Re: [Paparazzi-devel] lisa/m with mkk controller = problem
From:
Prof. Dr.-Ing. Heinrich Warmers
Subject:
Re: [Paparazzi-devel] lisa/m with mkk controller = problem
Date:
Fri, 28 Dec 2012 18:12:41 +0100
User-agent:
Mozilla/5.0 (Windows; U; Windows NT 5.0; de-DE; rv:1.4) Gecko/20030619 Netscape/7.1 (ax)
Hi
are you sure that there is not hardware problem on the i2c bus?
In the diploma work from the HB-Autopilot i read:
"1.1.1Störungen
auf dem I²C Bus
Während der ersten Inbetriebnahme und Tests der
Motoransteuerung viel ein stottern der Motoren auf. Während der Tests
der
Motorsteller war jedoch ein gleichmäßiger Lauf der Motoren festgestellt
worden.
Die Überprüfung des Signalverlaufs am I²C Bus
ergab, dass
der Signalpegel nicht immer 0V erreichte sondern nur etwa 1V. Der
Verdacht viel
hierbei auf die ungewöhnlich niedrigen Pullupwiderstände, welche im
Datenblatt
für den vorgeschlagen wurden. Es was durchaus möglich, dass die auf
den
Motorstellern verbauten ATMEGA8 Mikrocontroller nicht in der Lage waren
den
Signalpergel vollständig auf 0V zu ziehen.
Da bereits ein Prototyp des Pegelwandler mit
eigentlich
falsch bestückten PullUp Widerstanden einwandfrei funktionierte, wurde
beschlossen die funktionierenden Werte von 3,3k Ohm zu übernehmen.
Nach Austausch der Widerstände waren keine
Störungen mehr
messbar, die Funktion somit einwandfrei".
On the HB Autopilot 3.3k Ohm for the i2c pull up
where used for correct working with the mikrokopter controlles.
On the lisa 2,2k Ohm are used.
On the mikrokopter 1k an 100 Ohm series resistors
are used.
I had problems with the wires on the mikrokopter i2c bus.
Therefore is used 3 wires twisted, one additional for gnd.
The gnd on the controllers where sometimes more than 300mV away from
the gnd on the FC controller pcb.
Heinirch Warmers
Eduardo Lavratti schrieb:
Well,
with i2c new driver the mkk driver work much better but the problem
occur .
way the i2c1 clockspeed of newdriver is set to 40khz and i2c2 to 300khz
?
On Thu, Dec 27, 2012 at 12:19 PM, Eric
BAHIER <address@hidden> wrote:
Hi,
It is not a I2C bus speed problem.
The problem is probably, (it was for me) we try to send to many byte on
the I2Cbus, and the I2C bandwith is too small.
We must refresh the order to the MKK less time, you must reduce the
quantity of data below the bandwith.
it is pssible, but I don't remenber the correct parameters in the
airframe.xml file.... I will need it in few days...
Regards,
Eric
Le 27/12/2012 14:45, Eduardo Lavratti a écrit :
I think Its not buffer size ... its speed
problem.
Anyone know if ppz i2c bus run at 100 or 400khz ?
On Thu, Dec 27, 2012 at 6:48 AM, Eric
BAHIER <address@hidden> wrote:
Hi,
In the past with booz (and near futur), I fly an hexa.... I had the
same problem.
The problem was an over load of the I2C bus, we put slower the data
refresh for 6 mkk esc...
I dont rember where we make it slower....
Have nice day,
Eric
Le 27/12/2012 01:54, Eduardo Lavratti a écrit :
Hello,
anyone using 6 ou 8 MK esc with lisa/m ?
Today i put my lisa/m board on my BiG hexa.
This hexa fly very well with MK boards.
i have an mini octo that fly very well with mk boards too.
with lisa_m it fly some time and the lisa HANG resulting in crash.
in second test crashed again.
i make some tests in the bench.
i tested with big hexa setup and mini octo, all with i2c escs.
the result:
with 4 esc the lisa never hang.
With 6 esc lisa hang after some time
with 8 esc lisa hang much faster than with 6.
i think it is a i2c buffer problem again.
I remember i configured the buffer last year an tell felix about this
problem.