Sunday, April 4, 2021

Programming Alternatives


The alternative to automatically changing from passthru mode to gyro-accelerometer mode during takeoff is to make the pilot do the change. This means to add the following 4 OpenTX mixer configurations, 
  1. Collective pitch suppression in passthru mode This is what prevents a takeoff at passthru mode This is not used in the automatic programming because the automatic swithcing happens at very small collective pitch
  2. Tail power proportional to throttle in passthru mode This is not used with automatic programming with multiple betaflight profiles. But this is used also with single betaflight profile.
  3. Fixed RPM for accelerometer-attitude mode This is not used in automatic programming because the same swith C position is used for both spool-up and acc-att mode, which means that as soon as throttle position is below 1/8 , rotor spools down.
  4. Arming guard This is also used with automatic programming
The problem with non-automatic programming is the point 3 configuration. Point 3 configuration allows accidental RPM jump when the pilot learns that the switch C needs to go up for the craft to takeoff, but he/she can flip the switch C before the spool up, resulting in tip-over. We may argue that the pilot may just as well accidentally pushes the throttle up so quickly that the craft also tips over. But the human brain is customed to analog control, meaning that it hard to form habit of flipping the switch C just before takeoff than to form the habit of pushing the throttle slowly. 

We may wonder if the accidental RPM jump can be prevented by checking if RPM is already full speed before accelerometer-attitude mode can enter, but the OpenTX logical switch has no such condition of changing switch position to middle. Instead, the condition can only be the middle position itself. Now, if middle position means to have accelerometer-attitude stabilization only after spool up, and at the same time collective pitch becomes curve-defined, this entire manual programming becomes tuning in the field because the throttle point at which the stabilization is engaged determines the amount of tail wag and rotor wobbling.  Is it desireable to have field tuning? Or have a optimized build for consistent operations? 

We may argue that we can program point 3 to have a curve from zero throttle to fixed RPM when switch C is flipped on. But then this is becoming an quasi automatic programming. Further more, if this quasi automatic programming does not allow acc-att mode in craft during the rise of throttle to prevent the capsizing problem, this quasi automatic programming becomes fully automatic. But, if this quasi programming allows acc-att mode, the capsizing problem of other helicopters becomes the problem of this craft, meaning that our helicopter has the same ubiquitous problems as others.

Re-examine Statements About Engaging Stabilization

So, the OpenTX transmitter should engage a stabilization mode when ever the rotor spools up. 
Is that statement true? What if I spool up the rotor but keep the craft on the ground just to go through some check lists by keeping the collective pitch very low or zero? 
The answer is that you are adding manual controls that the programming can't handle without requiring 4, 5 modes. For example, you may want to disable just the I-term , but that becomes the 4th mode. And you may want to enhance the tail I-gain while loitering, and that becomes 5th mode. 
The core of the problem is whether engaging the full PID while the collective pitch is so low, nearly zero, can capsize the heli. If the answer is no, the statement is very true, and the solution is very simple. If the answer is yes and no, meaning the wobbling sometimes ends in capsizing, then the very low pitch or the spool up curve needs to be tuned and landing characteristics (sudden torque loss) suffers. The core problem is finding the optimum tuning and curving. 

The alternative answer is that, once we have a good tuning, you can spool the rotor up without taking off and performs your checklist with relatively minor wobbling.  

We need to separate the definitions of capsize and tip-over. Capsize means the forceful, erratic I-term winding, over correcting the craft. This can damage the rotor.

Tip-over refers to the forceful rotor starting torque from standing still. The unbalanced rotor knocks the heli over. This does not damage the rotor because the rotor is just starting up with a very small momentum.

Deep down, ultimately, the problem is the difference between CC3D and FD411 flight software. CC3D has a integral reset mechanism that is used by many manufacturers. The integral reset cancels integral term on the cyclic control momentarily when the attitude deviates from the set point to some threshold, and it is always active. FD411 allows this optional configuration as "I-term relax".  When tuning is off and the craft is capsizing, CC3D allows the pilot to reset the integral by simply yanking the roll stick to the opposite extreme. FD411 without I-term relax, needs the extreme stick roll to accumulate to the exact amount of the deviation, which is not possible to achieve by hand.  



-------------------
solution,either the I term or all the 3 PID terms needs to be canceled during spoolup time. CC3D readily allows an additional mode to cancel the I term while FD411 has github branches of switching PID profiles. But canceling the entire PID process is the simplest, straightforward solution, which is the chosen solution. Tail still needs counter torque proportional to rotor RPM with such solution in the spoolup state.

The racing drone flight computer software is modified to allow changing flight characteristics mid-flight for utility use, downloadable here betaflight_4.2.5_STM32F411_.hex . After the software is flashed onto the flight computer, the manufacturer default configuration, dump2.txt,  needs to be applied before adding our specific configuration. Our Betaflight diff config file for Betaflight Configurator 10.7.0 is download able here Converged-IoT.txt .  The configurator needs to have PID profile switching feature and it is downloadable here betaflight-configurator.pid_profile.tar.gz .

Our conclusion is that there is really no alternatives. 
The question is whether such a stringent criteria will force us to use power tools to build it?
Yes, power tool is required. The rotary cutter is required.

How about a 10 pond vise?
No, not a vise needed. All tools weigh under 1 pond.

Soldering required? 
Yes.

Binary compiler required?
No.

  

1. 50-gram motors powerful enough for 250 gram helicopter direct drive acrobatics is available from multiple vendors.

2. Single rotor does have slightly better fuel economy and endurance than quadcopters, as calculated according to mechanical physics.

3. Main rotor cage and servo cage geometry is congruent between Blade 230S V2 and BSR helicopter frames. Interchangeable parts are readily off the shelf.

4. Sub-6 gram servos for acrobatics is available, but only 1 brand has integrated fast response fast enough for comprehensive acrobatics and terminal speed diving .
    The servos are compatible from different vendors. Needed physical hacks already done.

5. Blade 180 and Oxy 210 rotor head parts are compatible with Blade 230, exception DFC link that needs hacking.

6. Receiver needs anti-noise arrangement for the direct drive motor system.

7. 1105 low rpm motors are sufficient for this 250 gram heli with clever prop selection.

8. Servo saver mod suffices for craft durability comparable to quadcopters.

9. Blade 230S V2 body suffices for craft durability comparable to quadcopters.

10. Blade balancing and stabilization tricks enables unprecedented FPV maneuverability for helicopter. All it takes is 2-3 times on-field training for a person to hear the difference between the fine dynamic balance and the coarse balance in an outdoor setting. However, this balancing process can not be further abridged to fully match quadcopter's build/operating simplicity.

11. At least 2 very good radio TX/RX combinations are available.

12. PID for longitudinal pitch is 3 times larger than craft rolling PID. I gain needs to be very large for cyclic control. D gain is not appropriate for cyclic control, but large D gain is needed for yaw control.
      Cyclic PIDs are inversely proportional to rotor RPM.

13. The elevated RPM for video presentation trades off flight endurance and flight time.

14. Clever TX curves and setups allows pleasant landing, approaching the ease of use of quadcopters.

15. Extreme light weight FPV transmission for urban adrenaline stunts is quite affordable.

16. Conventional rotor hub setup suffices and trumps tricky retaining mechanisms.

17. Conventional ESC suffices and trumps bleeding edge development products.



No comments:

Post a Comment