Copter: Autotune and big props = problem?

[Stefan Gofferje]

Hi,

is anybody aware of potential problems with autotune and big props?

I noticed the following:

When I flew my new copter with 15" "T-Motor style" props, it was very
stable. Video from the onboard camera was almost flawless. Then I
changed to 17" Multirotor props and the copter was shaking a lot,
something like around 5-10Hz intermittently. It's easily visible in the
videos. I just ran autotune and the general behavior is much more stable
(yesterday, it almost fell out of the sky when hit by gusts) but it's
still shaking.

Of course the props are all balanced and the interesting thing is that
the IMU logs show vibrations well within boundaries (+/- 1.8 around the
baseline in hover), so I was suspecting that the shaking comes from bad
tuning. Autotune did some clear changes to the params but - as I wrote -
the copter still shakes.

Now the question is why :). I could of course just mount the 15" T-Motor
style props and be happy with it but with the 17" multirotor props my
copter pulls about 1A less plus I always wanna understand what's the
issue, especially if it's something weird like in this case.

Photos of the frame (with the 15" props) and selected logfiles are at
http://home.gofferje.net/copter/)

-S

--
(o_ Stefan Gofferje | SCLT, MCP, CCSA
//\ Reg'd Linux User #247167 | VCP #2263
V_/_ Heckler & Koch - the original point and click interface

[Jesus Alvarez]

Could it be related to a "way too big propeller" (inertia) for your given motor?

Maybe the motors are not able to track the autopilot demanded motor speeds without introducing lag 

[Stefan Gofferje]

On 06/15/2014 08:05 PM, Jesus Alvarez wrote:
> Could it be related to a "way too big propeller" (inertia) for your
> given motor?

Interesting idea! But the prop/motor combination is pretty proven. It's
RCTimer 5010-14 360kV. Pretty many people who build long endurance
copters use this combo. At least according to the current, I'm running
the motors well within specs (~3A/motor vs. 10A max rating).

But yeah, I do see some discrepancy between desired roll/pitch and
actual roll/pitch even after autotune.

Could it be that the ESC can't keep up with the commands? It's a Q-Brain
/ Hobbywing 4-in-1 4x20A ESC.
I don't see any sync losses, though. It's just obvious that the
controller is struggling to get the copter to do what it wants. It also
sounds pretty aggressive.

[Holger Steinhaus]

Hi Stefan,

just in case you are flying the narrow-tip RCTimer 17x5.5 props: these props are crap - they have different pitch angles on both blades. Even if they are perfectly balanced, they produce lots of vibration due to the different angle of attack and therefore the lift produced. The problem gets worse the more power you put into this prop. Below 500g of thrust per prop might be marginally ok, as the vibration is not to bad yet here.

I did run autotune several times with 16" props. The resulting PID values are ok, but far the ones I got by manual tuning. I usually end up with more rate D, less rate P and much less stab P when tuning manually.

Holger

[Stefan Gofferje]

Hi,

On 06/16/2014 09:56 AM, Holger Steinhaus wrote:
> just in case you are flying the narrow-tip RCTimer 17x5.5 props: these
> props are crap - they have different pitch angles on both blades. Even
> if they are perfectly balanced, they produce lots of vibration due to
> the different angle of attack and therefore the lift produced. The
> problem gets worse the more power you put into this prop. Below 500g of
> thrust per prop might be marginally ok, as the vibration is not to bad
> yet here.

That's the props, yeah. I just bought mine from the eBay-seller
happychoice2009 which seems to be a manufacturer. Anyways, I just
checked my props very thoroughly after reading your mail and they seem
totally symmetrical to me.

> I did run autotune several times with 16" props. The resulting PID
> values are ok, but far the ones I got by manual tuning. I usually end up
> with more rate D, less rate P and much less stab P when tuning manually.

That's a good point, I'll try that. However, when flying the 15" props
again yesterday, I noticed that also with those, the FC seems to be
fighting and the copter was shaking a bit - much less, though, primarily
in wind. The video I posted recently on DIYD is from the copter
perspective with the 17" props is gusty wind.
I also copied a video with the 15" props now to my home server copter
directory.
In the latter, you can pretty clear the FC fight.

I also noticed that when I ran autotune with the 17" props, it took much
longer and the overall movements were less "sharp". With the 15" props,
the movements in autotune were very "sharp" and it was done in about 2.5
minutes.

Which brings me back to wondering if the ESC is maybe too slow or
otherwise unsuitable for big props. I didn't see any sync issues, though.

[Andrew Chapman]

Hi Stefan,

There is/was an issue with autotune on multi-rotors spinning large slow props, in that they might not be able to achieve the angular rates commanded by autotune. I'm just catching up after a bit of a hiatus so I'm not sure if it has been improved in the code since, but Leonard had this to say a little while back:

On 2014-03-20, 2:25 PM, Leonard Hall wrote:

Hi Andrew,
Sorry to hear about your crash. This post has just come to my attention and I am finding your autotune log very interesting.
I have been aware that it may be possible for autotune to fail like this but I have not seen it until now.
Basically the problem is the copter accelerates so slowly that it can't achieve the desired rate of 90 degrees per second and finish it's rate response before hitting the 20 degree limit.
I am building a hex with 18 inch props and will be able to play with this myself before too long. However, until then I could use some help from you if you are willing.
Would you be willing to do an auto tune with some custom code? I would reduce the rate request to 60 degrees per second to see if your copter is able to achieve a proper autotune with that.
Thanks for bringing this to our attention!!!
Leonard

AC.

[Craig Elder]

Holger if you could post your logs from an autotune with a vehicle with large props that would be helpful

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[jdennings]

Just my experience, but I've auto-tuned many times an X8 with 17" props (CF paddle type, 17x5.5) without no problem.

[Stefan Gofferje]

On 06/17/2014 04:58 AM, Craig Elder wrote:
> Holger if you could post your logs from an autotune with a vehicle with
> large props that would be helpful

My logs include autotune cycles with 15 adn 17" props on 360kV motors
:). I just flashed my ESC with BLHeli. As soon as the weather gets
better, I*m going to try with the big props again.

-S

[Holger Steinhaus]

Yes, will give it a try if there is suitable weather tomorrow.

Holger

[chandruth]

Hi

I fly a 830mm Quadcopter with RCtimer 5010 360kV motors, RCtimer 17x5.5(paddle type) Carbon fibre props, RCtimer NFS 45A ESC SimonK firmware, Pixhawk with External Compass+GPS. I also run 2 3S 5000mAh batteries in series to get 6S 5000mAh. I installed a fresh copy of Arducopter 3.1.5 and autotuned from the stock Pixhawk gains. Just to note, with the stock gains the copter was flyable. But after the autotune the copter feels over responsive. I have another smaller 550mm sized quad, when that autotuned it felt very stable and smooth. I have attached the log and some pictures of the quad for you guys to see. I hope this data helps you find the problem and a solution to the large propeller autotune issues. If you need me to try anything for you guys I am more than willing.

Regards

Chandruth

[Jason Short]

There is the RC feel option which should help. 

Btw what are your new gains?

Also I've been thinking of building something similar. What is the current draw on your motors at hover? And what prompted you to chose such a high current esc?

Jason

Sent from my iPhone

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

<Large Quad data.zip>

[Holger Steinhaus]

https://www.dropbox.com/s/729483vo32e4z5a/oh2q-20140619-59-autotune.BIN

Here the promised autotune log. The results were quite good this time, probably the change to faster motors (470 rpm/V now instead of 360) made up for the difference. 

The values found by manual tuning were set before starting autotune. 

Airframe: 16" quad (2.1kg, 4x Tiger MN3110-26, 4s-12500). The copter is quite rigid as is designed to fly as X8 @ 5kg. The large battery (1kg, 10cm below the frame) and a large landing gear probably move the COG down quite a bit.

Holger

On Tuesday, June 17, 2014 3:58:23 AM UTC+2, Craig@3DR wrote:

[Holger Steinhaus]

 

That's the props, yeah. I just bought mine from the eBay-seller
happychoice2009 which seems to be a manufacturer. Anyways, I just
checked my props very thoroughly after reading your mail and they seem
totally symmetrical to me.

May be there are different qualities around. Mine are symmetrical in shape as well, but the sanding of the leading and trailing edges of the blades is completely different. This leads to a different airfoil shape and different effective angle of  attack. One of these props almost destroyed my test bench, causing a really bad resonance effect (of course after static balnacing).

Holger

[Stefan Gofferje]

On 06/18/2014 11:54 PM, Holger Steinhaus wrote:
> May be there are different qualities around. Mine are symmetrical in
> shape as well, but the sanding of the leading and trailing edges of the
> blades is completely different. This leads to a different airfoil shape
> and different effective angle of attack. One of these props almost
> destroyed my test bench, causing a really bad resonance effect (of
> course after static balnacing).

That sounds like a bad batch. Mine are pretty ok. Not perfect but good
enough. But happychoice has the T-Motor style also in 16", 17" and 18" -
even 20". I was thinking of getting the 16 and 17" for testing. That
blade shape IS much more silent and if the 17" are similar in efficiency
to the triangles, I'm happy.

-S

[chandruth]

I don't really like using RC feel. So I just leave it at 100. I don't have current sensor with me to check the amp draw yet. But the copter flies for 27minutes with 2 3S 5000mAh batteries in series to get 6S 5000mAh. The weight is about 2.7kg with batteries. As for the choice of ESC's, I wanted it to have SimonK firmware and 6S support. I went for these because of that and the 16MHz crystal to prevent temperature drift. I live in Singapore and it gets very hot here. Also the 30A and the 45A versions had very little price difference and weight so went for the 45A. I have attached a screenshot of the gains. It really is too high I feel. Also another thing i noticed during the autotune, initially the copter would roll/pitch to the normal angle during autotune. But when it gets closer to the end of tuning for each axis. The angle is significantly reduced (maybe 10-15degrees of roll/pitch only). 

Regards

Chandruth

[Stefan Gofferje]

I did another test today. A few days ago I flashed my ESC with BLHeli.
Today the weather allowed for a new test. I ran another autotune in low
winds with 17" props and just visually, the behavior was totally
different. The copter moved much "sharper" in autotune but also overshot
and oscillated quite a bit.
After autotune it flew much nicer than before with the stock ESC FW
after autotune but it still felt very soft. The low frequency vibrations
are almost gone now.

Log file here:
http://home.gofferje.net/copter/2014-06-19/

Video is currently uploading to Youtube.
Link follows.

[Stefan Gofferje]

On 06/19/2014 05:29 PM, Stefan Gofferje wrote:
> Log file here:
> http://home.gofferje.net/copter/2014-06-19/

Video: http://youtu.be/W199Knx7XVE