Re: [Paparazzi-devel] Lisa/L Octocopter with MMK Power Board Reseting Intermittently

[Piotr Esden-Tempski]

Hi,

One Idea about Lisa/L, you could try to make the i2c pullups stronger. (lower 
resistance) That may help with the MK controllers. The best thing to 
investigate this problem may be to use an oscilloscope and check the quality of 
the i2c signal coming out of lis/l vs. lisa/m.

Cheers Esden

--
Transition Robotics Inc.
http://transition-robotics.com
the makers of the Quadshot
Co-Founder, Embedded systems Engineer
Mobile: +1 831 440 7454
Skype: esdentem

On Jan 27, 2012, at 5:32 PM, Felix Ruess wrote:

> Hi,
> 
> better late than never...
> This could possibly be an i2c issue, maybe something else (as well?)
> ... or did you find out more already?
> 
> Regarding issue 2): is the whole lisa resetting, or just the motor
> controllers? If the whole lisa is resetting it might be a power issue.
> 
> Issue 3) could very well be an i2c issue, I have experienced similar
> things where it got stuck in an i2c interrupt routine with the current
> i2c driver on the stm32.
> 
> See https://github.com/paparazzi/paparazzi/issues/71 for the current
> status on the new i2c driver..
> 
> Cheers, Felix
> 
> On Wed, Jan 4, 2012 at 1:37 AM, Simon Wilks <address@hidden> wrote:
>> Good new/bad news.
>> 
>> First the bad:
>> After looking through the i2c_arch.c code and trying out a lower clock rate
>> on the I2C2 bus (Aspirin) it seemed to run longer until an issue popped up
>> but I would need to test more before I could be sure of this. I started
>> adding some code to indicate i2c errors using the LEDs but haven't finished
>> that yet. I still have the feeling that this is somehow related to the
>> Aspirin as unplugging it stops the failure occurring.
>> 
>> And now for the good:
>> I now have solved the problem a different way...kinda. Today the Lisa/M I
>> ordered arrived (I didn't get it because of this) and I have set it up and
>> connected it and it is running smoothly since about 15 minutes now without
>> the slightest issue! So this is clearly something Lisa/L related. The only
>> difference at the moment is that I don't have the GPS and XBee connected but
>> I tried the Lisa/L out without these too with no change.
>> 
>> So why is this? I would like to get the Lisa/L going but will settle for the
>> Lisa/M.
>> 
>> Thanks
>> 
>> Simon
>> 
>> 
>> On Mon, Jan 2, 2012 at 8:37 PM, Simon Wilks <address@hidden> wrote:
>>> 
>>> When the motors are idling (they are drawing around 1A) if something goes
>>> wrong everything resets. If leave them running with more throttle (drawing
>>> 2A) the motors will hiccup but everything keeps working! Still not good, but
>>> maybe a hint. Still digging...
>>> 
>>> 
>>> On Mon, Jan 2, 2012 at 4:42 PM, Simon Wilks <address@hidden> wrote:
>>>> 
>>>> 
>>>> On Jan 2, 2012 4:13 PM, "Eduardo lavratti" <address@hidden> wrote:
>>>>> 
>>>>> Simon, i am not sure but i think the ESC use I2C1_bus and
>>>>> sensors(aspirin / baro) use I2C2_bus.
>>>> Correct.
>>>> 
>>>>> Soo, the i2c queue for motor not use the same queue as sensors.
>>>>> I not have any schematics here to confirm this question.
>>>> It's just odd that when it's connected things screw up so fast. It could
>>>> of course be something else though. The power distribution board was
>>>> preassembled but it still could be something there.
>>>>> 
>>>>> I remember that my lisa hang too on the first tests.
>>>>> The problem occour because the 5v regulator work very HOT with 12v.
>>>>> When i put a 7v regulator on V_in the problem are solved.
>>>> 
>>>> OK. Worth a try. I am testing with a lab supply and I can try a lower
>>>> voltage... or with a BEC.
>>>> 
>>>>> 
>>>>> I not know anything about LISA/L board. This is my next aquisition if
>>>>> lisa/m past on my tests .
>>>>> 
>>>>> regards.
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> Date: Mon, 2 Jan 2012 14:58:41 +0100
>>>>> From: address@hidden
>>>>> To: address@hidden
>>>>> Subject: [Paparazzi-devel] Lisa/L Octocopter with MMK Power Board
>>>>> Reseting Intermittently
>>>>> 
>>>>> 
>>>>> What I have noticed so far:
>>>>> 
>>>>> 1) Momentary Failure on All Motors
>>>>> There will be a clicking sound, coming from no one particular motor and
>>>>> the LED strips flash off and on very quickly and all continues normally. 
>>>>> You
>>>>> can see this happening in this video.
>>>>> 
>>>>> 2) Reset
>>>>> All idling motors stop as what appears to be a reset is issued on all
>>>>> controllers (Lisa/L LED activity changes as well). This usually happens
>>>>> within about 30-60 secs of arming the motors, but not always.
>>>>> 
>>>>> 3) Lisa/L Freezes
>>>>> If I leave the motors running idle eventually all motors shutdown and
>>>>> the Lisa/L freezes, ie: LED 2 and 3 remain on and there is no more 
>>>>> activity
>>>>> in the GCS/Messages. Only a power cycle will fix this.
>>>>> 
>>>>> What is interesting so far is that only 1) and 2) above happen when the
>>>>> Aspirin IMU (v1.5) is connected. If I disconnect the IMU then I have only
>>>>> noticed 3) happen once so far after about one hour testing. I tried a 
>>>>> second
>>>>> Aspirin and nothing changed. To make sure this was not because the motors
>>>>> were all running at a constant speed seeing the IMU was not picking up 
>>>>> small
>>>>> movements I moved the roll/pitch controls without an issue.
>>>>> 
>>>>> Is this an I2C issue? It seems like it could be the Lisa/L freaking
>>>>> out. Perhaps too much is happening on the bus with 8 motors plus the IMU?
>>>>> 
>>>>> I currently have I2C_TRANSACTION_QUEUE_LEN = 16.
>>>>> 
>>>>> I will do more testing later in the day.
>>>>> 
>>>>> Simon
>>>>> 
>>>>> _______________________________________________ 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
>>>>> 
>>> 
>>> 
>> 
>> 
>> _______________________________________________
>> 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

signature.asc
Description: Message signed with OpenPGP using GPGMail