At the moment I do not have Booz IMU + Booz connected. The behavior was in the messages the z axis was not reporting. Just 0s. All the other axis were reporting values. When I used an older paparazzi3 source it worked. All 3 axis. Nothing in the hardware was changed. I even used the same airframe XML file. Only difference was the source.
Since my hardware is not connected I am unable to verify myself. Too many other things I am working on ATM.
Regards,
David
On Jun 17, 2012, at 12:53 PM, Christoph Niemann wrote:
Hi Gautier and David again,
I tried it also using the last SVN checked into savannah (r6430). The MS2100 didn't work, but the I2C1 define didn't prevent the AP from starting. I also tried to use the driver for the v1.2 but it didn't build since the compiler somehow didn't find the defines for the signs of the gyros and accels. I didn't investigate this further, since that was a really old version.
it seems to be an I2C issue. I tried it again with the configfile below and it doesn't initialize. Its enough to just define USE_I2C1 and not load any modules (I tried also to load I2C modules as well, but as expected it didn't change anything). Perhaps you could try this file tomorrow.
Maybe I have a booz/imu_v1.2 at work, so you will have to wait
tomorrow before I can test it.
Can you try to load one of the i2c sensor, no matter which,
configured on i2c1. It can help to tell if it is a i2c or peripheral
issue.
Gautier
Le 17/06/2012 11:14, Christoph Niemann a écrit :
Hi David,
sounds interesting. The current behavior, trying to use the MS2100
is, the IMU initializes (the ADC gives values and the LEDs glow
continuously) but the MAG_RAW message only gives (0,0,0). Do you
think this could be a new bug? Which SVN revision do you use? (http://svn.savannah.nongnu.org/viewvc/paparazzi-software/trunk/?root=paparazzi&pathrev=6430)
? Which IMU version do you have? Could you supply the config-file,
so I could do some HW-testing with the original mag?
Could someone who has a Booz-Mainboard try, what happens if the
define USE_I2C1 is set?
Christoph, Have you tried
with older paparazzi3 source? I found if I used older SVN
paparazzi3 the single axis mag does work. With the newer Git
code it does not. AJ was going to create an issue in Git. I
have confirmed it many times but not until now did you also
see the same issue with v1.1 Booz IMU.
-David Conger
On Jun 16, 2012, at 9:44 AM, Christoph Niemann
wrote:
I forgot: It starts with
v1.1 and v1.0 since there is no USE_I2C1 in their
makefiles.
Yep, it also crashes
without anything plugged. As I wrote, normally
its a v1.1 IMU with the MS2100 (potentially
broken, the SPI works, since the MAX1168 does
- I disconnected the DRDY from the MS2100)
placed. Currently I build it using the driver
for the v1.1. If I load the HMC-Module for
I2C0 it doesn't crash. If I don't load the
module and do the define for I2C1 it crashes.
It doesn't matter if the HMC is plugged or
not.
I've got a LA but unfortunaly currently I'm
not at home. Do you have a suggestion what to
look for?
Does it also crash on the booz board
without the HMC plugged ? Does it
start with imu v1.0 or others
without the magnetometer code ? By
any chance, do you have a logical
analyzer ?
Gautier
Le 16/06/2012 16:43, Christoph
Niemann a écrit :
Hi
Gautier,
thank you very much, but if I
use v1.2 (which already has
the I2C1 define) or v1.1 (in
this case by loading the hmc
module and changing the I2C to
I2C1 (because it is connected
to this port, because I want
to use I2C0 for the MKK ESCs
and the hmc is 3.3V I2C)) the
Booz-Main board doesn't flash
the LEDs anymore, when
aligning the AHRS and there is
no telemetry. I tried to
reproduce it using a TWOG and
the fixiedwing fw and there is
still telemetry. I have
absolutely no clue, what could
be the problem, so if somebody
has an idea, I would be really
happy! The fact the HMC isn't
needing the DRDY-line is
really good :-)
The HMC58XX driver is
polling the sensor, so
their is no need of the
DRDY pin. Also the flag
USE_I2C1 is already
define by the subsystem,
so you should not be
necessary to define it
again.
Gautier
Le 16/06/2012 00:13,
Christoph Niemann a
écrit :
Hi
Felix,
thank you very
much. As usual
your response was
very helpful.I was
just using a
really old
version, since
then it didn't
build. Now I'm
using the current
v3.9 and it
builds. But there
are still some
problems: If I
define USE_I2C1
the LPC won't
start at all, I
wasn't able to
reproduce this
with a TWOG. Might
it be a
HW-Problem? I
think I should use
i2c1 since it
should be 3.3V
I2C.
Second: I'm using
this (http://www.sparkfun.com/products/9371)
Breakout Board
from Sparkfun. It
doesn't have the
DRDY-Pin. I was
able to solder a
wire to the pad
and could solder
it accordingly,
but: Will it work
if I define
HMC5843_NO_IRQ
without DRDY for
the AHRS?
Last but not least
I saw the
"original b2_v1.2
has a 1.8V Vreg.
Is it really
needed?
If there is
somebody, who has
answers for this
questions help
would be highly
appreciated.
the driver for
the booz imu
v1.2 on
LPC21xx uses
the arch
independent
driver
hcm58xx, so
you don't need
an arch
file...
See
conf/firmwares/subsystems/shared/imu_b2_v1.2.makefile
Cheers, Felix
On Thu,
Jun 14, 2012
at 9:53 AM,
Christoph
Niemann <address@hidden>
wrote:
Hi
everyone,
currently I'm
trying to get
some old booz
Hardware to
work.
Everything is
fine and
works, but the
mag doesn't.
It's a Booz
PCB set v1.1 .
So the mag is
currently a
MS2100. I can
see the LPC
sends
somesthing
like 0x4B to
the mag, but
it doesn't
pull DRDY. So
I guess the
MS2100 is
broken. I have
a HMC5843
here. So what
I want to try
to do is to
just change
the type of
the IMU from
v1.1 (MS2100)
to v1.2
(HMC5843). But
the arch file
for the HMC on
the LPC seems
to miss. Am I
making a
misstake or do
I have to
implement it?
Or has
somebody
already
implemented
it? Every help
would be
appreciated.