Vibration, frustration and crash with 450mm

Hi there,
First of all, while I’m not new with GPS quadcopters, this is my first quad that is running Arducopter on a Pixhawk clone. So a few weeks ago I finished building this 450mm quad and installed balanced 10" props, upgraded the firmware to 4.0.3 and did a few test flights. While it did flew somewhat decent, I had issues with oscillations and vibrations. Mainly in the Yaw axis.

Then the frustration began. For weeks i tried to tune the filters and PIDs, did AUTOTRIM and AUTOTUNE but i had the feeling, that it was flying worse and worse with everything I did.

Just right now the “crash” happened . When I wanted to take off in alt hold mode, the quad just tipped over and shredded 3 out of 4 props and the gimbal to bits, severed some cables and also probably damaged the Lipo. Disarm didn’t work and I had to disconnect the battery, so i dont have any logs, but the Pixhawk probably wouldn’t have saved anything anyway because I didn’t left the ground.

As of right now I’m close to disassembling the quad and dropping the project. I only had one spare pair of 10" props so I guess I won’t be flying this quad for the next month anyway. But I still want to know what the hell happened, why it just tipped over, and how I can prevent it the next time.2020-04-09 17-12-10.tlog (765.2 KB) parameter backup.param (17.9 KB)

Can anyone just take a quick look at the logs. Maybe there is an easy to see mayor flaw in my parameters?
regards Tav

Edit: Seems like i can’t upload all my logs. Because of size?

That .bin log is corrupt. Do you have another one? The Tlog’s are of little value, don’t bother with those.

450 size quads usually will fly OK on default parameters if the vibrations levels are not high. Dshot600 isn’t a wise choice and on a quad that size it will do nothing for you. Dshot150 if you get it flying again.

Well, one thing I can see from the Tlog is it looks like you had the motors connected to the Pixhawk main outputs and were configured for Dshot. You won’t get Dshot form those outputs.

Before doing any type of autotune you should follow the tunning instructions

Your parameters look OK, only I’d change definitely change:
And maybe these:
ATC_RAT_YAW_FLTE,3.868905 to about 2 or 2.5, lower in stages

It would be good to see that .bin file. Have you got an earlier .bin file?
Sometimes if you power up the flight controller it might close off that last .bin file and it will become usable - here’s hoping.

These are some values from a quad 450 with 9inch props, 1175 KV motors and 3 cell battery. It’s a very stiff frame too. Be aware every frame/motor/prop/ESC combination can give wildly different PID’s of course.

Hi Dave,
Do you think Dshot600 is overkill or works the ESCs and motors too hard for larger quads? I know on smaller quads faster is better.
I’m interested in your opinions and findings on Dshot - I plan to start using it but haven’t had a chance yet.

Hi Shawn,
I have tested Dshot on a 3", a 5" and a~500 size with 13" props and low kV motors. I’m no Racer so maybe I don’t have a good feel for this but I notice no difference on the 5" between Dshot150, Dshot300 or Dshot600. Dshot1200 did funny things on the bench, never tried it flying. I think I do notice a difference between PWM and any Dhot. On the 500 I don’t notice any difference in any protocol from PWM, Oneshot, Oneshot125, Dshot150 or Dshot300. Not in flying performance and if the motors are getting hotter with any of them I haven’t noticed that either. And I do tend to fly this quad around aggressively, for a 500 anyway. The 3" was only flown with Dshot300 when it was flying.

If it makes any difference the 5" and 500 have PixRacers, the 3" had a Omni F4 Nano.

Here’s to flying a 500 aggressively!
Thanks for your input.

This is our 450, it’s had some more tuning since then, and even been set up for Acro. It’s done it’s fair share of rolls, flips and loops all on good old PWM.

Sorry to derail the thread

Hi there,
Thanks for your answers.


Do you have another one?

Sadly no. It seems like Pixhawk only creates a log file if you are flying for at least 1 min or so. I have a log file from monday but I tried to tune the filters since then (which probably made the flight characteristics worse)
So as long as I am using the main outputs, it doesn’t matter if I change the protocoll to Dshot150 or oneshot?


Before doing any type of autotune you should follow the tunning instructions

I did that. It made things worse.

I will keep the parameters in mind. I have 2212 kv980 Motors on this one and use a 4 cell Lipo (wanted to change it with Li-ion after I tuned the quad). I do have a set of 9" props left, but I never tried them out. I would like it, if it stays that way cause they seem to be made from a rubber like material. unbreakable silicone props

Another thing I should add: during the last flight (28.bin) it seems like the quad was resonating during 40% trottle. I don’t thing this is the reason it flipped over after that, but I think this isn’t the first time the quad did that. Just stronger this time.

That’s what I’m talking about :grinning:

I don’t know what the deal is with your logs but that one is problematic also. Sd card maybe.

Right, on a Pixhawk you need to use the Aux outputs for Dshot and configure the servo out mapping accordingly. Chan 9-12 instead of 1-4 on a quad.

A 450 size with 9-10" props and those motors is pretty much the reference for the default parameters. By all means use the Tuning parameters advised and as Shawn has noted but it should fly on defaults. I have built many a 450 that has. Is it one of those cheap F450 plastic frames? How is the Flight Controller mounted?

It is a no name SD card,which I didn’t pay much attention to until now. So I did test it now by copying random data to it. Seems fine but I can try another one when the quad is flyable again.

Well the frame is out of plastic, but everything except the landing gear seems sturdy enough. Of course not as sturdy as carbon fiber or aluminium, but i try to stay under 2kg. The flight controller is mounted with these silicone stoppers that came with the kit.

How do you analyze the log files? INav can overlay the video feed with the log data, so vibrations in the gyro log data are visible in the video feed in real time.

If you use the main outputs, you’ll be using PWM and you’ll need to make sure your ESC calibration is done. Even if you have Dhsot set, Arducopter knows it cant do it on those outputs and does PWM instead.
As Dave says, to use Dhsot move to the AUX outputs.

Here’s some values from a 520 with 910KV and 11inch props, 3 cell.
You can see they’re in the same ballpark as the 450/1175kv/9inch

For 4 cell LiPo you’d want to set these:
And it’s best to do the Compass/Motor calibration, although not really related to you current issues. You’ll NEED to adjust these and redo the compass/motor calibration if you change to Li-ion because they have a very different magnetic field to Lipos.
You cant just swap between the battery types without redoing some calibrations unfortunately.

That FC mount is OK.

I don’t see any practical use for iNav overlay. Unscaled squiggly lines crossing the screen. Cool :smirk:

Mission Planner and APM Planner 2 have good analysis tools.