MPU6k failure related to main battery voltage going below 11V

[Randy Mackay]

 

     Another MPU6k failure has appeared on the AC3.3-beta testing thread from a user called Frederic Jean.  This user’s MPU6k’s Z-axis accelerometer suddenly develops a 3m/s bias when the main battery voltage drops to below 10.3V.  He can easily reproduce the issue and in the logs provided it always happens at the same main battery voltage.

                http://diydrones.com/forum/topics/copter-3-3-beta-testing?id=705844%3ATopic%3A1964465&page=174#comments

 

     The MPU6k failures are not new - we saw many due to a manufacturing issue of the Pixhawk boards between June 2014 ~ Feb 2015.  That manufacturing issue has been fixed and I expect many of the bad boards have been RMA’d but some are still out there.

 

     Anyway, what’s new with this report is the strong correlation with the main battery voltage.  Also when I checked another log from earlier this week from another user I see the same correlation with main battery voltage (although it’s at 11V).

 

-Randy

[Tom Pittenger]

Could there be an issue with the power module?

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Robert Lefebvre]

My off-the-cuff guess is that as the voltage gets low, the PM voltage outputs starts to get a bit rough, and may be causing problems with the MPU6k.  The question is if this is an issue for all, or only some PM's, or only some MPU6K's, or only a combination of certain units.

I don't think it's likely to be a problem directly related to battery voltage level, as I have a number of helicopters running the Pixhawk off a 2S battery supply (6-8.4V) without issue, but using a linear regulator.  And of course, the Pixhawk Vcc has no direct connection to the battery, save for through the voltage divider circuit.

[Philip Rowse]

can you please post this log?

Philip Rowse

Lead Systems Engineer 

3D Robotics Australia

+61-(0)4-2732-6366

website | facebook | instagram

[Jaime Machuca]

Logs  are on the DIY Thread:

https://www.dropbox.com/s/gr2jipj4u1jwnxx/282.BIN?dl=0

https://www.dropbox.com/s/r5948y8ovswripn/284.BIN?dl=0

Best regards,

Jaime Machuca Mercado

CTO | Droidika | www.Droidika.com

Mail. Ja...@Droidika.com

Tel. +52 (33) 3777 5327

Cel. +52 1 (33) 3945 3350

Este correo electrónico y cualquier archivo transmitido en él, son confidenciales y para uso exclusivo de los individuos y entidades a quienes está dirigido. Si usted no es el destinatario previsto o la persona encargada de recibirlo, y tiene por error este mensaje, queda prohibido y sin validez el uso, difusión, re-envío, reimpresión o copia. Toda oferta y/o aceptación de propuestas comerciales, celebración de contratos u otros actos tendientes a la adquisición de bienes o servicios, así como el establecimiento de cualquier clase de obligación legal para Droidika S.A. de C.V., deberá confirmarse por escrito firmado autógrafamente por funcionario competente, excepto que se cuente con un contrato vigente que autorice el uso de este medio para tales fines. Si usted recibió este correo por equivocación, favor de notificar inmediatamente por este medio a su remitente, y después borrarlo de su correo

[Philip Rowse]

something else is going on here.... 
It starts as the failsafe engages... but what is the pulsing?

Philip Rowse

Lead Systems Engineer 

3D Robotics Australia

+61-(0)4-2732-6366

website | facebook | instagram

[Thorsten]

This somehow reminds me of my buzzer induced magnetic field and vibration problem: https://groups.google.com/forum/#!topic/drones-discuss/7MIpOZJY7SQ

[Robert Lefebvre]

Nice Thorsten!  That's a definite possibility.  It would explain the periodicity of the IMU error.

On 8 July 2015 at 17:53, Thorsten <thorsten...@soilution.de> wrote:

This somehow reminds me of my buzzer induced magnetic field and vibration problem: https://groups.google.com/forum/#!topic/drones-discuss/7MIpOZJY7SQ

--

[Philip Rowse]

we need to replicate this....

nice theory....

Philip Rowse

Lead Systems Engineer 

3D Robotics Australia

+61-(0)4-2732-6366

website | facebook | instagram

[Jason Short]

if the failsafe is triggered what is the frequency of the buzzer beeps? Does the buzzer draw a lot of current?

[Randy Mackay]

 

     Yes, it must be this!  Nicely done Thorsten.

 

     It can’t possibly be a coincidence.  That pattern on the z-axis accel looks just like the buzzer sounds.  The buzzer is placed directly above the Pixhawk in an IRIS+ (I just checked mine) and there’s not much clearance so a little mass of wires or something moved about could cause them to touch.

 

     I’ve asked the user, Frederic, to disconnect the buzzer and try again.

 

-Randy

[Thorsten]

This is what happened in my case:

(The pink line is MagZ)

Vibration damping is pretty good on this system. But the buzzer was mounted on the same vibration damped platform as the autopilot. So any vibration induced by the buzzer were directly transferred to the IMUs. But I suspect the offsets between the IMUs are coming from magnetic interference. Although I am not 100% sure. It might also be the specific frequency of the buzzer sound.

However, the effect the buzzer has when mounted closely to the autopilot can be different depending on its position relative to the sensors.  Depending on the location of the buzzer I saw jumps and drops in altitude of different strength when changing flight modes. 

It may also effect the baro due to sonic effects.

[Daniel Frenzel]

Then the accel works fine. The Buzzer sounds would create a high frequency vibration, which is just detected. 

[Thorsten]

Yes, the accels work fine and AC3.3 as well. It is hardware setup problem which is easy to solve. Anyway, I hope not too many systems are affected since it could easily lead to a crash.

[Jonathan Challinger]

My bad, I picked the sound :)

On Jul 10, 2015 2:37 AM, "Thorsten" <thorsten...@soilution.de> wrote:

Yes, the accels work fine and AC3.3 as well. It is hardware setup problem which is easy to solve. Anyway, I hope not too many systems are affected since it could easily lead to a crash.

[Philip Rowse]

Lol! 

[Thorsten]

My bad, I picked the sound :)

I know! :-)

Good news: Federic Jean made some tests and changed the location of the buzzer. All is fine now. 

Question: vibration issue or magnetic interference?

[Randy Mackay]

 

     This is definitely vibration.  One last concern on this is whether this new tone will affect all IRISs or if this was a special case.  I can do some tests of invoking the buzzer on my IRIS+ and/or others can see if their IRISs misbehave when the battery alarm goes off.

 

-Randy

 

From: drones-...@googlegroups.com [mailto:drones-...@googlegroups.com] On Behalf Of Thorsten

Sent: 13-Jul-15 3:34 PM
To: drones-...@googlegroups.com

--

[Philip Rowse]

and if there are tones that can affect solo.... the buzzer is right next to the pixhawk 2

Is it possible to have a tone that runs through the full frequency range while on high rate logging?

Gotta love unintended consequences...

Phil

[Thorsten]

One thing I am wondering about and what makes me think it has also to do with some magnetic effect is that if I move the buzzer on my vibration damped platform the behavior changed from falling to climbing. But the intensity of the vibration should have been the same because all is hard mounted on that platform. However, the magnetic field has moved. I am a little scared to make some systematic field tests. But maybe if I find some time...

Also, why do the accelerometers diverge? Is this due to the specific frequencies?