APM Copter V3.2, bugs/issue tracker before the first beta release exit

[Marco Robustini‏]

Hi all, I'd like to focalizare the attention of testers in this thread in order to take stock of the situation on the APM Copter code, that Randy is evolving in a superb manner.
Too many threads that are creating a bit of confusion imho.
I therefore ask the testers to do a report here of any problems identified by testing the latest commits.
Personally i have not encountered relevant problems, just an annoying fast twiching to the motors during all the nav function (Auto, Guided, RTL, etc.), already reported to Randy.
EKF seems to give good results, although personally with the advent of the Onion discloses a slight "lag" in the fly in Stabilize, but it can be just my feeling.
I wait your feedback here, thanks in advance!

Bests, Marco

[Robert Lefebvre‏]

Marco, I thought that the settings I pointed out to you with regards to Stabilize sharpness helped the lag?

Rob

--
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.
For more options, visit https://groups.google.com/d/optout.

[Marco Robustini‏]

Yes Rob, i've tried but the feeling imho is a little different.
But I think that about 1/100 users probably will notice this thing.

Marco

[Matthias Badaire‏]

Hi Marco,

I will have some time this week-end (if weather permits).

It would be nice that people checks the existing list of issues for 3.2 before submitting to avoid duplication 

https://github.com/diydrones/ardupilot/issues?milestone=4&page=1&state=open

Thx

[Roberto Navoni‏]

Hi Marco i can doing some test with my micro quad it is very usefull
for testing any kind of new functionality at very low risk for the
frame :)
My opinion and the opinion of Emile ,too is about the EKF log , so
it's a lot could be better if is possible to activate and de activate
by standard method in cli non always active .
This first my 2 cent about new revision of code .
best
Roberto

[Xerr Avon‏]

Hi,

Only one thing I have seen a couple times with different copters running 3.1.3-RC1 and newer 3.2-dev versions is a small "twitch" (the number 2, and 4 motors kind of  drop quickly then return to normal) when switching from loiter to land. it hasnt caused me any trouble and I haven't looked into it at all so I am not sure what is causing it.  I just thought I would mention it.  If anyone wants I can see if it happens again and save the log to post, let me know.

I have been using EKF too, seems good to me so far.

John

[Randy Mackay‏]

John,

 

     Thanks for the feedback.  I’m having a look at twitches at the moment and it’s on the issues list as well in a couple of forms:

        https://github.com/diydrones/ardupilot/issues/970   ß twitch when entering RTL below RTL_ALT

         https://github.com/diydrones/ardupilot/issues/1034   ß twitch when entering loiter

     ..and some of the hybrid testers also reported a twitch when entering loiter mode from Hybrid.

 

     I fixed up some issues yesterday which could have caused twitches.  Julien (one of the developers of Hybrid) has a fix as well that I may apply.

 

     For those really into the details, the twitch is probably caused by the sudden jump in the target position from the previous mode’s final target position to the new mode’s target position which is calculated using a “projected stopping point” which is based on the vehicle’s current velocity.  In a windless environment they should be almost exactly the same but in real life they may not be.  Julien’s suggested fix is to simply use the previous mode’s final target position instead of using the projected stopping point.  I agree but just want to be sure that we understand the cause and that the fix works in all situations.

 

-Randy

--

[Peter Plischka‏]

Hi Randy,

the twitch when switching from hybrid to Loiter and vice versa is still there.


Peter

[Randy Mackay‏]

Peter,

 

     Thanks for testing this.  I was keen to know if the twitch was gone or not and it’s very clear from the logs that the twitch is in the throttle.  This is totally consistent with what Julien found.

     Fyi, looks like you were switching between AltHold <-> Hybrid but the twitch is there so still very useful!

 

-Randy

 

From: drones-...@googlegroups.com [mailto:drones-...@googlegroups.com] On Behalf Of Peter Plischka
Sent: May 1, 2014 6:48 PM
To: drones-...@googlegroups.com
Subject: [drones-discuss] Re: APM Copter V3.2, bugs/issue tracker before the first beta release exit

 

Hi Randy,

the twitch when switching from hybrid to Loiter and vice versa is still there.


Peter

--

[Matthias Badaire‏]

Marco / Randy ,

I did some test at the field today. TBS Disco dji 920KV/OPTO 30A/Lipo 3s 4400mah with yesterday latest firmware

1/ Testing Hybrid - really cool new mode and it worked perfectly. I would prefer to have harder stop but I need to check which parameter to change.

2/ EKF - I did not see any change of behavior . it just works. I used CH8_OPT=20 to activate / deactivate EKF

3/ twitch between loiter and hybrid is here , it is quite small though.

For some reasons, I had to set RC_FEEL to 100 multiple tie as well as I would stay to 0 as well. I manage to have RC_FEEL to 100 eventually. Check that MP takes your params before flying.

I reloaded some of my parameters in MP and LOG went to 0 for some reasons so I have no logs for the field testing. 

Now the less good part . 

Yaw Issue : I do not kmow what is up with this but this copter has been yawing perfectly before and now it is very sluggish only on yaw (pitch and roll are fine) in all modes tested (STAB, HYBRID, LOITER).

As I had no logs of the field testing, I flew in house for a few mins to generate logs.

I checked three times all the Yaw related parameters with no luck. Looking at the logs, it seems that DesYaw is reacting very slowly to my ch4 input.

I hope that it is not because I have set a wrong parameter and make you loose time.

3.1.3 was working fine on this quad.

Matthias

On Tuesday, April 29, 2014 5:30:15 PM UTC+2, Marco Robustini wrote:

[Matthias Badaire‏]

forgot the logs

On Tuesday, April 29, 2014 5:30:15 PM UTC+2, Marco Robustini wrote:

[Charles Taylor‏]

Take a look at the ATC rate, slew, and accel gains for yaw.  They probably need to be increased.  I'll let someone who knows more about them explain which ones and how to adjust. There are similar gains on roll and pitch.  It adds a big improvement for slow moving copters.  Like my 16" prop 3s quad.  Not very nimble but the new code made it fly like a dream.  Big step in helping smooth out camera ships.

[Matthias Badaire‏]

Hi,

I increased STB_YAW_P to 10 and now it is much better. I thought of this as AutoTune set 11.25  for STB_PIT_P and STB_RLL_P (which is above max values...).

For some reason, I did not had to do this in 3.1.x

M

--

[Craig Elder‏]

---------- Forwarded message ----------
From: Matthias Badaire <mbad...@gmail.com>
To: drones-...@googlegroups.com
Cc: 
Date: Fri, 2 May 2014 20:44:31 +0200
Subject: Re: [drones-discuss] Re: APM Copter V3.2, bugs/issue tracker before the first beta release exit

Ok,

I set STB_YAW_P back to 4.5 and change ATC_ACCEL_Y_MAX (which is new param in 3.2) to 90000 (default is 18000) and now I have the same yaw behavior as 3.1.3 with all the other parameters identical.

It means that the current default parameters do not give the same behavior as before.Fine with me but it might be worth writing here somewhere

thx Charles for the tip

[Matthias Badaire‏]

Hi,

I have had a battery failsafe as well during my flight using 3.2-dev and it switched to OF_Loiter , so says the bin logs.

checking the autotest, it has the same issue at line 33264

http://autotest.diydrones.com/ArduCopter-log.bin

I do not know if it is real or not. My copter seems to have try landing so it might only be a mavlink message error.

Should I log a gihub issues for those small problems ?

Thx

[Randy Mackay‏]

Matthias,

 

      I found that bug in mission planner and reported it to MichaelO.  It’s fixed in the current Beta version so the next official version should have the fix.

 

-Randy

[Matthias Badaire‏]

Thanks
I thought the mode was a string in mavlink and the message was issued by the firmware. My bad.

What do you think about ATC_ACCEL_Y_MAX default value that I talked about above. On my setup, the default value was not giving me the same behavior as 3.1.x. I had to bump it to 90000 (same as rp). Is it a problem of my setup or should we advise users to do the same ? Or default value should be change.

M

[Marco Robustini‏]

Hi Matthias, is not a problem with your setup, see my first post here.
Probably something has changed now with Onion, increased stability at the expense of responsiveness on inputs.
As I've said many times i think this thing is not tangible on some setup and in other yes, on my quad I noticed right away, I also made ​​a video to report it.

Bests, Marco

[Matthias Badaire‏]

Thx Marco
The sluggishness I found on my setup was really specific to yaw and changing that parameter fixed it.

As for twitching, I have some as well after doing an autotune with 3.2. My PIDs were completely different . I use the former PIDs calculated in 3.1 and now it is gone. I did not check if there was any change on autotune code or if it is my setup ( or too much wind the day I did it)

One last thing I found today as well is that ekf altitude hold is quite bad (it drops 3 or 4 meters when using ekf). I need to try some of the ekf parameter to see if I can improve it. I am reading the exchange between David and Paul to understand what to change.

M

[Randy Mackay‏]

Matthias,

 

     Re the ATC_ACCEL_Y_MAX, thanks for looking into that.  We should find the lowest value that results in the same performance as AC3.1.3 and use that as the default I think.

 

     By the way, did you try changing the SLEW_YAW value as well?  I was thinking that could also affect how quickly it can yaw.

[Peter Plischka‏]

Hi Paul,

I flew with two Copter the version with the patch "add protection for badly conditioned innovation variances."

The first flight there were several short drops and some fast Throttle Peaks.

Unfortunately, no log.

On the second flight I see in 9390 one  peak down.

In EKF = off everything is normal.
I'm trying to get a better log file. 

 

regards Peter

[Paul Riseborough‏]

Peter,

Thanks a lot for testing that for me, I don't get much of a chance to test during the week.

Can you confirm that you flew my master at commit 3db00da74c36b3c428ea1a919ad393d6bfef60a2

Regards,

Paul

[Charles Taylor‏]

Has autotune been updated, or going to be, for the ATC RP gains? I remember another post talking about changing autotune so it would set those gains to a certain percentage of what the copter is capable of.  Is there a guide available for tuning these manually? And will autotune take care of the rate/stab gains properly after setting ATC manually?  I have my quad tuned fairly well by hand but I'm sure autotune could do better.

[Xerr Avon‏]

Hi,

I had a problem today, I am kind of doubting it is a bug but I have flown this copter quite a few times with this FW ver on it with no problems but today is the first time I tried a 6s batt (usually use 4S)  I took off put it in althold, it was about 4 feet off the ground when it did an unexpected flip just perfectly landing on its feet. right before this happened I heard a tune play (maybe the gps glitch tone) and I did get a bad gps health on the hud but I am not sure if it is from before the flip or after.   right after it crashed/landed I heard the pixhawk start up tone, like it reset itself. My first thought was brownout but from what I can tell in the logs the 5v looks good.

I attached a log from the flight.

Not sure if it is a bug, which I doubt, but I thought I should post it just in case

John


On Tuesday, April 29, 2014 10:30:15 AM UTC-5, Marco Robustini wrote:

[Peter Plischka‏]

Yes Paul.

3db00da74c36b3c428ea1a919ad393d6bfef60a2

[Paul Riseborough‏]

Peter,

The EKF behaviour looks normal - some GPS jumps at the start (it had 5 satellites when EKF started), but otherwise OK, with no problems in altitude. See plot of demanded, EKF and baro alt.

However EKF didn't start running until after you were in the air because you took off before you had GPS. Then as designed, EKF was initialised 10 seconds after getting 3D GPS lock, and the code immediately started using it - this produced a step in measured height and vertical velocity and a corresponding throttle glitch.

The GPS locked up with 5 satellites and then managed to get 6 for the rest of flight. 

This is the first in-air EKF start I have seen from a copter log, so thanks for the interesting data.

Regards,

Paul

On Tuesday, 6 May 2014 00:44:56 UTC+10, Peter Plischka wrote:

[Charles Taylor‏]

Were you using the stock 3DR power module when flying 6s? I believe it is only rated to 4s.  I haven't looked at your logs but it was the first thing I thought of.

[Xerr Avon‏]

Hi,  I am not using the 3dr module, I am using whe SBEC from the esc's for 5v and a 180a auttopilot sensor for current and voltage monitoring.

thanks,

John

[Matthias Badaire‏]

Randy,

I tried multiple values for ATC_ACCEL_Y_MAX and I found that anything below 40000 was too sluggish. Starting at 50000, it was responsive enough and anything above did not bring much better results.

For information, this was in house testing (some props and furniture did not like it ..8) . I did not go to the field yet.

 SLEW_YAW description says that it is only for AUTO modes, my problem was already happening in STAB. I will do more test in loiter but is it worth trying changing it for STAB ?

Thx

[Chester‏]

Hello everyone,

couple days ago i decided to test 3.2dev on my px4 hexa. Copter was not realy sable. Althold didn't work. Didn't test loiter beacouse althold didn't work.

At althold it fires into sky. After senod fly (in same log) hexa become oncontrollable i reduce thotle and fall upside down (gps errors in logs occures after fall).

As i can understand i didn't used EFK so why px4 was so unstable and althold fires like a rocket?

[Marco Robustini‏]

Yesterday I did a lot of testing with the latest commit and the problem of "motors twiching" during Auto / RTL if EKF is enabled has not been resolved.
The problem is evident with SimonK ESC, where micre corrections are more tangible.
I confirm always "lag" in the ALE/ELE commands, the flying feeling in Stabilize is not like version 3.1, exacerbating some parameters I managed to make it quite similar but then significantly lowering STAB_P or gets really bad to maneuver.
I have not noticed any other particular problems, Alt-Hold and Loiter on my setup I would say perfect, better than the old release.

Bests, Marco

[Peter Plischka‏]

Marco, I can confirm your statements.

 

 "motors twiching" for all auto modes. This can not to be used for a camera flight.
The lag in Stabilize I do not notice.
Hybrid, Loiter and  AltHold are very fine.


regards Peter

[Svein Alexander Frotjold‏]

I Also notice some wierd unresponsivenes in ALE/ELE. (APM 2.5)

ATC_ACCEL_RP_MAX,180000

ATC_ACCEL_Y_MAX,80000

ATC_RATE_RP_MAX,250000

ATC_RATE_Y_MAX,200000

ATC_SLEW_YAW,1000

EKF N/A on APM

RC_FEEL_RP,75 (i suspect this is the cause)

What do you have RC_FEEL at Marco?

i'm gonna try it later at 100

[Peter Plischka‏]

Try 100, Svein.

[Paul Riseborough‏]

Marco,

Do you have a flashlog for a flight log where you have switched EKF on and off so it shows the 'twitching' vs not' twitching' behaviour?

I can then work out which control channel (height, yaw, roll/pitch) the twitching is being produced on so I can work back to the specific EKF measurement fusion that is causing it.

Regards,

Paul

[Paul Riseborough‏]

Peter & Marco,

So the twitching doesn't happen in loiter? If it was due to measurement noise coming through the EKF I  would it expect it to affect loiter as well. 

Regards,

Paul

[Peter Plischka‏]

Hi Paul,

I have the "motors twiching" in all the AUTO modes.
With and without EKF.

regards Peter

[Marco Robustini‏]

Hi Paul, give me the time and tomorrow i send you a flashlog with this "twiching".

Marco

[Peter Plischka‏]

Hi Paul,

here is a log with two short Auto Mission and RTL. 
V. "5e21111076a2d93a2977368ef549e0230eb15d93"

 

regards Peter

[Jaime Machuca‏]

Hi guys, 

Today I flew master with EKF, without EKF and in Loiter, Hybrid, Stabilize, Alt hold. Hybrid, Alt hold and Loiter all had a drop in altitude after forward flight, no mater what speed, when stopping the copter lost about 1m or so in altitude. Also there is a slight twitch (pitches forward) when switching between loiter and hybrid. This was with both EKF on and EKF off. Then when trying an Auto mission, the copter flipped on its own on takeoff right after activating Auto and throttle up. I have looked at the logs but I can’t figure out what happened on take off. Here are the logs for your review. 

https://dl.dropboxusercontent.com/u/42723480/EKF%20Crash%20on%20Auto.zip

Jaime Machuca 

--
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.
For more options, visit https://groups.google.com/d/optout.

<2014-05-21 16-34-38 IRIS FSC Auto.zip>

[Paul Riseborough‏]

DCM and EKF angles both agree. EKF data looks good with small innovations on all measurements so no problems there.

Demanded roll went to -35 degrees as soon as you switched to AUTO. It looked like the attitude control follows the demand initially, but then keeps rolling (maybe at that stage it caught a leg?). Question is, why was it demanding -35 degrees on entry into auto given the difference between demanded and achieved position was less than 0.5 m and the demanded velocity was less than 0.2 m/s ?

[Peter Plischka‏]

Hi Paul,

Here s Log with "4be98dddcbbb96027f905dde164a05faf7c2cde0".

    AP_NavEKF: Increase divergence test....

 

AutoTune and then fast flying.

regards Peter

[David Pawlak‏]

I'd get the same after RTL in hybrid takeoff. Any kind of take-off attempt after the RTL (flip). Resetting the machine gets things back to normal. Haven't tested this lately though, so might be different issue.

[Jaime Machuca‏]

Thats interesting, where do you see the demanded roll? is that on the NTUN message? I’ll try looking at the code see if I can figure something out, but I don’t think I will get very far… :)

Jaime Machuca Mercado

[Maurice Barnes‏]

I have had two incidents like this and it turned out that my Mission got messed up and the takeoff command ended up being after the first waypoint. So when the copter was set to auto it immediately attempted to go to the waypoint and a steep angle and flipped. Not sure if this is related but thought I would throw it in.

Regards,

Maurice

[Paul Riseborough‏]

The ATT message shows demanded vs measured angles.The NTUN message shows demanded vs measured positions and velocities.

[Jesus Alvarez‏]

I have had the same problem as Maurice states.

Mission planner has messed up some missions this past weekend.

If I upload a survey mission with DO_TRIGG_DISTANCE and auto take off and land and just after saving it, I try to read it back, It gets messed.

And then if you try to load it again, it messes it again , etc etc.

APM Planner read mission from copter correctly, but Mission Planner was always messing waypoints or missing the last LAND waypoint

I think there is a bug in either Mission Planner or Pixhawk (APM).

Will try to debug mission planner...

[Xerr Avon‏]

Hi,

I am not sure if this is the place to post this, I am running 3.1.5-RC2 on a Hexa with a PX4 and PX4IO. I had a kind of flyaway. the copter was about 5 feet high when it droped, I think it toched ground then took off strait up, not sure how high but looked like about 100 feet. I had no control so I shut the tx off to try to get an RTL. I m not sure if it stoped before or after I shut the tx off but it did and it hovered in place. I turned the tx back on and still no control (this is because I had the switches not set so the tx did nto really turn on.  When I noticed the switch warning and corrected it the copter fell, it seemed like I had no control but how knows I was kind of panicking at this point, it fell pretty close to where it took off from and really not too much damage for falling that far.

what I think happened is, I had been having problems with a loose battery connector so instead of replacing it I just bent the pins (I will replace from now on) so because the copter fell at first I think it may have been a brown out but I am not sure, ad if it was that loose I would be surprised it would have reconnected and stayed for so long.

I really dont know what happened, I do see a fence breach happens on the hud but I never got any RTL as far as I can tell. I will post the logs for anyone that wants to check them out, I would love to know what happened if possible.

the dataflash logs are kind of weird, the last several are really small like 1k so I will include up to the last one of any size.

thanks,

John

[Paul Riseborough‏]

Bug in acceleration between waypoints.

Flew master at cde7d31dadabe1846a422e3c5b9c94d3b0c16700  +  a minor EKF default parameter patch this morning in a 3DR quad. I took-off in auto, flew the mission, put it into a loiter at the end of the mission when it started its landing, restarted the mission using mission planner and sent it off again, When it finished the mission the second time, I restarted the mission, but this time did not put into loiter first. When it flew the mission for the third time I noticed that it was gaining speed and braking much more aggressively between waypoints (looked quite good actually). Here's a plot of demanded North and East velocity.

You can see the higher (higher than specified by the parameter) acceleration in the third section.

The flash log is attached,

Regards,

Paul

[Peter Plischka‏]

Hi together,

I had during testing with my T810 a sudden crash.
After 7 minutes the copter suddenly fell from the sky.

In the logs I can not see what the cause was.
Maybe one of you can look over.


regards Peter

[David Pawlak‏]

Looks pretty good to me! I see that there was a descent from about 80m and you tried to compensate for it. But it "looks" like the landing was without much damage.

It would be a shame, that's a pretty nice looking hexa.

Which battery monitor are you using. I guess the correction constants are all well calculated. You don't have battery failsafe set, and, if your constants are good, you should have still had battery (something like 3.45 on a 5S battery), If you go by the parameter settings, you are drawing 50 amps and have a capacity for 5000mAHr. Gives 6 minute.

The only thing close and that could well do what you are seeing is a weakening battery. But the numbers don't really jive.

[Peter Plischka‏]

Hi David,

I measure with the Attopilot 90 The voltage was set exact.
After the crash there were still 65% voltage on the 10,000 mA / h battery.

The damage was not as high. GPS demolished Gimbal defective and broken an arm.

Much worse is I do not know what happens.
If I find an error, I would have a better feel.

regards Peter

[David Pawlak‏]

It's a shame. Must be pretty hard to break one of those arms. I don't see a crash detect... I guess it didn't tip over the "magic mark"

It sounds a litle more reasonable with 10,000 ma / h you should have a voltage of more like 17,75 after 6 min. But that is  also more consistent with what I see.

Your starting voltage is 20.3, and a normal start for a 5S is 21.0. so translating the offset to the final voltage, you´re where you should be.

If you started partly charged, that would show a sag a lot more pronounced on initial start up, so my feel is that maybe a slight adjustment in the voltage multiplier is require. About (21 / 20,3) = 1.034 more.

In any case, doesn't explain the loss. At least I can't see anything else.

[Robert Lefebvre‏]

Peter, did the motors shut off in the air and it just fell like a rock?

--

[Peter Plischka‏]

Hi Robert,

Unfortunately I was a little distracted.
I suspect now also the sync problem. The controllers are Turnigy Kforce 40A BL-Heli.

I have just tested on the ground. 20 times the motors start immediately and then suddenly a motor 

stutters 

I had Demag=off Timing=high.

Now ich have Demag=high Timing=high

I test 50 Times and no stutters. I think that was the mistake.

regards Peter

[Robert Lefebvre‏]

What is the Demag?

--

[Peter Plischka‏]

Some motors can be very difficult to drive, because it takes a long time for current to decay after a commutation. And current must decay, in order to read motor position by sensing the voltage on the now non-driven winding.

This decay is often called the demag time. Pancake motors are notorious for having long demag times.

Peter 

[Peter Plischka‏]

If you started partly charged, that would show a sag a lot more pronounced on initial start up, so my feel is that maybe a slight adjustment in the voltage multiplier is require. About (21 / 20,3) = 1.034 more.

In any case, doesn't explain the loss. At least I can't see anything else.

I check the voltage again David.
I belief more and more that there is a sync problem of ESCs.
  With the new settings I have now over 100 times given full throttle without stuttering.

regards Peter

[David Pawlak‏]

OK, great... Nice hex BTW. 

[Robert Lefebvre‏]

Interesting.  OK, so what does the Demag setting in the ESC do exactly?

Is it the induction in the windings that causes the demag to be long?  Or is the problem actually in the iron core?

--

[Peter Plischka‏]

Sorry Robert, I do not know. I have not yet employed extensively with the ESCs.

 

I'm just looking for the cause of my crash. 
If I know why that happened, I'm already satisfied.
But it looks like the sync problems. 

    with "Demag = high" they no longer occur. 

 

regards Peter

[r. k.r‏]

hi Robert

notes taken from BLheli manual (https://github.com/bitdump/BLHeli/blob/master/SiLabs/BLHeli%20manual%20SiLabs%20Rev11.x.pdf):

Demag compensation is a feature to protect from motor stalls caused by long winding demagnetization time after commutation. The typical
symptom is motor stop or stutter upon quick throttle increase, particularly when running at a low rpm. As described earlier, setting high
commutation timing normally helps, but at the cost of efficiency. Throttle change rate can also be reduced, but at the cost of slower throttle
response.


--Richard

[Randy Mackay‏]

 

      I had a quick look at Peter’s log and I wonder if it’s the correct log.  It all looks fine.  Attached is a couple of graphs, the first is desired Roll & Pitch vs actual roll and pitch and it’s all tracking quite well.  The 2^nd graph is the Desired Alt vs Actual alt and again they’re all together and it shows a pretty reasonable climb and decent.    Nothing that jumps out to me as a crash.  Am I missing something or maybe we’re looking at the wrong logs?  It is an 8min flight and it’s nearly all in Hybrid mode.

 

-Randy

[Peter Plischka‏]

Hi Randy,
.
thank you. It is the right log. I can not understand, everything looks normal. The software has not made any mistakes.

I can only imagine that it was a sync problem. Maybe it was too windy above the copter had to compensate the position too quickly.

The copter is ready again and I'll try it out.

FYI: the ESCs are Turnigy Plush 40A with BL-Heli firmware
I've attached an Excel Sheet from the programmer of the BL-Heli software.
This could be interesting for some.


regards Peter