David and Chad answer some viewer questions and share a little about what's coming up in the weeks ahead!
First off, we want to mention that we are planning to attend The Toledo Show: R/C Model Expo in April.
Learn more about the Weak Signals RC Club Event HERE
The articles mentioned in this episode are listed below in the 'related articles' section.
We will be reviewing the Skywalker X8 in the near future! David's looking forward to loading it up with all his FPV gear.
We're excited about some new multirotor components that will be available in our STORE very soon! Keep checking back for when these are available to purchase.
We'll be featuring more multirotor stuff in the episodes ahead, quadcopter, tricopters, t-copters, build videos and more!
Multirotors like this 3DR ArduCopter Quad-D will be coming soon!
We've been busy keeping up with all of the speed build orders, more of our swappable designs will be available in the months ahead. The Baby Blender is available as a speed build kit in the store!
The FT Cruiser (twin engine) swappable will be available as a speed build kit and the scratch build tutorial will be available very soon. Josh recently designed some cool accessories that can be added to this twin engine design that we think you'll really like!
David's been busy working on this custom made Saab 37 Viggen in answer to all of the viewer requests asking for more EDF stuff. We read all the comments, so please continue sending in your requests and comments on what you'd like to see.
We would like to also thank Flite Test viewer Teddy C. who designed these amazing snow skis for the Super Cub. Unfortunately, they arrived just after the snow melted so we haven't had a chance to try them out yet. We're looking forward to the maiden these first thing next winter! Thanks for sending these out Teddy keep up the great work!
David share more details about what was involved with his FPV to Space and Back project.
David is teaching Josh how to launch a Discus glider and that episode will be available in the weeks ahead!
Check out the articles mentioned in this episode:
Log In to reply
anyway i would really like to see that show on the Discus Launched Glider, i think that would be amazing to learn about.
Log In to reply
This is the best thing I have ever seen in the RC world….. I hate this saying but…. ‘You inspire’ and this brings out the Neil Armstrong that is inside us all… totally inspiring mate…….
Log In to reply
Log In to reply
Speaking of guys.... maybe you need to find a girl with as much interest in the hobby as the rest of us and put on the show!
Log In to reply
Log In to reply
Log In to reply
Log In to reply
Log In to reply
Log In to reply
Log In to reply
Log In to reply
Log In to reply
One thing though: You MUST do a VTail multirotor frame piece. PLEASE! It could be optionally used in place of the rotating yaw mechanism.
Log In to reply
A PID controller operates by calculating error and reacting to it. Error is the difference between the command from the user and the actual state of the device. A good example is a thermostat. If you set 72 °F, and the room is 68 °F, the error is 4 °F.
In the case of multicopters, they use gyros which report rotational speed. (Some controllers use accelerometers, which actually do give angle relative to gravity, but for this example I'll only talk about the simplest controllers, which use gyros only, like the KKboard). Thus you're actually commanding a certain rotational speed with the sticks if you use a gyro-only controller. The error in this case is the difference between the commanded speed and the actual speed reported by the gyro. The P,I,D constants adjust how the controller responds to the error:
Let's call the error "e", and call the controller output "T" (the controller outputs an throttle percentage, from 0-100, effectively)
P = "Proportional" This is a constant that gets multiplied by the error and sent to the controller output, so the output is 'proportional' to the error. It's simply a multiplier. If as the error increases, the controller will output more throttle. In math terms : T=P*e. Some may notice that if the error is zero, the controller isn't trying to correct anything. This is an important point. The P term can also be viewed as controlling based on "where you are right now". If the error right now is large, try harder to correct.
I = "Integral" This is a constant that gets multiplied by the integral of the error. Integral is a calculus term that simply means the product of the error and the time the error has existed. For example if I've had an error of 2 for 20 seconds, I'd have an integral of the error of 40. In math terms the controller uses the integral term like this: T=I*e*t. Integral(e) is the same as error*time. You can see where this would be useful. Imagine you want the copter to start rolling right. If you give 1/2 right roll on the stick, the proportional term will cause the controller to output a constant throttle value. But what if something is blocking the copter from rolling right, like wind? Well if you keep holding 1/2 right roll on the stick, but the copter isn't rolling, the integral of the error will build as time marches on. The integral term will make the controller try harder and harder to achieve a right roll. Integral action is thus necessary to to null out disturbances. Remember how proportional action outputs nothing if there is no error? This is ok if everything on the copter is perfect. What if you need some small adjustment just to keep the copter level? Integral action can do this because it lets the controller sustain an output even when the error is zero. In control engineering we call this elimination of "steady-state error". The I term can be viewed as controlling based on "where you have been in the past". If the error has accumulated, try harder to correct.
D = "Derivative" This is a constant that gets multiplied by the rate of change of the error. Derivative is another calculus term that simply means rate of change (with respect to time). In math terms, the controller uses derivative like this: T=D*e/t. Derivative adds a "damping" effect. It can be used to slow down changes in the movement of the copter. If you make the P term too high, the copter will oscillate. You can actually get rid of this oscillation by increasing the D term! This means you can keep the strong fast response of a high P term, but get rid of the wobbling it can cause. The D term can be viewed as controlling based on "where you are going in the future". If the error is on an upwards trend, try harder to correct so it doesn't overshoot. To me the D term is quite important to getting stable copter control. It's often overlooked but it can really make a significant improvement. For example if you use only P and I terms, when you roll to 30 ° and want to stop, the copter will bounce a little when it tries to stop. By adding a D term, you can get it to roll and then stop instantly without any bouncing. In control engineering we call this "critically damped". It's the fastest path to a new setpoint without any overshoot.
The controller adds all three of these terms, and then outputs to the ESC:
T=P*e + I*e*t + D*e/t
There are a lot of way to tune, if you have a mathematically model of the device. Thus we usually rely on empirical tuning methods. Most common is :
Set I and D to zero.
Increase P until oscillations start, then reduce by 20%
Increase I to get rid of steady-state error
Incrase D to remove overshoot when making setpoint changes.
Log In to reply
And thanks Flitetest crew for the comments on my dumbed down version of applying glue to foam board. :-)
Log In to reply
Log In to reply
Log In to reply
Thank you for an awesome show!! I have followed both flitetest and RCexplorer a while now and couldn't be happier when you started making videos together! I always look forward to the next episode!
I, myself, fly a quad and are just getting into FPV. My long term goal is to mimic the latest mars landing (curiosity rover) but I am sure you would do a much better job! That includes releasing a quad from high altitude in a parachute, cut the cord(!) and then hover down and, before touching the ground, lower a rover and then take off again. To me, that is the ultimate RC mission! I hope I planted some seeds there.
This is how it was done in reality http://youtu.be/Q7k_LsBWcCc
Thank you again for a good show!
/Tobias
Log In to reply
http://www.instructables.com/id/My-Space-Balloon-Project-Stratohab-Success-High/
http://www.hobbyspace.com/NearSpace/
https://sites.google.com/site/ucsdnearspaceballoon/
http://makezine.com/24/weatherballoons/
Log In to reply