Copter jerks extremely

[SoundMTB]

Hello,
I have the following problem with my copter:
I use the software MultiWii_dev_20111029
The gyros and the ACC is calibrated
The Coper floats 100% without problems
When only pitch or roll, everything is OK
However, if one wants to fly properly and uses all the control functions, the copter jerks extremely
especially in stick movements with YAW

I'm not looking further, already tried everything possible: for example, propeller balancing, engine balancing, set PID values, etc.
What can be causing this?

The second problem is:
I get the 1.9 software does not function.
It is impossible for the engines to start with the stick.
What am I doing wrong?

regards
Oliver

[Noctaro]

Hi,
i think you are talking of I2C glitches.
As these errors seem to be fixed with the new 1.9 release, i strongly recommand you to use this one.
If you upload the new 1.9 version, you also have to use the 1.9 GUI to get feedback.
Did you make the right changes to config.h before uploading?
btw. what sensors do you use?
The arm motors procedure is the same as ever. But you may change this by applying the ARM checkbox to one of your AUX channels via GUI.

Hope this helped a bit.

greetz
Noc

[UndCon]

I get the 1.9 software does not function.
It is impossible for the engines to start with the stick.
What am I doing wrong?

regards
Oliver

The problem is that Alex uploaded fixes a few times and didn't change version number from 1.9

actually I think in current stat it should be:

1.9 release
1.9.1 (first bugfix)
1.9.2 (second bugfix release)

I think we are at 1.9.2 or even 1.9.3 atm...the latest is a minor fix only with with the latest i2c debug values enabled for GUI presentation

Solution is to download 1.9 again and make your needed adjustments to config.h - compile and upload to Arduino

Try again and see if i2c errors appear in GUI debug value (a counter that should remain at 0)


If you have a LCD you can also monitor i2c errors there - very useful in the field.

//UndCon

[SoundMTB]

Hi

I've mede tests with 1.9 to see the i2c error, there is nor error in my system the debug window shows 0.
I use the WII Motion Plus and BMA020
I use pin 12 for AUX2
the config.h in many times cheked by me and there isn't a fault.
Of course I use always the right GUI.

regards
Oliver

[SoundMTB]

SORRY I HAVE TO MAKE A CORRECTION

I've just test the 1.9 a few minutes ago
and I found a lot of i2C errors 20000 and more.

What I have to do now??????????

regards
Oliver

[SoundMTB]

Hi

I made the correction posted by Alex in the Sensors.pde in MultiWii_dev_20111029

The first flight was very good no vibes no jerks but at the scond flight I have again extream jerks when i control yaw,pitch and roll at the same time.

Oliver

[UndCon]

So you did find some errors...good or bad...
Did you download the latest version of 1.9 to find out?

earlier "twitchy" versions of 1.9 crashed my 1 GoProHD and Quad

Today I started to fabricate 6 more arms (4 goes to a new Quad to a customer) and 2 for my own Quad
All share same measurements and it's quite easy to make more when the tools are set up.

//UndCon

[mr.rc-cam]

SORRY I HAVE TO MAKE A CORRECTION
I've just test the 1.9 a few minutes ago and I found a lot of i2C errors 20000 and more. What I have to do now??????????

You need to fix your hardware. There are several recent I2C discussions about troubleshooting/solving I2C errors. For example, here is discussion for the WMP/BMA020 combo: viewtopic.php?f=6&t=935

I made the correction posted by Alex in the Sensors.pde in MultiWii_dev_20111029

That file is old and does not have any of the latest I2C code. So you will never see the I2C errors at the GUI's Debug2 window because the older versions do not report them. To troubleshoot the I2C problems you must use ALL the latest V1.9 files (dated Nov 12).

[SoundMTB]

Hi
I've read all the Postes reagrding WMP/BMA020 combo, I've attached pullup resistor 2k2 to the SCL SDA line an comand out the internal pull up.
I modified the I2C Code to monitor the Errors by my self. I din't use the 1.9 version becouse I'm unable to arm the motors by Stick bt I need thi funktion the Aux1 and 2 are blocked by other funktions.
regards
Oliver

[mr.rc-cam]

I modified the I2C Code to monitor the Errors by my self. I din't use the 1.9 version becouse I'm unable to arm the motors by Stick bt I need thi funktion the Aux1 and 2 are blocked by other funktions.

The older code does not have the I2C driver fixes that Alex incorporated into the Nov-12 V1.9. So you may continue to have I2C errors, even with the correct pullups.

One of the many reasons for not being able to stick Arm is due to a poor sensor calibration. Bad I2C communication can cause that. However, if you are using the Nov-12 V1.9, and the GUI's Debug2 does not show any I2C errors, then the arming problem is something else; For example, insufficient servo pulse range (which can be confirmed by observing the pulse values in the GUI). Bad sensors can cause this too.

So ignore the Arming problem. Does the Nov-12 V1.9 report I2C errors? Use the results from this test to direct your troubleshooting efforts.

[Niclas Hedlund]

I just found my jerk-problem today.
Everything was fine until i let the quad loose in the air.

I even held it with two arms and a friend gave it full throttle and I shaked it and really tried to get the problem to appear.
It sounded just like the ESC´s stopped for a short while and the quad started "jump/shake" in the air.

I spend some time with the scope today and found out that the i2c signals looked fine on the "3.3v side" of the 3.3<->5v converter, but they looked like a mess on the 5v side.

So adding two 2,7k resistors on the 5v side solved my problems! The weird thing is that it worked like a charm with 0 i2c errors when it wasn't free in the air.

So take a look with a O-scope on the signals!

[SoundMTB]

Hi
I have the same presumption will now measure time with the evening Ozziloscope the signals.
I used it generally 5V.

Oliver

[SoundMTB]

Hi

I've measured the voldage without motors and with motors no errors or peaks.
the VCC of arduino is 5,28V in all situations.
I've installet 2.2K; 2.7K and 4.7K pullup allhave the same resault no changes.

I can see some smal error cound in debug2 about 68 but I can'd fly with 1.9 i fly with dev 29.

My las option is to change the BL Controller to the motors.

I'VE NO IDEAR WHAT CAN BE THE PROBLEM!!!!!!?????????

regards
Oliver

[mr.rc-cam]

I can see some smal error cound in debug2 about 68 but I can'd fly with 1.9 i fly with dev 29.

If you see I2C errors reported in the GUI's debug2 window then you need to fix something. It could be due to the schematic you used, an intermittent solder connection, or a defective sensor. I'm not saying that having a zero debug2 count will solve ALL your problems, but ANY I2C errors can contribute to poor stability.

So before you start replacing other parts I think it would be wise to fix the I2C first. You should have zero errors after a 10 minute bench test, including when the motors are running. Keep in mind that the I2C test is very primitive, so it cannot test for bad performing sensors and it does not detect every type of I2C communication problem. So zero errors does not mean you don't have sensor R/W problems. But it is the best I2C test we have for now.

Also, you should use the GUI to watch the sensor data. If you are having jerky movement then I would expect that you will see the problem in the GUI data. This may help you ID a bad sensor if you have one that needs replacement.

My las option is to change the BL Controller to the motors.

Poor performing brushless motor controllers can cause stability problems. But they are just one of the many things on the list of possible causes. So be prepared for anything and everything.

[SoundMTB]

Hi

I've tested several Pullups 2.2k up to 4.7 no change the error is still there.
I've chaged the BL Controlers but nothing change.

I thnik I will put this copter into the trash

I realy din't know what is the problem.

regards
Oliver