My name is Ben. I am building a Drone for forestry observation form my Bachelor thesis.
I have a problem with the copter flying in auto mode.
On the way from one waypoint to the next the copter flys extremly wobbly and has problems to hold the direction.
On the turning points the copter flys forwards and backwards and rolls to the sides till it has more or less the heading oft he next way point.
I don’t know witch parameters i have to tune now.
If i understand right what is written in the documentation, the „auto mode“ use the basic PID parameters, the alltitude controll from the „alt-holde mode“ and the position controll from the „loiter mode“?!
So i tuned the basic PID in „stabelize mode“ and with the help oft the log-files.
Then i set up the „alt-holde mode“ and then i tuned the „loiter mode“ by changing the Loiter PID parameter (extended tuning page in mission planner) and the WPNAV_LOIT paramters.
Now it’s flying well in all of these three flight modes. Also no problems in „pos-hold mode“.
But when i start a mission it seems that nothing is working. Evertime i have the overtake the copter in stabelize mode to avert a chrash.
Does anyone have a tip for me what i have to change?
The best way to understand the problem is by analyzing a flight log. Can you upload a .bin log of a good flight in PosHold and also a log of a bad Auto mode flight?
You can download the logs with Mission Planner or APM Planner, and upload them to an online storage like Google Drive or Dropbox and link it here.
A newbie should never build a craft with 29" props and 100kv motors.
That setup is very unstable and prone to getting knocked all over the sky by winds.
I am guessing you have a Quad, at 4.8kg the craft will act like a leaf.
The high moment of inertial of those props will make flying in auto mode unbearable to watch, as the motors/props cannot react quick enough to stabilize your craft.
My advice, scrap your build and rebuild it with 17-18" props, before someone gets hurt.
I’m surprised it’s still flying. The logs looks really bad. Note your desired roll and pitch aren’t following well and diverge at times. The tune is not good. I’d recommend an autotune, but you have some problems with your general setup you should address. As others have said, large props and low KV is difficult to tune…I know lol! Personally I run 28" props in an X8 config which is a little easier to tune than say a quad that large. I’ve tried 170kv at 6s with really poor results. 8s was a little better, still not good in wind. I’ve finally swapped the motors to TMotor U8II at 100kv, Alpha ESC’s and 12s. Rock solid now. You may want to consider going to 12s but not knowing the pitch of your 29" props, you may run into heat issues. Yes, flight time is great with these setups… And that’s about all unfortunately. OMG, I just realized your weight! You’ll need to add batteries. You need those props to spin fast. Even though I’m using an X8, my weight is 16kg. What is your battery setup?
Having designed professional multirotor systems for the past 10 years, we know how difficult it is to tune in 100KV motors and 29" props. As X1Aero stated, you have made it even more difficult on your self by running 6S and your total weight is low. When designing multirotor systems it is quite simple to get a quick idea as to how it will perform in the wind. Take the total area of the propellers X the number of booms to get the area each boom covers. Then divide in the weight of the craft. The higher the weight to area, or to quote in aviation terms the higher wing load, the better it will deal with wind. Notice I did not say the number of motors, because an X8 and a X4 behave similar in the same winds. If you are dead set on using your motor/prop combination, heres our advice. Go to 12S as X1Aero stated, select a good ESC, look at the ones at Tiger Motors for those props, and increase your weight, I would recommend you increase your weight and increase your voltage by using two 6S batteries in series. When you are done, your system will fly ok in low winds, just stay below 15mph. And never expect it to fly as stable as a smaller quad.
PS if anyone wants Tiger Motors 29" props and motors for recreation and fun, let me know.
We have tons of them on our shelves, as they are not safe for professional use.
I’ll back you up on this. I’m surprised he’s able to fly at all. I’ve tried 24 inch on 170kV U8s on 6S and the craft had a tendency to start oscillating then lose yaw authority in a simple stabilize hover.
What I always wanted to try and couldn’t find time building, was a counter-rotating pair of big props to do 80% of the lift at center, surrounded by a regular quad running smaller, faster props.
I also saw the problem with the desired Roll and pitch but i don’t know how to fix the problem. do you have an advice?
Did you also see the vibration value? Could these also cause problems withe the accelerometer an compass orientation? or are they i an aceptable amount?
I tried autotune several times but there were no acceptable results. i guess it’s somehow not performing in the right way.
how do you think i have to change the tune?
I am using 28"*9,2 T-Motor Props
The weight is without payload so i will add around 500g with the camera.
I am using a 6s 22000mah battery from GanceAce-Tattoo
@All
If i understand you all right you see the problem in the general setup. The Copter it self is to light and the probs spinn to slow? because of the light weight and the big probs i will get also problems with the wind (this problem and the calculation i knew before)
I will try your advice and increase the voltage and the weight. But i am more or less dead set to use this combination. So unfortunately i have to make this system flying.
I also guess i will become flight time trouble with the increased weight. At the end the Copter has to fly more than 50min.
One point i don’t understand, the copter is not flying that bad as you all say. It performs quiet well in stabilized and pos hold mode also in soft wind conditions. i now about the bad log datas, but why i have those big trouble in Auto mode?
Sorry I cannot help you as to why it operates poorly in Auto Mode, I only have 8 months experience with Arducopter but have designed systems around 5 other flight controllers. No mater what flight controller you use, I still stand by the fact that you cannot change the pure physics around the moment of inertia in that prop/motor combination. I have had many customers comment to us that they would rather have a craft that can fly in all conditions and lands after 20 minutes than to have a craft that can fly for 50minutes that has to wait for a calm day. Businesses cannot afford to wait.
Here’s a customer learning how to fly in winds gusting to 50mph
The Auto and PosHold flights look very similar as far as the data goes. Both attitude and altitude control is poor while the motor PWMs don’t change much. This shows that the attitude and altitude controllers aren’t being saturated, which is a strong indication that the motors simply cannot respond fast enough to meet the controller’s demands. This is a symptom of poor PID tuning or having large, slow props.
Vibrations are okay in general, but your Y vibration is higher than X, which is suspicious. Generally this means that the mount for the flight controller could be improved, as the damping is somehow different between those axes. Since this is a Pixhawk 2.1, I recommend just hard mounting it, as it has internal IMU isolation. (See VIBE.VibeX, VibeY, VibeZ. All should be <30, X and Y should be similar. Pixhawk 2.1 can achieve <10 for all axes)
Your copter is about 8% right-heavy, check weight distribution. Also, it has a mechanical right-yaw bias of about 10%, check motor alignment. (See RATE.POut, ROut, YOut. They should be centered around 0 in no-wind condition.)
I didn’t want to come to a conclusion before looking at the logs, but as the others have noted, it will be very difficult getting this configuration to fly well. I do have a few suggestions that could help without changing hardware, but that will only go so far.
Put a dummy weight on the copter that is equal to your intended payload. You need the weight, and you want to tune the copter for its intended mission anyways.
What ESCs are you using? You need regenerative braking, and some ESCs linearize the thrust-throttle curve, which needs to be taken into account. KDE and some T-Motor ESCs do this, which means you will need to change the MOT_THST_EXPO parameter (usually to ~0.25 or so for KDE and T-Motor).
Fix the motor alignment and weight distribution noted above. Doing some test flights and inspecting the rate outputs in the logs will allow you to verify good alignment.
Finally, run an auto-tune in no wind. This may be the best you can do without changing your powertrain.
And, of course, here’s a few options that you may want to consider regarding hardware:
Increase weight. The weight can be in the form of additional batteries if you want to keep your flight time.
Smaller props.
Higher KV motors.
Higher voltage.
The objective is to get your props spinning faster. This is important because controllability is strongly related to how quickly your motors can accelerate and decelerate. Big props have a lot of inertia, so your motors need a lot of torque to accelerate quickly, and there also needs to be a lot of drag to slow them down when needed (braking can only do so much). Higher voltage will get you the torque, and smaller props, higher KV, and/or more weight will get you higher RPM to get you enough drag.
@Infinite
You are totally right that it is better to have a copter flying 20min stable then to have 50min instable flight time. but unfurtunally it is a requirement of my university to reach more then 50min of flight time.
All other copters i bulit till now i did like you say and never had such trubble. But now i have to find a solution for this project
@Anubis
thanks for spendign so much time in my log datas
if i unserstand right the PID controllers can not work proper because the motors-prop combination reacts to slow?!
i am using a soft mount plate for the fc at the moment. I will hard mount the fc tomorrow and give it a try.
the mechanical right-yaw bias you mentioned, is it caused by a real mecahnical failure e.g. one Motor is spinnig to solw or is brocken or is it related to some parameters? I never had this problem with one of my copters till now.
i put a dummy weight of 500g on the copter yesterday and it performed a bit better even though the PIDs weren’t fitting. i will tune them tomorrow.
I am using “Kontrol-X 40LV” ECS from Sobek. i looked it up how this ECS behave with the trust throttle curve but i didn’t find anything. I will try the MOT_THST_EXPO paramter tomorrow.
I tryed to run Autotune several times. The log datas Show the Autotune is/was engaed. But while flying the Copter didn’t do anything and also the parameters didn’t change.i already set the RC_Deadzone above 70 but ist still not working. Do you have an idea?
Did you see the Err_EKF_check 0 and check0 message at the log datas? Do you have an idea how i could fix this problem?
About the Hardware changings, i am thinking About to use 2 of my 6s 22000mah batteries to get more weight and 44000mah.
You would probably be better off to put those batteries in series for 12S, replace the ESC’s with 12S capable units and then reduce the prop size.
What you have built is a Hover Machine which is the result when flight time is the primary goal. And for 50 minutes for a multirotor that would be the primary goal.
So 4.8kg is with one of those batteries? 7.92kg all-up-weight with 2?
I’m not familiar with those ESCs, but if they don’t mention thrust linearization, then they probably don’t do it. Leave the parameter alone unless you’re certain they do.
Not a failure per se, but definitely a real mechanical cause. This is almost always caused by the motors being slightly misaligned, i.e., not being level. When a motor isn’t level, it produces a small sideways thrust that causes the copter to spin, which makes the flight controller command a constant anti-torque to maintain heading.
The EKF error was produced by an uncertain position estimate. Eventually the EKF decided that the GPS readings were trustworthy and cleared the error. This isn’t necessarily indicative of a problem. Did this flight happen immediately after powering on the drone? Might have just needed a few more seconds on the ground.
Ah, I misunderstood the original post. I thought you did Autotune but it didn’t help, not that it didn’t even work. Usually the problem is the RC deadzones, but you tried that already, so I’m not sure. I’ll need a log of an attempted autotune flight to determine the cause.
Worth a try, but be conscious of your maximum takeoff weight. The logs posted above are already hovering at 50% throttle without payload. As Dave mentioned, 12S might work better, but you’ll need different ESCs (and maybe different BECs or other hardware) to handle 50V input.
The easiest first fix is to go as we have all said with 12S.
Two batteries will increase the weight and the higher voltage will make the motors more responsive.
If you go with 12S you will need new ESCs and make sure the other components on your craft can also deal with the increased voltage.
If you need to fly for 50 minutes and want stability I would recommend you scrap the multirotor and go with a single rotor craft. They can fly for hours and carry huge payloads. Just be careful around the large rotor, as they can cut your head off
If you need to fly for 50 minutes and want stability I would recommend you scrap the multirotor and go with a single rotor craft.
I suppose you are stuck with a multirotor but a Plane would be a much better choice also. It’s Trivial to get hardware in the air for 50 minutes even on a budget.
Well, your easiest solution would be to replace all your esc’s with HV versions and use two 6s packs in series. BTW, I still have a set of U8pro 170kv motors that would work for you on 6s with two 6s pack parallel. I’m selling these probably what you’d pay for a set of new tmotor esc’s : )
