king warthox !!

[Alexinparis]

he finally did it with a special ESC made by Felix and a small MWC mod
very impressive and innovative

"my first cautious attempts to fly a 3d capable quadrocopter."

[tovrin]

wow, i am full of aw and jealousy

[copterrichie]

Talk about pushing the Envelop!!! Kudos.

[NikTheGreek]

Impressive !!!

[alll]

And now mwc_esc / SimonK esc to do the same ..., and have a special mwc_3D with tunable PID via Aux-outputs )),
i will love to try it out.
manu

[crashlander]

Really Cool!
Any chance to get those mods for testing?!

Regards
Andrej

[Andy7]

Not that is the definition of awesome.

You know, I'd raally love to see some good tutorials on flying techniques. Has anyone ever made any decent ones? I've not seen many. If WartHox were to make one then it would be an immediate hit.

[ronco]

Hi,

i can give you the mod i did to MWC 2.1 .. the oneshot PWM part will be released by amperedieter soon.

in MWC.ino to be able to arm with a neutral throttle stick position: (you can only arm with a switch)

Code: Select all

#define RC_FREQ 50

  if(!rcOptions[BOXARM]) { ///////////////////// important to be always able to disarm
    f.ARMED = 0;
  }
  if (currentTime > rcTime ) { // 50Hz
    rcTime = currentTime + 20000;
    computeRC();
    // Failsafe routine - added by MIS
    #if defined(FAILSAFE)
      if ( failsafeCnt > (5*FAILSAVE_DELAY) && f.ARMED) {                  // Stabilize, and set Throttle to specified level
        for(i=0; i<3; i++) rcData[i] = MIDRC;                               // after specified guard time after RC signal is lost (in 0.1sec)
        rcData[THROTTLE] = FAILSAVE_THROTTLE;
        if (failsafeCnt > 5*(FAILSAVE_DELAY+FAILSAVE_OFF_DELAY)) {          // Turn OFF motors after specified Time (in 0.1sec)
          f.ARMED = 0;   // This will prevent the copter to automatically rearm if failsafe shuts it down and prevents
          f.OK_TO_ARM = 0; // to restart accidentely by just reconnect to the tx - you will have to switch off first to rearm
        }
        failsafeEvents++;
      }
      if ( failsafeCnt > (5*FAILSAVE_DELAY) && !f.ARMED) {  //Turn of "Ok To arm to prevent the motors from spinning after repowering the RX with low throttle and aux to arm
          f.ARMED = 0;   // This will prevent the copter to automatically rearm if failsafe shuts it down and prevents
          f.OK_TO_ARM = 0; // to restart accidentely by just reconnect to the tx - you will have to switch off first to rearm
      }
      failsafeCnt++;
    #endif
    // end of failsave routine - next change is made with RcOptions setting
    if (rcData[THROTTLE] > 1450 && rcData[THROTTLE] < 1550) { ///////////////////// if throttle is at middle
errorGyroI[ROLL] = 0; errorGyroI[PITCH] = 0; errorGyroI[YAW] = 0;
      errorAngleI[ROLL] = 0; errorAngleI[PITCH] = 0;
      rcDelayCommand++;

in output.ino .. half gains because the throttle signal is also half in each direction (for example cw -> 1500-2000 and ccw 1500-1000us)

Code: Select all

  #ifdef QUADX
  //------------------------------------------------------------------------------------//
    motor[0] = PIDMIX(-1/2,+1/2,-1/2); //REAR_R
    motor[1] = PIDMIX(-1/2,-1/2,+1/2); //FRONT_R
    motor[2] = PIDMIX(+1/2,+1/2,+1/2); //REAR_L
    motor[3] = PIDMIX(+1/2,-1/2,-1/2); //FRONT_L
  //------------------------------------------------------------------------------------//
  #endif

and other min, maxthrottle and armed settings -> the ESC's are set to start at 1525 CW and 1475us CCW so thay always spin

Code: Select all

/****************                Filter the Motors values                ******************/
  maxMotor=motor[0];
  static uint16_t minMotor = motor[0];
  for(i=1;i< NUMBER_MOTOR;i++)
    if (motor[i]>maxMotor) maxMotor=motor[i];
  
  for(i=1;i< NUMBER_MOTOR;i++)
    if (motor[i]<minMotor) minMotor=motor[i];
  
  for (i = 0; i < NUMBER_MOTOR; i++) {
    if (maxMotor > MAXTHROTTLE) // this is a way to still have good gyro corrections if at least one motor reaches its max.
      motor[i] -= maxMotor - MAXTHROTTLE;
      
    if ((rcData[THROTTLE]) > 1500){
      motor[i] = constrain(motor[i], 1530, MAXTHROTTLE); 
    }else{
      motor[i] = constrain(motor[i], MINCOMMAND, 1470); 
    }

    if (!f.ARMED)
      motor[i] = 1500;
  }

but without the "right" ESC's this wont help you .. and the even the first release of the UltraESC's wont support that .. because we need to prove it.. later we will give out an FW update.

regards

Felix

[subaru4wd]

SimonK firmware already supports this, so if you have any SimonK compatible ESC, you are ready for this. Just have to go change some code.

[ronco]

Hi,

i didnt tested simonk's reverse function .. but it needs to be very fast..

regards

felix

[copterrichie]

Question: How do one go about using the reversing feature of the SimonK ESC? Might be very useful in the land rover I plan to build.

[nicog]

simple you just change .equ RC_PULS_REVERSE = 1 ; Enable RC-car style forward/reverse throttle

and you build the target

[copterrichie]

Awe!!! Thank you.

[sismeiro]

If someone told me that a multicopter pilot did what I saw in the video only one name would came to my head: Warthox.

[timecop]

but without the "right" ESC's this wont help you .. and the even the first release of the UltraESC's wont support that .. because we need to prove it.. later we will give out an FW update.

Nice try. Sorry, it has been possible without your "right" ESC, for quite a whlie. Hope you've got a lot in stock, because something's going to happen very soon.

[ronco]

but without the "right" ESC's this wont help you .. and the even the first release of the UltraESC's wont support that .. because we need to prove it.. later we will give out an FW update.

Nice try. Sorry, it has been possible without your "right" ESC, for quite a whlie. Hope you've got a lot in stock, because something's going to happen very soon.

so you work on a ESC? toshiba FOC 32bit ??... and no we dont have a lot thats our problem

regards

felix

[scrat]

One stupid question. How FC knows when copter is upside down and when to turn motors in different direction?

[Lapino]

Warthox did this, not the FC just like 3D heli flying you have motors reversed at throttle down and normal full power at throttle up

[tovrin]

my brain has a hard time with this, switching motor direction seems wrong, your losing all kinetic energy and then reversing it. i would have guessed that he was using variable pitch. its awesome, regardless

[tovrin]

my brain has a hard time with this, switching motor direction seems wrong, your losing all kinetic energy and then reversing it. i would have guessed that he was using variable pitch. its awesome, regardless

[phenolic]

i can give you the mod i did to MWC 2.1 .. the oneshot PWM part will be released by amperedieter soon.

Thanks for providing information on your patch! It has been fun playing with it. I adapted it to MultiWii_shared and added a bit of code. I'm using it with simonk ESC's and it works (mostly) fine but my testing has been limited and still needs tweaking.

Some things I want to point out:

When the motors are reversed they are not controlled in the same way as normal. When upright the prop speed is increased on motors to change orientation. When inverted the prop speed is decreased to change orientation which makes it fly somewhat differently, it is less stable. In fact it becomes less stable the larger the control input when inverted (motors reversed). MultiWii is actually doing to the same thing and always increasing the throttle signal but when the motor is reversed an increase in throttle is actually a decrease in speed. I think the code needs to be reworked so that when the motors are reversed the logic in MultiWii in inverted. Obviously this would be a much more significant patch to MultiWii.

Does LEVEL mode work? Does it work when inverted? If inverted does it flip it back over upright but keep the prop rotation reversed? (ie. that would be bad ).

Watch out if using FAILSAFE because it's going to probably require a much higher throttle value than non-3D mode. This requires a working LEVEL mode as well.

[ronco]

Hi,

in gyro mode MWC wont recognize that it is upsidedown .. its just the same as normal flight .. so im sure there is nothing that needs to be changed (beleve me i and warthox tested it)

i think level mode is not needed when you want to use that..

regards

felix

[phenolic]

What I'm talking about has nothing to do with the orientation (eg. upside down). It's the gyro logic when the motors are running in reverse and MultiWii does know whether the motors are running forwards or in reverse. I know it works as is but it's less stable when the motors are running in reverse because it's reducing prop speed to drop one side instead of increasing speed it pull one side upwards like when the motors are running forwards. I tested this too. Sorry, maybe I'm not explaining it clear enough. Despite the fact that it works, the current logic is not as good as it could be.

I'm assuming you're talking about my first point, if not then disregard.

Edit:
OK, I'm going too far. Now with pictograms. Apologies if you understood me to begin with.

Normal motor running forwards:


With the current changes this is how the motors act when they are running in reverse. Again, this has nothing to do with orientation but I'm showing the craft inverted to make it clear.


This is how it should work which is the same as when the motors are running forwards.


Yes, at the end of the day any of these methods will work, but not exactly the same because decreasing speed on one side while keeping it the same on the other is less stable than increasing speed on one side while keeping the other the same.

Haha, confusing, eh?

---------------------

As a completely separate topic, an inverted level mode would be cool nonetheless.

[ronco]

Hi,

MWC does both. reducing the power on one side and increasing it on the other one .. so it doesent matter in which way you do it.

.. see the mix tables... every gain works in both directions..

[edit]

[/edit]

regards

felix

[phenolic]

I thought it only does both when the speed on the motors is maxed out (ie. extreme stick movement or when near max throttle). At least that's the way it looks.

When making less extreme stick movements and minor corrections it works similar to my pictures. One side will get more (or less when running in reverse) speed while the other does not change much.

I could be wrong but I'm going by what I see from the motor outputs.

Also when flying it's very definitely less stable when inverted. Could be because of props though, we need someone to start selling reversible straight props.

[warthox]

When the motors are reversed they are not controlled in the same way as normal.

how do you come to this conclusion? by pratically testing it [which can be very error prone due all the additional hardware which also have to be set right and may is not]?

Does LEVEL mode work? Does it work when inverted? If inverted does it flip it back over upright but keep the prop rotation reversed? (ie. that would be bad ).

why should the level mode dont work in 3d mode?
a normal mwc quad will flip when its inverted and acc is activated. i think a 3d quad will also do that, why shouldnt it do that. i will test it.

imho nobody who wants to fly acro or real 3d needs and will use an acc/level mode. these people are far away from using it. the difference between acc and 3d mode is like the difference between rockcrawlers and sportscars. nobody uses a sportscar for rockcrawling and also doesnt needs to use it for rockcrawling at any time.
also i think somebody who needs to use the level mode to fly is far away from flying 3d.

Watch out if using FAILSAFE because it's going to probably require a much higher throttle value than non-3D mode. This requires a working LEVEL mode as well.

you set failsafe like with a normal mwc quad. which means set throttle to a point where the quad slowly decreases. this value obviously have to be >1500 in 3d mode. for emergency landing set it to 1000
if you set failsafe throttle to low a normal quad will fall. in 3d mode it will just actively fall thats the only difference. but should be obvious to people which deal with 3d mode.


talking about the 3d mwc mod, the quad flies exactly the same upside down as upside up. theres no difference in stability. at least i didnt noticed it.
afaik mwc is adjusting motor rpm in both directions and the ultraesc is also able to actively adjust motor rpm in both directions.
if you use an esc which only actively adjust rpm in increasing direction [like every normal esc] it wont work properly.
for 3d its necessary to use an esc which can do it in both directions like its necessary to to hold your breath while diving.

[phenolic]

I was just pointing out my observations.

how do you come to this conclusion? by pratically testing it [which can be very error prone due all the additional hardware which also have to be set right and may is not]?

By looking at the actual motor outputs on the bench. That is, the raw values that MultiWii is outputting.

why should the level mode dont work in 3d mode?
a normal mwc quad will flip when its inverted and acc is activated. i think a 3d quad will also do that, why shouldnt it do that. i will test it.

I wasn't worried about level mode while doing 3D. I was wondering what happens if you (maybe accidentally) set level mode while the motors are reversed. My theory was the craft will flip itself upright except the motors will still be reversed, causing a fast, possibly unexpected decent.

imho nobody who wants to fly acro or real 3d needs and will use an acc/level mode. these people are far away from using it. the difference between acc and 3d mode is like the difference between rockcrawlers and sportscars. nobody uses a sportscar for rockcrawling and also doesnt needs to use it for rockcrawling at any time.
also i think somebody who needs to use the level mode to fly is far away from flying 3d.

The option is there, somebody will use it. Either by mistake or intention. I was just wondering what would happen. Maybe level mode needs to be completely disabled if it's dangerous when using this 3D stuff.

Again, I was just pointing out potential issues I see as this develops. Things to think about to make sure the whole system is consistent with no unexpected consequences. I'm sure you guys will figure it out.

[M4d_D]

hi, this seems very interesting, nice flying warthox !

you talked about the simonk esc and that a model is capable of doing this if simonk escs are installed. does that mean one just needs to reprogram the mwc board or do you also need to reprogram the escs ?

how does it do it with the direction change, since the blade has to stop to be able to turn in the other direction, and this within a few milliseconds .

at the end of the video it says that modified props were used , which modifications were done on these props ?

cheers

[Hamburger]

with the way an arduino behaves upon power-up and resets (like after uploading new MultiWii code), it becomes dangerous, because the MWC sends out a low signal, which with the new reverse-ESC will equal to full-backwards.

To avoid such dangerous situations during bootup of flight controllers the esc firmware could be changed to have an extra range to take care of that
- 1000 to 1030 : with motors off
- 1050 to 1470 : motor backwards; full back for 1050 to slow back at 1470
- 1470 to 1530 : deadband; motor stop or better slow rotation to avoid start problems
- 1530 to 2000: motor forward as usual

During flight you want to avoid the motor suddenly stopping when giving full back throttle (pulling throttle stick). To achieve that, you could change flight-mode in your tx to activate a shifted throttle curve that has its range going form 1050 to 2000. Most modern TX already provide this feature (flight modes) when model is setup in heli mode. The different modes are often called idle-up and 3d.

[Alexinparis]

I thought it only does both when the speed on the motors is maxed out (ie. extreme stick movement or when near max throttle). At least that's the way it looks.

When making less extreme stick movements and minor corrections it works similar to my pictures. One side will get more (or less when running in reverse) speed while the other does not change much.

I could be wrong but I'm going by what I see from the motor outputs.

Also when flying it's very definitely less stable when inverted. Could be because of props though, we need someone to start selling reversible straight props.

Hi,

ronco is right,
multiwii corrects both side at the same time, and with the same magnitude.
you can look in the code as the PIDMIX definition:
#define PIDMIX(X,Y,Z) rcCommand[THROTTLE] + axisPID[ROLL]*X + axisPID[PITCH]*Y + YAW_DIRECTION * axisPID[YAW]*Z

if axisPID in non null, it will apply a correction factor on every motor concerned by ROLL.

[Alexinparis]

with the way an arduino behaves upon power-up and resets (like after uploading new MultiWii code), it becomes dangerous, because the MWC sends out a low signal, which with the new reverse-ESC will equal to full-backwards.

To avoid such dangerous situations during bootup of flight controllers the esc firmware could be changed to have an extra range to take care of that
- 1000 to 1030 : with motors off
- 1050 to 1470 : motor backwards; full back for 1050 to slow back at 1470
- 1470 to 1530 : deadband; motor stop or better slow rotation to avoid start problems
- 1530 to 2000: motor forward as usual

During flight you want to avoid the motor suddenly stopping when giving full back throttle (pulling throttle stick). To achieve that, you could change flight-mode in your tx to activate a shifted throttle curve that has its range going form 1050 to 2000. Most modern TX already provide this feature (flight modes) when model is setup in heli mode. The different modes are often called idle-up and 3d.

or maybe just adjust MINCOMMAND to 1500 in the code

[Hamburger]

or maybe just adjust MINCOMMAND to 1500 in the code

are you sure an arduino will never output a pwm of 1000 on its pwm outputs during reset/reboot before MWii takes over control of the output value (again) when the pin was put into pwm mode on the prior run (right before the reset is triggered)?

[Alexinparis]

or maybe just adjust MINCOMMAND to 1500 in the code

are you sure an arduino will never output a pwm of 1000 on its pwm outputs during reset/reboot before MWii takes over control of the output value (again) when the pin was put into pwm mode on the prior run (right before the reset is triggered)?

By default, HW pwm outputs are all disabled at boot => no signal to the ESCs.
Once multiwii setups the needed PWM or interrupts, there is almost immediately this command in the code: writeAllMotors(MINCOMMAND);

[phenolic]

you talked about the simonk esc and that a model is capable of doing this if simonk escs are installed. does that mean one just needs to reprogram the mwc board or do you also need to reprogram the escs ?

I am flying with normal Blueseries 12A's and simonk firmware. I need to take video when I get time but I'm still tweaking the setup. The ESC firmware needs certain options enabled and MultiWii needs to be modified (so they both need to be reflashed).

ronco is right,
multiwii corrects both side at the same time, and with the same magnitude.

OK! I guess my bench tests are wrong. Maybe because it's not actually flying. When I push the stick I see dramatic power increase on one side and almost no effect on the other.

As far as instability I guess that means my props are no good, but I already knew that.

or maybe just adjust MINCOMMAND to 1500 in the code

Yeah, that's what I did.

By default, HW pwm outputs are all disabled at boot => no signal to the ESCs.
Once multiwii setups the needed PWM or interrupts, there is almost immediately this command in the code: writeAllMotors(MINCOMMAND);

If I power up my MultiWii board then press the reset button on the board my motors run (pretty fast too). It does this every single time. However, I'm still in the process of tweaking my simonk config along with the changes to MultiWii. I do think however that something will need to be done to prevent these unexpected spool-ups of the motors but I need to take some time and determine why it's happening in the first place. Hamburger's idea may be the solution.

[phenolic]

Plain HK BlueSeries ESC with simonk firmware. Running MultiWii_shared r1253. No more flying it outside for me until I have better props, the lack of thrust is disturbing.

http://www.youtube.com/watch?v=f6PavUVXFe4

[Hamburger]

or maybe just adjust MINCOMMAND to 1500 in the code

are you sure an arduino will never output a pwm of 1000 on its pwm outputs during reset/reboot before MWii takes over control of the output value (again) when the pin was put into pwm mode on the prior run (right before the reset is triggered)?

By default, HW pwm outputs are all disabled at boot => no signal to the ESCs.
Once multiwii setups the needed PWM or interrupts, there is almost immediately this command in the code: writeAllMotors(MINCOMMAND);

SimonK posted some info that might explain why we do occasionally see those nasty short power bursts of motors during reset. Unfortunately, it is only an intelligent ESC that could prevent this, so I understand. http://www.rcgroups.com/forums/showpost ... count=7306

[nicog]

It's a pity that simonk escs and naze32 can't do the same.
http://www.youtube.com/watch?v=V1E8lWlS ... e=youtu.be

[warthox]

why should the level mode dont work in 3d mode?
a normal mwc quad will flip when its inverted and acc is activated. i think a 3d quad will also do that, why shouldnt it do that. i will test it.

I wasn't worried about level mode while doing 3D. I was wondering what happens if you (maybe accidentally) set level mode while the motors are reversed. My theory was the craft will flip itself upright except the motors will still be reversed, causing a fast, possibly unexpected decent.

i tried the acc mode while inverted and it behaves as expected. the quad flips over.
if throttle is inverted before then its also inverted after and the quad falls 'actively'.
but the same would happen when throttle is low on a normal quad except it doesnt fall that fast.

[phenolic]

i tried the acc mode while inverted and it behaves as expected. the quad flips over.
if throttle is inverted before then its also inverted after and the quad falls 'actively'.
but the same would happen when throttle is low on a normal quad except it doesnt fall that fast.

Good to know, thanks for the update!

[nicog]

New test:
http://www.youtube.com/watch?v=-_AKs2zxy7Y

Ok, i don't have a filmer so sorry for the outside passing.

[crashlander]

Hello!
Nice flying, hope someday I'll be able to do something similar!
What kind of props/prop mods are you using for it?

Regards
Andrej

[nicog]

7x4.5x3 no mod, so extreme crappy reaction on backside. I'm modding some props now. We will see what happens

[crashlander]

7x4.5x3 no mod, so extreme crappy reaction on backside. I'm modding some props now. We will see what happens

x3 means 3 blades or something else?
Any specific brand/make/model?

Regards
Andrej

[nicog]

3 blades.
They were (cause all dead now) hobbykings. Orange ones.

Now I modded some 10inch 2 blades. Cutted down to 8 and flatten.

[phenolic]

Using the GemFam "new airfoil" props might work for cutting down. Is that what you used? They have a large root near the hub which means lots of surface area to work with.

[nicog]

i use cheap props.

You talk about the graupner copy? You can't find it here in france. Only graupner and are expensive.

[phenolic]

Yes. As seen here:
http://forums.openpilot.org/topic/16854 ... 0x5-props/

Look at the size of the root. Although maybe it would be hard to flatten.

[nicog]

that are the ones warthox has flatten i think.
it's the graupner copy.


Latest test from 10 minutes ago.
Fly on back works now.

http://youtu.be/8mLkR96kavE

[ronco]

https://vimeo.com/62015977

[nicog]

Better to use a fart iron (for ski farting) from TOKO. taht is my weapon of choice.