hi,
I am back on the paparazzi list, Xavier has
decided to left the team this week.
The bug still alive....
This morning I have updated my system to 3.92 (up
to date version), I have made some tests.
My IMU is an Aspirn 1.5 with larger PCB, the
exact size of the GPS, to be mounted directly on the LISA/L
without any adaptor.
I also have a LISA/M with the stamp size Aspirin
1.5. it have realy fly
To test my IMU, I have removed the stamp aspirin
from the LISA/M, and solder some wire and a molex 12 pin
connector.
First time I connect my IMU, it was wrong, it was
a mistake on filtering cap on the power lines.
Now if I connect my IMU to the Lisa/M, everything
is ok (with the standard i2c driver), I can go to fly, if
mecanical fix My imu one the quad...
With the new driver
(<configure
name="USE_NEW_I2C_DRIVER" value="1" />) the mag and the
gyro run, after about 8 secondes they are died. The ACC still
to run (SPI)
Test on the Lisa/L.
With the standard driver
with
my IMU plugged at power up, All
four leds still off... nothing run !
power-up
without IMU, the first led flash at 1 sec (OK!) the 3 and 4
leds flash few sec, and still on.
I plug (hot plug!!!!) my IMU. The second led flash...
and still on.
On GCS the PFD is ok !
Evry thing feel good, the motors can be start with the
RC in ATT mode.
On message, with raw_sensor option the ACC is ok, the
Gyro is OK, the
MAG is died.
With the new driver
with my IMU plugged at power up The mag and the gyro
run, after few seconds they are died.
The ACC still to run (SPI)
GCS still
in « Kill » mode.
nothing better with "hot plug"
the tests have same result with « lisa/L my
aspirin1.5 » or Lisa/L stamp aspirin 1.5.
I have compile error with the option
« newdriver » :
arch/stm32/mcu_periph/i2c_arch.rewritten.c: In
function 'i2c_error':
arch/stm32/mcu_periph/i2c_arch.rewritten.c:667:11:
warning: variable 'err_nr' set but not used
[-Wunused-but-set-variable]
I will try to became friend with "git", and try
move in th past, just before the change master/3.9
I still on line and I am open for all new test,
or questionsss….
Regards,
Eric
Le 02/07/2012 17:34, Felix Ruess a
écrit :
Hi Xavier,
you don't have to use github, but could you please make
some patches and post them here? That would make things a
_lot_ easier...
Cheers, Felix
On Fri, Jun 29, 2012 at 9:34 AM,
Xavier RIVIERE
<address@hidden>
wrote:
Hi,
I already post a lot about my I2C problem with Lisa/L +
aspirin v1.5 + v3.9 configuration. After checking my
electronics, I check the software. And finaly, it's
working !!
Well... I change so much things that I'm not sure what is
the main part... As I'm not in the dev team, and I don't
want to open my own branch, I'm posting here my change.
----------------------
First, in my conclusion, the problem don't come from the
I2C driver... I made little change here. I just put this
line
while ((regs->CR1 & 0x200) == 0x200);
before each I2C_GenerateSTART and I2C_GenerateSTOP, to
avoid double send, as describe in STM32 datasheet
Add this line if need in the function :
I2C_TypeDef *regs = (I2C_TypeDef *) periph->reg_addr;
----------------------
So, for me, the problem come from the I2C driver call...
that's really a mess. There is T.H.R.E.E. calling function
:
send_i2c_msg_with_retry
i2cSubmit
I2CTranmit
I put i2cSubmit everywhere -____-
----------------------
I change the function imu_periodic, in imu_aspirin.c this
way :
http://pastebin.com/aWNAXQf1
I change the function imu_aspirin_event this way :
http://pastebin.com/vvXkNnkD
I change the interrupt function in
stm32/imu_aspirin_arch.c :
http://pastebin.com/Ar0gD9fr
I activate the interrupt for the mag.
How it works ??
- Captor send interrupt
- Interrupt handler happen, and send i2c reading
- imu_aspirin_event catch the reading end, and send
according handler
- imu_aspirin_periodic works as a watchdog. Mag timeout
must be around 30+. Accel and Gyro timeout must be around
3+
------------------------
Hope this will help. If someone can test it before
validate, it would be great ;)
Xavier
_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
--
************************************
Lez Concept EURL
Eric BAHIER
15 chemin neuf
34 980 Montferrier sur LEZ
France
Tel : 33 4 67 59 92 53
Fax : 33 9 59 15 92 53
************************************
_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel