14 | Measure angles with the MPU6050 accelerometer

Full code and manual on GitHub: github.com/Car...
In this video, you will learn how you can use the MPU6050 accelerometer to measure angles.
The purpose of this video series is to learn the basics behind a quadcopter drone and enable you to build one yourself, by dividing this challenging project in several easy-to-understand parts. You use the capable Teensy 4.0 microcontroller together with the easy-to-use Arduino language.

[@markmaker2488]

This was the best explanation of how to implement the data sheet information into code I have ever seen. Simple and clear instructions. Thank you.

[@nafs314]

This was so informative - you are genuinely great at explaining things!

[@robdevoer1]

Superbly presented, you spend a lot of time making the design easy to understand. Thanks for all your efforts and then sharing it here, it has made a big difference for my project.

[@johndarking8078]

I should add that I was always impressed by the quality of my Belgian engineer coworkers.

[@johndarking8078]

Further on pitch/roll reversed: This obviously depends on orientation. I have implemented on a breadboard style veroboard so the Y-axis is in the forward direction, hence Y is roll and X is pitch. On AccZ at 1.6g: I bought a batch of 3 cheap mpu6060 boards (GY-521). The other 2 all give AccZ within 0.06 of 1g. I am a little suspicious of the 0.6 variation. I shall keep digging.

[@theblobewab6844]

yoo this series is exactly what I was looking for thank you so much! underrated channel. One thing I don't understand is that if there is an acceleration happening that is not equal to 1g then wouldnt that make the accelerometer readings invalid for using for calculating final angle?

[@carbonaeronautics]

Thanks! If accX for example is 10g and accY and accZ are ≈ 0g, the result will be exactly the same as when accX is exactly equal to 1g. You can calculate this example through: pitch angle=atan(accX/sqrt(accY^2+accZ^2))=atan(1/(≈0))=atan(10/(≈0))=90°.

[@theblobewab6844]

@@carbonaeronautics Makes sense, thank you

[@АК-32ОлексійКозакевич]

Can you explain how you are able to use this measerments while quadocopter is moving in space. Because movement will lead to incorrect angle measures(It can show that quadocopter has non-zero pitch and/or roll while just accelerating in some direction)

[@HenishViradiya]

I was stuck with this thing. Thank you very much for helping in this.

[@velocitylaunchsystems]

When will be the next video and do you have any other videos that fully explain a kalman filter, this video was extremely helpful and for the first time i understood why simply using acceleration cannot work in a system that is constantly changing acceleration

[@carbonaeronautics]

Yes next video will include a full mathematical derivation of a one dimensional Kalman filter, in which we will use the gyroscope to predict the angle and the accelerometer to correct the prediction. I hope to finish it by the end of this week!

[@nassimsadikou1699]

hey, it seems the YAW angle computation based on Accelerometer propery isn't working, i mean after i got the value of AccZ who is equal 1, the Angle based on trigonometry should be : atan(AccelZ / sqrt(pow(AccelX , 2) + pow(AccelY , 2))) /(PI/180);

[@DavisidMatius]

Thank you very much! I have a question for you, if one decided to use merely the angular accelerations from the sensor and estimating the angle using numerical integration. If instead of using the rectangle numerical method (the one you showed in the video), one uses a more complex but more reliable numerical method, do you thing the drifting of the error would slow down?, and also, do you think that changing the sample frequency of the data acquisition from the sensor (and the Arduino) can improve the results?

[@dabunnisher29]

This is a great video for explaining this sensor. Subscribed and Liked your video. Thank you.

[@P0K0]

Great sir, thanks a lot for explaining everything simply ❤️

[@jiriminarik6702]

The accelerometer angle measurements will work only when the device is stationary. Whenever you try to move, you will lose angle precision.

[@markmaker2488]

Did you watch the video to the end?

[@BmsquareC]

we use it just for making attitude alignment

[@viktor_pavlov_channel]

A very strange idea about finding the roll and pitch vectors at each iteration are different, which means they have different lengths. Yes, in the picture that we see visually the angles are similar, can you provide proof?

[@brassmonster]

Thanks a lot man, you are a legend

[@johndarking8078]

On trying this example I seem to have pitch and roll reversed. (MPU6050 in same orientation as example.) Also AccZ was 1.6 by default, had to correct by 0.6, however AccX and AccY similar to example. [I used to work in Antwerp back in the 1980s. Nobody told me an alien spaceship landed on the Port Authority building.]

[@samurai4k393]

From where we get the rotation Rate, is it from the Gyroscope ? 🤔and From where the blue signal of the gyro at the End get from ?.

[@aliali-qd3rk]

well done bro , keep the hard working

[@borensztejntitouan736]

Hello, your Z axis is pointing to up but you have 1g ? You're not supposed to have -1g because gravity acceleration is pointing down ?

[@alpxka5801]

Yes, gravity accelerates the module downwards (-1g). So, in order to keep its horizontal position, it has to be accelerated upwards by the surface it is mounted on (+1g). Therefore, the net acceleration equals -1g + 1g = 0g

[@meme新聞]

Great lecture!!! Can you give me the material for this subject? In Vietnam it is very difficult to find documents like yours. Thanks Teacher.

[@carbonaeronautics]

Sure, it will be published on GitHub with the other parts (currently published up to part 13): github.com/CarbonAeronautics/Manual-Quadcopter-Drone

[@meme新聞]

@@carbonaeronautics Thank you Teacher!!!

[@AngilynMaeReyes]

great idea sir, may I ask how to get the yaw angle from this? thank you very much!

[@henksteenfrits3290]

I like to know this to

[@techtheguy5180]

@@henksteenfrits3290 and Angilyn you can't get it using the accellerometer. You need to integrate the gyroscope.

[@Rafaeldelucca-bd2vk]

Thank you

[@zoeyzhang9866]

Informative tutorial it is! Not sure if any custom PCBs/ 3D printing parts may help for any upcoming content? If so, would love to supply freely, if possible, will be cool if any collabs could be achieved together! (PCBWay zoey)

[@Tom-wd2ee]

because of you i will pass exams :D

[@alank51709]

Hi , I want to know why 5:17 theta roll in the middle is equal to others , can someone explain, thanks .

[@DarthVader-sm5wc]

exactly i am also getting confused did u get to know?

[@sidhaksingh7675]

I didn’t get this as well. They leave us with half reasoning

[@viktorpavlov7763]

Same thing i miss understand.

[@mahmoudsaid762]

Same also confused

[@Auddy_s7395]

What about the yaw angle? Slow Rotate horizontal Problem while the drone is still?

[@matiasbravo4482]

i hve the same question, did you found any answer?

[@RixtronixLAB]

Cool video shot, thanks for sharing, keep it up :)

[@alphaparticle2823]

Hii Where did u find that Ts=0.004 i.e for 250Hz cycle .how did we select that

[@stargate-offic]

please , i want to detecte the apogee of a rocket , mpu6050 or a barometre bmp280 who is the best one in this case ????

[@truongthong7513]

Thanks so much !

[@brotherhood4627]

Thank you so much really helpful!!😇

[@romainbayol]

Hi, I have a question : why are you choosen 8g for the accelerometer value ? Why not 2,4 or 16g ? Thanks

[@carbonaeronautics]

2 or 4g would be too coarse for control purposes, but you could try 16g if your application requires a very precise value

[@hocuspocus1126]

I want to ask, base on that equation wouldn't the tan(angle) of roll change if it starts accelerating in the x axis. Thanks

[@OscaKholopha]

The approach he used to calculate the roll and pitch is only valid for system that is not in motion other than rotating. Once the system starts to accelerate in the x and or y directions, the results become incorrect

[@covingtonkua9404]

How did u specify which SDA and SCL pin is used?

[@ThorAirlines]

I dont know if you will reply this, but i am using your code in a RC plane. Using this to know if im upside or downside. When you turn the sensor downside stills having 0 roll angle. did you know this "bug"? Its just for notice you, i am trying to solve it. (apologizes me for my english).

[@spazmoidectomorf6209]

Did you manage to fix it, if not you can use the 2arctan function, but you will have to make you own custom function to recognise when the numerator is positive or negative to determine what quadrant the angle is in

[@richterd21]

What about to use Kalman filter to eliminate vibrations?

[@carbonaeronautics]

Exactly, its covered in part 15!

[@vayunandanakishore6652]

Good information...

[@dangata3358]

Could you please help me to find (ADXL335, HMC5883L, Tilt sensor) libraries for proteus

[@kinhkong69z62]

how to calculate yaw ?

[@igort5418]

Why don't use the dmp to get fused data ?

[@hamedtaran2102]

I also use this library, but the z-axis value increases little by little. It has a little drift, help me please

[@ranitroychowdhury7084]

What is the Project 4?? Can you please tell me?

[@afankhan7732]

How do we adjust the Sample rate?

[@GaganD-h2s]

How to measure yaw angle, please help me

[@pramsss4422]

how to read pan or yaw angle?

[@thorverhoeven3426]

i can't find the full code on github please someone help

[@thebigelectrode]

can anyone please explain the formula derivation for pitch and roll angle if someone can please point me to relevant resources that would be very helpfull i literally asked everyone my professors and spend a lot of time please help

[@alenninan5524]

Hi , can you send me the code of this video ?

[@ahmedmoustafa6829]

8:26 it is better to use atan2 instead of atan.

[@tongvy5294]

i read it and now i don't remember me who i'm! @.@

[@MariaAnderson-e4i]

Hall John Taylor Susan Walker Dorothy

[@pykid1915]

hey carbon how can i contact you? discord or smthng ??

[@carbonaeronautics]

You can open an issue on my Github page and I will have a look at it!

[@pykid1915]

@@carbonaeronautics 😁