thought it looked cleaner, but I found out cf is conductive, so that probably helps keep the rf out of the apm.. 

Standing on the other side is good. It's how Intelligent people further an Idea.

On my system I was chasing a different problem (and I had basically the same ESC's) I did not see any problems with all of them hooked up, When I unhooked them I saw no change however from what Ive gleaned from the wiki and other sources, your only drawing power from one of those hooked up. (and in any case your only going to get the 2 amps, or in his case 3) Now before in the Wiki there was note of some debate, whether or not to hook up multiple sources (like one from each ESC'c built in BEC) but in the case of a turnigy plush, which uses a linear BEC there was some advantage.

Now that is gone from the Wiki and it only mentions using power from the break out board from an arducopter kit.  Which I suppose is what you get when some people who buy the whole setup from diydrones, then insist that the instructions should fit there needs perfectly, and be less confusing....(at the expense of mentioning other methods that may work better for other setups) It is now of course less confusing for those running all there equipment.

Anyways I don't know enough to speculate on what could be the difference from using a linear to non, and having them all hooked up.. I still think your only going to get 3 amps. Also I asked Chris a question about this awhile back, noting I had a BEC and four ESC's, and he said I would better of to just use the BEC.

Oh and I believe the way the flash helps is to get the settings, and timing more in sync with what a quad needs.

Now I just read through the wiki, and Im going to take some qoutes from it that seem relevant to power issues causing problems.

If you experience spurious resets or other odd behavior it is most likely due to noisy or insufficient power to the APM. As with all logic boards, electrical noise from the motors, servos, or other high current devices on the power source can cause unpredictable behavior. It is recommend that a power filter such as this or this be used in such conditions.

Power source problems are common and can be insidious and frustrating. Be meticulous. Any autopilot or flight controller is useless and potentially dangerous without good clean power source.

Too short or long power wires, bad or old connectors, or insufficient current capability of the APM power source can result in a "brown-out" situation resulting in unpredictable operation. This is particularly true in traditional helicopters where the collective servos can draw 3-20 amps in short bursts. The power source must be able to accommodate this without voltage droop or voltage spikes. A quality switching type BEC such as one of these or one of these can be a solution depending on overall current requirements. Many of these type of regulators are programmable so remember to program them with in the safe operating range of the APM2. Linear voltage regulators are not recommended as they are inefficient and prone to overheating and heat induced failure

It's also possible to power APM 2 from two separate sources, one powering the RC system on the input side, and the other powering the output side (servos or ESCs). This is determined by a jumper on the JP1 pins (see below). If the jumper is on, which is the factory default, the board is powered from the Output rail. If the jumper is off, the board is powered from the Input rail, but the Output rail will need its own power source. This configuration is used if you want to have two seperate power sources in your aircraft, one powering the servos and the other powering the electronics.

Reply to Discussion



Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2020   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service