I also have a similar problem. Whenever the crazyflie was working, it was working fine with MPU9250, AK8963 and LPS25H all OK. But every now and then the red LED would start flashing five times and then it would take a lot of effort to get it out of that state again. Switching off and on wouldn't help. First I thought it might be a sensor calibration issue. So I tried the shaking as recommended in another thread and after some time it would work again. However, there were also the "I2CDEV: Error callback nr: 1024" error on the console with fail for MPU9250, AK8963 and LPS25H. In particular, it happened every time I used the Turnigy Nano-Tech 300 mAh instead of the stock battery. Now I assume this is because the Turnigy is narrower than the stock battery and therefore doesn't put pressure on the sensor like the stock battery does.
After a while I found this thread, perfectly matching the behavior I have seen. So I tried to reflow the pressure sensor (U9 on the board) with a hot air rework station, but I only managed to make the behavior more stable: instead of occasionally not working it now never works.
With the latest cf2-2016.02.bin firmware release, when I connect the Crazyradio PA to the Crazyflie right after I switch it on, I now always get:
Code: Select all
I2CDEV: I2C timeout callback dev: 2
IMU: MPU9250 I2C connection [FAIL].
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
I2CDEV: I2C timeout callback dev: 2
IMU:
I2C connection [FAIL].
I2CDEV: I2C timeout callback dev: 2
IMU: LPS25H I2C connection [FAIL].
DECK_DRIVERS: Found 6 drivers
DECK_INFO: Found 0 deck memory.
DECK_CORE: 0 deck enumerated
EEPROM: I2C connection [OK].
I also flashed the manufacturing test firmware (cf2_run_imu_manufacturing_test.bin). With that, I get:
Code: Select all
SYS: I am 0x3132383231334715360022 and I have 1024KB of flash!
IMU: MPU9250 I2C connection [FAIL].
IMU: AK8963 I2C connection [FAIL].
IMU: LPS25H I2C connection [FAIL].
EXPBRD: Found 0 memories.
30 78 42 43 0 50 0 0 0 0 0 0 0 0 0 7D
Status 0
EEPROM: I2C connection [OK].
Actually, I do not need altitude hold. So it would be perfectly fine to just disable the pressure sensor functionality.
I recompiled the crazyflie firmware with "#define IMU_ENABLE_PRESSURE_LPS25H" commented out in hal/src/imu_cf2.c as described in
viewtopic.php?f=5&t=1125&p=6141&hilit=E ... back#p6215
Unfortunately, that also doesn't work. With this self-compiled firmware I get:
Code: Select all
h!
IMU: MPU9250 I2C connection [FAIL].
IMU: AK8963 I2C connection [FAIL].
EXPBRD: Found 0 memories.
EEPROM: I2C connection [OK].
OW: Found 0.
30 78 42 43 1 50 0 0 0 0 0 0 0 0 0 E7
Status 0
30 78 42 43 1 50 0 0 0 0 0 0 0 0 0 E7
Status 0
E7 E7 E7 E7 1
Status 0
Does this have to do with the AK8963 also failing? AK8963 also failed before my reflow attempt whenever MPU9250 failed. But whenever MPU9250 was OK, AK8963 and LPS25H were also OK. Do I need to make additional changes in the firmware source or do I have to get the reflow working?