AC3.1-rc1 now available in Beta firmwares via MP

[Randy Mackay]

Beta Testers!

     We've just pushed the 1st release (AC3.1-rc1) into the beta downloads area.  The changes are not as fundamental as the last release but include the following improvements over 3.0.1
1) Support for Pixhawks board
2) Arm, Disarm, Land and Takeoff in Loiter and AltHold
3) Improved Acro
       a) ACRO_RP_P, ACRO_YAW_P parameters added to control speed of rotation
       b) ACRO_BAL_ROLL, ACRO_BAL_PITCH controls speed at which copter returns to level
       c) ACRO_TRAINER can be set to 0:disable trainer, 1:auto leveling when sticks released, 2:auto leveling and lean angle limited to ANGLE_MAX
       d) Ch7 & Ch8 switch to set ACRO_TRAINER options in-flight
4) SPORT mode - equivalent of earth frame Acro with support for simple mode
5) Sonar ground tracking improvements and bug fixes that reduce reaction to bad sonar data
6) Safety improvements
       a) motors always spin when armed (speed can be controlled with MOT_SPIN_ARMED, set to 0 to disable)
       b) vehicle's maximum lean angle can be reduced (or increased) with the ANGLE_MAX parameter
       c) arming check that GPS hdop is > 2.0 (disable by setting GPS_HDOP parameter to 900)
       d) slow take-off in AUTO, LOITER, ALTHOLD to reduce chance of motor/esc burn-out on large copters
7) Bug fixes:
       a) Optical flow sensor initialisation fix
       b) altitude control fix for Loiter_turns mission command (i.e. mission version of CIRCLE mode)
       c) DO_SET_ROI fix (do not use "ROI")
8) Distribute Loiter & Navigation calcs over 4 cycles to reduce impact on a single 100hz loop.  This may resolve some APM1 users alt hold issues.

All feedback welcome!  Remember it's release candidate #1 so there will be at least 2 more release candidates.  The most important things we need to get done before the big release is the GPS glitch protection (assuming we can get it to work) and the TradHeli improvements.

-Randy

[Robert Lefebvre]

Hi Randy,

I guess 6d) is related to the problem you were telling me about last night?  Good idea what you've done.  Have you made it so that the throttle_accel PID controller will not ramp up the I-term while this is happening?  Will this change affect TradHeli?  We don't really need it since we have our own motor start feature.  But, this might actually be nice anyway.  There is often a pretty good jerk when doing an auto-takeoff.  The auto-throttle controller holds the collective pitch down while the motor is spinning up, and then raises the collective sharply once the motor_runup_complete flag is set.  I should try and implement something similar to what you've done here.  So it will do the motor runup, but then slew the collective up gently.

Just a nice-to-have feature.  Icing on the cake of auto-takeoff.

--
You received this message because you are subscribed to the Google Groups "ArduCopter Testers Team" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arducoptertest...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Gary McCray]

Hi Randy and all involved,

Congratulations to you all.

I know you have all spent a lot of time working to ensure this release fixes a lot of loose ends, improves safety and adds some great new functionality like Sport, takeoff and land in Loiter and Alt Hold and the Pixhawk.

I think the previous version worked remarkably well and this should make it even better.

I will start adding the necessary information to the wiki (with temporary Beta note).

Best Regards,

Gary

[Robert Lefebvre]

I think Note 3 should be expanded a bit, as the biggest part is not covered.  IMO, this is the most important part of 3.1, but maybe I'm biased. ;)

3a) should be:

New body-frame acro mode allows full 3D acrobatic capability.  Flips and rolls are possible. Not just a simple rate controller, but features error tracking and correction in body-frame. 

Ok, here's my write-up for what the Tradheli changes are/will be.

9) TradHeli Changes

a) Major pitch/roll rate controller improvement.  Eliminates excessive filtering on the rate feedback into the PID control loop.  Will greatly increase the responsiveness to rate errors.  NB: Will require PID retuning!  D-term will likely be unusable, please set to Zero before attempting take-off.

b) Dynamic Flight detection automatically determines if heli is in fast forward flight and changes to a non-leaky-i-term for better control responsiveness.

c) Auto Mode yaw behaviour now defaults to Yaw-Look-Ahead.  Heli will yaw into the direction of flight automatically.

d) When in Auto Mode, down collective pitch is limited to Stab_Col_Min until takeoff is complete.  This avoids excessively pushing the helicopter down onto the ground.

e) Yaw Rate controller uses Leaky-I-term until motor has run up.  This helps avoid excessive I-term buildup when on the ground.

f) add check to Auto_Armed status to wait until motor runup time is complete before auto mission will begin.

g) add support for Direct Drive tail rotor control.  Both fixed pitch with variable speed, and variable pitch with fixed speed (governed ESC) are possible. Caution: Avoid conflicts with generic Ch7 Output functions!

h) bug fix to External Gyro control.  This function may work properly now but unconfirmed.

i) Collective-Yaw compensation does not react when main motor is not running.

j) Piro-Comp and Acceleration-Cross-Coupling code added, but not yet functional on APM due to insufficient system resources.  

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Leonard Hall]

Yeh, I would agree with Robert here. It isn't really an improved ACRO it is a completely new ACRO. I would also add a warning along the lines of "ACRO lets pilots crash copters, if you can't handle crashing your copter quiet a few times then learning ACRO probably isn't for you"

I am personally very happy with this release :)

[Craig Elder]

Please note:

6) Safety improvements

       c) arming check that GPS hdop is > 2.0 (disable by setting GPS_HDOP parameter to 900)

should read  GPS hdop is < 2.0

On Fri, Sep 13, 2013 at 4:42 PM, Leonard Hall <leonar...@gmail.com> wrote:

Yeh, I would agree with Robert here. It isn't really an improved ACRO it is a completely new ACRO. I would also add a warning along the lines of "ACRO lets pilots crash copters, if you can't handle crashing your copter quiet a few times then learning ACRO probably isn't for you"

I am personally very happy with this release :)

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Robert Lefebvre]

Though I found Stab as a bailout worked totally fine.

On 13 September 2013 19:42, Leonard Hall <leonar...@gmail.com> wrote:

Yeh, I would agree with Robert here. It isn't really an improved ACRO it is a completely new ACRO. I would also add a warning along the lines of "ACRO lets pilots crash copters, if you can't handle crashing your copter quiet a few times then learning ACRO probably isn't for you"

I am personally very happy with this release :)

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[john...@gmail.com]

Not to sound elitish but ACRO (unlimited rate mode) just means you actually have to fly the thing like every R/C pilot has done for decades..

[Leonard Hall]

Hehe, Exactly, but those RC pilots learned to do it by crashing. The problem is that there are many users who expect the code to be crash proof. So much so I feel we should point out to them that ACRO deliberately lets you write off your frame with all the fury your motors can muster upside down at full throttle :)

[john...@gmail.com]

That is true. In my opinion this false sense of security and belief that the system will always save you cause more danger then it does good. People are sidestepping basic learning curves making them totally reliable on a 100% working system at all times. The sense of security also makes some want to push boundaries leading to more dangerous situations when (not if) something go wrong.

And the sad part is, that by never trying and learning ACRO they are missing out on probably the most fun you can have with a copter. Flying ACRO is much more engaging and satisfying in the long run. You can do much more complex and aggressive maneuvers (like fast dives by gliding down at aggressive angles to avoid flying into your own turbulence) and the feeling of mastery if very rewarding. Just pulling sticks or clicking at a screen to point the copter in a direction gets old, real fast.

[Robert Lefebvre]

Preaching to the choir brother...

--

[Jonathan Challinger]

Can we do an altitude floor recovery maybe? Predict based on angle when the copter will hit the ground and go into recovery if it's going to happen.

[Leonard Hall]

Yeh we could, this bloke had done a very good job of implementing it.

http://www.youtube.com/watch?v=xIrLBZttYAU#t=296

[Jan Willem van den Barg]

Where are we supposed to post questions and observations about the 3.1 rc1 firmware?
I see posts appear on the DIY Drones forum in the 3.01 thread and in this google discussion group.
Maybe it would be good to ask people to post only in this discussion group.

Best regards,

Jan Willem
The Netherlands

[Marco Robustini]

As I already reported the parameter "MOT_SPIN_ARMED" has a maximum range too low for some motors with Simonk ESCs and 3s LiPo, i mean with this RC1, also with 3.0.1.
"130" is the max value, the default is "70": here with "70" the motors doesn't start, and with "130" (if you write the parameter and then read the value go to "127") the pwm is too low.
In my video (PX4) one motor runs very low, almost stopping, the parameter is "127":

https://www.youtube.com/watch?v=CMMmS9fL-UU&feature=c4-overview&list=UUPcCCiqkWYZmvxVgHhDZDHA

As I had already mentioned to Randy would be useful to have a maximum value configured at least "200" or "250", now is currently unusable, please fix it!

Bests, Marco

[Randy Mackay]

Marco,

     Ah right.  We made it an Int8 when it needs to be an Int16.  We will fix that for -rc2.  Thanks for the report!

-Randy

---

From: Marco Robustini <robusti...@gmail.com>
To: drones-...@googlegroups.com
Cc: ArduCopterTesters <arducopt...@googlegroups.com>; Randy Mackay <rmac...@yahoo.com>
Sent: Saturday, September 14, 2013 8:06 PM
Subject: [drones-discuss] Re: AC3.1-rc1 now available in Beta firmwares via MP

[Robert Lefebvre]

That would definitely be pretty cool. 

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Jan Willem van den Barg]

Tested firmware 3.1 RC1 on my quadcopter today.
After performing live compass calibration I tested the following:

• Arming and disarming in Loiter mode --> works perfect.
  
• TakeOff and Landing in Loiter mode --> works perfect.
• Flying around in Loiter mode --> works perfect, just like 3.01.
  
• Motors spinning when armed --> only three of my motors spin when armed. I use SimonK flashed ESC's. Same problem as Marco I guess.
  
• Tested several missions, differing in complexity --> all executed as planned.
• Used circle mode / loiter turns in a mission --> works perfect.
• Tried to use DO_SET_ROI in a mission --> couldn't find the DO_SET_ROI command in the drop down list. SET Roi only adds an empty waypoint to the list, but no symbol on the map. Seems to be a omission in Mission Planner.
  

Best regards,

Jan Willem

The Netherlands

On Friday, September 13, 2013 4:01:07 PM UTC+2, Randy Mackay wrote:

[Robert Lefebvre]

Leonard, I wonder if it could be as simple as an option where, you can change to Acro above a floor (like 20, 30, 40 feet), and then fly however you want, but it you fall below the floor, it automatically switches to Stabilize mode.  Maybe we'd need to lock-out the user input for a bit, because they'll be busy on the sticks.  I wonder how reliable it would be?  We won't be running INAV, correct, so we'd be relying completely on the Baro?

I think it would be easy to implement, as we simply don't allow Acro mode below the min altitude.  So you take off in Stab, fly up, flip to acro, lose it, and it will flip back to Stab automatically.  Maybe not lock out the inputs, but just tell users to let go of the stick for optimum recovery.  And give it throttle.

The better helicopter flybarless controllers are coming out with this feature called "hard deck".  I haven't seen multirotor controllers coming out with it yet however?

On 14 September 2013 01:01, Leonard Hall <leonar...@gmail.com> wrote:

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Randy Mackay]

     That sounds a lot like a minimum altitude for the Fence so it might be best to include it there...

-Randy

---

From: Robert Lefebvre <robert....@gmail.com>
To: drones-discuss <drones-...@googlegroups.com>
Sent: Sunday, September 15, 2013 10:02 AM
Subject: Re: [drones-discuss] Re: AC3.1-rc1 now available in Beta firmwares via MP

[john...@gmail.com]

Yes please make the altitude ACRO override part of Fence. It makes sense and also makes it possible to fly ACRO all the way to the ground for better or worse if you want to.

[Leonard Hall]

We could simply use the height that we switch to ACRO as the height where we switch to Stabilize minus 1m. But the question is how would we implement it. We already have 3 acro levels of trainer and I don't think this should be active for full ACRO without able to be fully disabled. Maybe we could simply add a parameter in the mission planner that enables or disables it, or as Randy suggested make it a Fence option.

In any case it is very important that the user can choose to not use this feature. The last thing we want is the ACRO pilot says this is good for beginners but you still can't fly in full ACRO.

[Robert Lefebvre]

Yeah, absolutely, I wouldn't want it to be always active.  Or... if it was, give it it's own parameter that can simply be set to zero.

I wouldn't like it to be part of fence, meaning you'd have to have fence on to use it.  I don't like to use fence, but I would like the hard deck.

[Marco Robustini]

Hi Randy, thanks for the RC2 update and this fix:

https://github.com/diydrones/ardupilot/commit/d3fcfe669d9c099bdf4837cce2da62ac3a9a0b72

Anyway bad news, not solve the problem, same issue, i use "150" and the motors start too slow like in my old video, too bad.
PS: from 100 to 150 i can't see any difference, have you checked the pwm output to the motors?

Bests, Marco

On Friday, September 13, 2013 4:01:07 PM UTC+2, Randy Mackay wrote:

[Olivier ADLER]

I did get a flip during flight and a crash with version 3.02 pre RC1.

I think it is related to this new feature.

During this crashed flight, i did use "MOT_SPIN_ARMED" = 0 because i wanted the older behavior.

On the ground when motors where armed and running at slow speed i did notice that if i did bank a bit the copter, one or more motors did stop.


I think that something is wrong at this level. I did keep the log if necessary. I'm waiting for a new frame so that i can make new tests (this time at a lower altitude !)


Olivier

[Jonathan Challinger]

if your motors are stopping from input your min throttle is way too low and/or you need to recalibrate your escs. can't rule out a bug though.

[Olivier ADLER]

I never had this problem before and i did not change anything to the setup except the firmware version. I will make more tests as soon as possible to make things clearer.

Something is sure, this is the first time i did saw a couple motors stopped on the ground and other ones running. My min throttle is set to the default (130) and i have the default non reflashed ESCs from 3DR.

It seems to me that the "MOT_SPIN_ARMED" = 0 parameter is causing a problem at least on the ground when armed and rotating slowly. We must not see motors stopping when banking the copter on the ground. This mean that the throttle value for those motors is falling below 130.


I would be interested to know the results of other testers when using "MOT_SPIN_ARMED" = 0.


Olivier

[Randy Mackay]

Marco,

     The MOT_SPIN_ARMED is limited by the THR_MIN.  So if the THR_MIN is only 150 then you can't make MOT_SPIN_ARMED higher than 150.

     I'm pretty sure it works properly but I can double check.

-Randy

---

From: Marco Robustini <robusti...@gmail.com>
To: drones-...@googlegroups.com
Cc: ArduCopterTesters <arducopt...@googlegroups.com>; Randy Mackay <rmac...@yahoo.com>

Sent: Friday, September 20, 2013 3:53 AM
Subject: Re: AC3.1-rc1 now available in Beta firmwares via MP

--
You received this message because you are subscribed to the Google Groups "ArduCopter Testers Team" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arducoptertest...@googlegroups.com.

[Marco Robustini]

In my latest tests with PX4 and the RC3 i'm noticing that during the takeoff in Loiter, when more than half of the throttle stick the pwm to the motors is too fast, I was asked to put some sort of "linear ramp" of about one second to make it a little smooth, currently the multirotor off the ground too hard.
With small quad thing you do not notice so much, but with heavy configurations, where the hovering throttle is between 30 and 40% this thing is horrible.
Randy and Leonard, please watch the video of takeoff in position hold with Naza or WooKong, in that case this "ramp" is present and the multicopter off the ground fairly gently.

Regards, Marco

[Leonard Hall]

Randy added a 1 second ramp to full throttle but maybe it should be a 1 second to hover throttle ramp instead. That should be much more consistent from copter to copter.

--
You received this message because you are subscribed to a topic in the Google Groups "drones-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/drones-discuss/L9vlwqYGijo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to drones-discus...@googlegroups.com.

[john...@gmail.com]

Not sure I followed the last question completely, but when taking off under difficult conditions with stuff like uneven ground and/or side winds. You need to be able to "punch" the throttle for a quick takeoff.

- JAB

[Peter Plischka]

Thank you for the good work.
I fly every day with the latest trunk from the master.

The versions of the past few weeks flying extremely stable.

Here's a video from today.

http://www.youtube.com/edit?video_id=eD60K1rVDYs&video_referrer=watch&ns=1

The video shows the X600 with 40 A controllers, HP2814-710KV motors (97 g) and 10-inch propellers fitted.

The battery is a 6S 6000 mA / h LiPo, the flight time is 17 min 30 sec

Regards Peter

[Peter Plischka]

Forgotten:

The new Arduino gcc 4.7.2 compiler is also great.

"Binary Sketch Size: 210 048 bytes (of a maximum of 258,048 bytes)"

Peter

[john...@gmail.com]

Fixed the youtube link:
http://www.youtube.com/watch?v=eD60K1rVDYs

[Olivier ADLER]

3.1 RC2 test report :


I did test 3.1 rc2 on my 3DR Y6.

- 4S 4500 mAh batteries

- 10x4.7 APC SF propellers

- Simonk 3DR ESCs

- Compassmot value : 7%


All seems to work nicely. The Simonk ESCs seems to enhance stability. it's really better than with my older Hexa B without Simonk.


I did test Stabilize, Loiter, Position-Hold, Alt_Hold, Simple.


Yaw was a bit unstable, so i did rise Yaw rate P from 0.15 to 0.17. Seems almost OK now. (my frame is perfectly straight, but all my motors are not new, i have 4 old motors with tired ball bearings and 2 new motors)


At the same time, i did change those parameters that should be changed i think for the release :


LAND_SPEED : from 50 to 75

PILOT_VELZ_MAX : from 250 to 500


PILOT_VELZ_MAX is the most important to change, the actual setting can be dangerous specially when the baro gives under or overpressures during forward flight.


LAND_SPEED is a bit slow actually, if you are landing from a 15 meters RTL altitude, it takes to much time to go on the ground.



Olivier.

[Olivier ADLER]

Something i'd like to emphasis on is the Alt_Hold dead band.


It is certainly nice for beginners, but for anything else it is really to large, specially when using a mode 5 RC transmitter (throttle stick with return to center spring). It's a shame to have such a nice Alt_hold mode and a so high dead band ruining the altitude control precision we can have in this mode.


Would it be possible to add a parameter to set this dead band ?


A solution would be to use the RC3_DZ parameter, actually it is not used.


Olivier

[Robert Lefebvre]

I agree, would be nice to have this as a parameter.  I'd prefer we make a proper parameter for it however.  We've run out of processing, but we have some Memory left, might as well make good use of it?

At the same time, would be good to fix the issue where the Alt_Hold climb rate is also limited by Thr_Max, and it should not be.

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Randy Mackay]

     Sure, Leonard and I have discussed adding this in the past because some people want it bigger and some want it smaller.

     There's too many change requests coming in each day for me to keep track of them all so if you could add it to the issues list that would be great..

-Randy

---

From: Robert Lefebvre <robert....@gmail.com>
To: drones-discuss <drones-...@googlegroups.com>

Cc: ArduCopterTesters <arducopt...@googlegroups.com>; Randy Mackay <rmac...@yahoo.com>

Sent: Wednesday, October 2, 2013 2:01 AM
Subject: Re: [drones-discuss] Re: AC3.1-rc1 now available in Beta firmwares via MP

[Olivier ADLER]

Done, Issue #586

Olivier

[Robert Lefebvre]

I just created 591 about THR_MAX and DCR.