I am new to the scene and on a steep learning curve. I would be most grateful for assistance in addressing my current hitch. I have fitted an Ardupilot 2.6 to a quad along with the 3DR compass/GPS unit. I have installed Mission planner 1.2.93. I am able to connect and install firmware and followed the wizard initially to run through the mandatory hardware installation. However I am unable to achieve a successful accelerometer calibration. I have watched the accompanying link and have tried quite a few times now without success. I have also tried calibrating without the wizard but continue to get the calibrate fail notice. I have also noticed several times when on the final position in the calibration sequence the procedure ends, without a key stroke and results in a failed calibration.
Thanks Madhamster. I did have an issue with a shifting USB cable the first couple of times and was disconnected. I have now taped it up so that it cannot move and it will no longer disconnect when carrying out the calibration.
I guess the question is whether the accel calibration routine is failing or whether there’s a communcation problem between the APM and the mission planner.
If it displays “Calibration Failed” (or something to that effect) on the Accel Calibration screen or on the HUD (the artificial horizon) on the Flight Data screen then it’s a problem with the calibration. If you’re not seeing that then it’s a communication problem.
It may be that there are extra characters being sent from the mission planner to the APM and that’s making it think that it should move onto the next stage of the calibration procedure before it’s ready. If this is the case it may be the USB cable (as suggested above). You might also try going into the CLI (terminal window in the mission planner), press Connect, then enter “setup” and “accel”. Then follow the regular procedure of pushing a key and changing the position of the vehicle.
Hi rmackay9
Thank you for your assistance. I have followed the method that you suggest, however to no avail. I do get the 'calibration failed message on the MP screen and on the HUD. I am going to buy another USB cable to check if that makes a difference.
ozyfly,
If it says “Calibration Failed” which I think is what you’re saying then it’s most likely that the accelerometers are out-of-tolerance which basically means that they’re leaning on the board or they’re broken. Normally if you use the CLI (aka mission planner’s terminal window) it will print a little message with the results which could be NaN or some numbers that show that it’s out of tolerance. could you post a screen shot of what it says at the end of the calibration?
I strongly suspect now that it’s not a USB cable issue. If it’s just barely out of tolerance then we could provide a modified firmware with a slightly wider tolerance…or if it’s a 3dr board you could probably get a replacement.
I have a new USB cable. Gets me away from the computer and screen. I have attempted setup via the CLI. The following is what I get. What does it all mean?
setup] acell
Invalid command, type 'help’
setup] accel
Initialising gyros
Init Gyro**Place APM level and press any key.
Place APM on its LEFT side and press any key.
Place APM on its RIGHT side and press any key.
Place APM nose DOWN and press any key.
Place APM nose UP and press any key.
Place APM on its BACK and press any key.
Offsets: 0.07 0.06 -5.16
Scaling: 1.00 1.00 0.97
Calibration FAILED
INS
I have APM 2.0 with firmware 3.1.1 and when I try to Calibrate Accelerometer always obtain a error message (Calibration FAILED) in GUI interface even CLI interface of Mission Planner (version 1.2.x and new version).
The same procedure with CRIUS AIOP v1 and MegapiratesNG 3.1-beta works fine with the same USB cable.
I have found through the raw sensor view that zacc seems to be broken.
All the values change when I move the controller as in accelerometre calibration sequence except zacc that alwais remain at aproximately -1000 even if controller is flipped upside down.
Could be a hardware problem? Although I have the board since two years ago I only have try to fly in my house’s garage. I could reproduce the problem with other firmwares.
There is an alternate to complete the calibration procedure?
Is the board broken?
What does I should do?
It could fly without z accelerometer or disabling accelerometer?
There are any solution?
There are some values obtained in calibration procedure.
I have APM 2.0 with firmware 3.1.1 and when I try to Calibrate Accelerometer always obtain a error message (Calibration FAILED) in GUI interface even CLI interface of Mission Planner (version 1.2.x and new version).
The same procedure with CRIUS AIOP v1 and MegapiratesNG 3.1-beta works fine with the same USB cable.
I have found through the raw sensor view that zacc seems to be broken.
All the values change when I move the controller as in accelerometre calibration sequence except zacc that alwais remain at aproximately -1000 even if controller is flipped upside down.
Could be a hardware problem? Although I have the board since two years ago I only have try to fly in my house’s garage. I could reproduce the problem with other firmwares.
There is an alternate to complete the calibration procedure?
Is the board broken?
What does I should do?
It could fly without z accelerometer or disabling accelerometer?
There are any solution?
There are some values obtained in calibration procedure.
I have a new USB cable. Gets me away from the computer and screen. I have attempted setup via the CLI. The following is what I get. What does it all mean?
setup] acell
Invalid command, type 'help’
setup] accel
Initialising gyros
Init Gyro**Place APM level and press any key.
Place APM on its LEFT side and press any key.
Place APM on its RIGHT side and press any key.
Place APM nose DOWN and press any key.
Place APM nose UP and press any key.
Place APM on its BACK and press any key.
Offsets: 0.07 0.06 -5.16
Scaling: 1.00 1.00 0.97
Calibration FAILED
INS
I have APM 2.0 with firmware 3.1.1 and when I try to Calibrate Accelerometer always obtain a error message (Calibration FAILED) in GUI interface even CLI interface of Mission Planner (version 1.2.x and new version).
The same procedure with CRIUS AIOP v1 and MegapiratesNG 3.1-beta works fine with the same USB cable.
I have found through the raw sensor view that zacc seems to be broken.
All the values change when I move the controller as in accelerometre calibration sequence except zacc that alwais remain at aproximately -1000 even if controller is flipped upside down.
Could be a hardware problem? Although I have the board since two years ago I only have try to fly in my house’s garage. I could reproduce the problem with other firmwares.
There is an alternate to complete the calibration procedure?
Is the board broken?
What does I should do?
It could fly without z accelerometer or disabling accelerometer?
There are any solution?
There are some values obtained in calibration procedure.
Yours is a hardware failure for sure if the Z axis always reads the same. There isn’t any easy solution that will keep you flying.[/quote]
Hi jschall,
After a reset & erase throw CLI I have seen that z accel varies in the interval [-2655, -2520] it could be possible to calibrate manually to transform this interval in [-980,980].
I have attached a file with the log of last checks I have done.
After a reset & erase throw CLI I have seen that z accel varies in the interval [-2655, -2520] it could be possible to calibrate manually to transform this interval in [-980,980].
I have attached a file with the log of last checks I have done.[/quote]
Yeah it can be done but you’re still getting 2 m/s^2 of noise. Don’t fly that APM. It is broken. Attached is a graph comparing to the Y accel.
If you want to try to set it manually, your INS_ACCOFFS_Z needs to be set to 40580 and your INS_ACCSCAL_Z needs to be set to 15.68. Keep in mind that the accel calibration routine will never work without code modification, and that the copter will probably refuse to arm with those settings. Who knows when or if it will fail completely and stop working or change again.
Contact help@3drobotics.com if your board is a genuine 3dr board and is under warranty.
After a reset & erase throw CLI I have seen that z accel varies in the interval [-2655, -2520] it could be possible to calibrate manually to transform this interval in [-980,980].
I have attached a file with the log of last checks I have done.[/quote]
Yeah it can be done but you’re still getting 2 m/s^2 of noise. Don’t fly that APM. It is broken. Attached is a graph comparing to the Y accel.
If you want to try to set it manually, your INS_ACCOFFS_Z needs to be set to 40580 and your INS_ACCSCAL_Z needs to be set to 15.68. Keep in mind that the accel calibration routine will never work without code modification, and that the copter will probably refuse to arm with those settings. Who knows when or if it will fail completely and stop working or change again.
Contact help@3drobotics.com if your board is a genuine 3dr board and is under warranty.[/quote]
Hi jschall,
It’s possible to use a external accelerometer like this ebay.com/itm/MPU6050-6000-6- … 416d7d221a ?
If it is possible, how could attach it to the board and how I have to configure?
It’s possible to use a external accelerometer like this ebay.com/itm/MPU6050-6000-6- … 416d7d221a ?
If it is possible, how could attach it to the board and how I have to configure?
Your sincerely,
Miguel[/quote]
The response from the dev team is that it would not be easy to do that. For starters, MPU6050 is i2c only, you would need to attach it to SPI, so you need MPU6000. Otherwise, if you can get it on the SPI bus and run it at 3.3v and disconnect the old device by cutting traces, it might work. Probably easier to solder on a new MPU6k if you have the tools.
Some dev quotes:
[6:50:02 PM] Philip Rowse: must be MPU6000, not 6050
[6:50:33 PM] Randy Mackay: yes, my guess is that it’s possible if he also does the soldering on the bottom of the APM to make the spi pins 3.3V instead of 5V
[7:27:00 PM] Randy Mackay: Jonathan, the spi bus runs at 3.3V but there’s a level converter between the AVR and the SPI pins on the left side of the APM. So to disable that level converter the user needs to cut a trace and put a solder blob on a couple of pads on the back of the board. Instructions are on the optflow wiki page: copter.ardupilot.com/wiki/optica … _the_APM25
I have removed the APM from the copter and connected to Mission Planner. The values via MP in the status screen and from the parameters list are attached. Can the value be measured anywhere else.
Attached is an extract from the parameter list.[/quote]
Hi ozyfly,
Please contact help@3drobotics.com to get a replacement ArduPilot. That offset is well out of spec for the MPU6000.
If you’d like to get a better idea of what the accelerometer is doing, try setting the ArduPilot on its back (pins down) and checking the Z accelerometer (az) value. The difference should be 2000 (2g). The correct behavior is for the Z accelerometer to read -1000 when right side up and 1000 when upside down.
Thank you for your advice. I will get into contact with 3DR. I previously contacted an Australian distributor from whom I purchased the APM and have not received a reply. I will give 3DR a try direct.
Thank you for your advice. I will get into contact with 3DR. I previously contacted an Australian distributor from whom I purchased the APM and have not received a reply. I will give 3DR a try direct.
Cheers
D[/quote]
3DR will tell you to get in touch with the distributor. Who is the distributor?