Repeated mid-air APM reboot - only happens at one flying site - possible weather radar interference

I have three flying sites that I use depending on weather and availability. My primary flying site is located only 800 meters from a weather radar and I have been having repeated in-air reboots when flying from that flying site only. Fortunately, being new to the ardupilot system, I put it in a slow flying high wing airframe and have enough years of RC experience to catch it and go manual each time.

Given that so far I have only experienced the reboots at the one location and the weather radar is the only powerful emitter I am aware of operating in the area, the weather radar is now the leading suspect. I was wondering if any other users have had experience making their hardware a bit more resistant to EMI (Electromagnetic Interference). As the hardware comes out of the box it contains little to no protection.

Knowing what mechanisms could cause the processor to re-boot would help me implement a more targeted solution than adding chokes to leads and wrapping leads and the processor in alumnimum foil.

Views: 371

Reply to This

Replies to This Discussion

More information to support the weather radar theory. The reboots were mostly 5min or 10min apart. The Met bureau states that the unit scans every 5min, and  operating on S-band (10cm wavelength) and has a 1deg beamwidth. The power output will be over50kW for that type of radar.

I'll have a go at adding chokes and foil. but suspect I may have to avoid the area altogether.

Have you analyzed the logs for the flights in question?

If it is truly rebooting, as in a power-on reset, I would think a new log would be created while in flight. Your observations do lend credibility to the site as the problem, it only needs to be determined which part of your electronics is suffering from the interference.

That kind of energy could be simply overloading the rejection capability of the receiver and not the APM. It could also be soaked up by the servo leads and causing spurious/erratic function.

More data will be needed to be sure.

Because S-Band operates in the range of many 2.4 GHz Tx/Rx systems, your site is definitely suspect - if your RC gear is 2.4 GHz. From the Wikipedia article:

The S band is part of the microwave band of the electromagnetic spectrum. It is defined by an IEEE standard for radio waves with frequencies that range from 2 to 4 GHz, crossing the conventional boundary between UHF and SHF at 3.0 GHz

 

-=Doug

New log on the AC - yes. On the GCS logs the millis since reboot counter goes back to zero, and the AP goes into initialisation so I'm certain its a reset. I've added chokes to both power leads and  twisted and shielded the GPS and radio modem leads so we'll see how it goes next weekend. The servo leads from my Spektrum RX will be next once I have enough chokes to go round.

Are you running APM off one supply, or are you powering each rail separate? It would be good to try powering each rail separately at that site to see if it happens, this is assuming you are powering off one source now !

I'm using a 3DR power module for the APM and the ESC BEC for the servos with the jumper removed. Assuming the radar is putting out 50 kW and has an antenna gain of 30 dBi (typical for a parabolic dish of ~1.5 m at S-band frequencies), the electric field strength at 800 metres is in the order of 50 V/m. This means several volts can be generated on even short leads.

You probably will have to shield any long wire runs (more than a few inches) and use some .001 bypass capacitors on the DC lines.

Or, as you suggest, just don't fly in that area.

I am leaning toward Stephen's 2nd suggestion unless the suspect site is a premium place to be.

If the site is a must-do, you will have to start the process of hardening your aircraft to EMI. It will be a time consuming process but not necessarily an expensive process. As Stephen suggested, sheilding the long wires is a start. How to do this?

Try self-adhesive copper foil tape. This would allow you to also solder capacitors to the shield for additional decoupling. 3M has a line products made just for this. Click HERE.It is easily cut with scissors and can be shaped to many surfaces.

The goal is to attentuate the unwanted energy and also maintain your control of the aircraft through the original RC system... which you still have not mentioned what gear/frequency you are using.

-=Doug

I'm using a Spektrum AR7010 receiver. I've had no problems with genuine Spektrum gear at this site, but some of the cheaper orange receivers will come back with the led flashing.

Have you checked the 'vcc' value in the logs?

VCC completely normal up to the point where logging stops.

Hi Paul, do you have a confirmation about weather radar afecting APM2.5? This really interest me.

I did a fly last week with APM2.5 powered by 3DR module, and I did have the same in-air reboot as you, see my link below. At that time I did not make any connection to weather radar site, since I did fly before on this site, but with APM2.0 and all was OK. Maybe radar was powered off...

Do you have some tlog or log from that time where did you have the in-air reboor because of weather radar? You said in another post you did see a counter reseting in log which give you the confidence it was an in-air reboot. Which param is this?

I want to compare your logs with mine.

Disarm at 80 m, close to a weather radar

Look at the time from last reboot im ms. There are a number of other signs of a reboot (battery status goes back to 100%, airspeed autozeros, etc.

 

This has been confirmed with another club member with an APM 2.0 having the same problem.

 

The resets occurred when the weather radar did its scan which corresponded to when the rain map was updated on the web-site.

 

Lining the equipment bay with foil and adding chokes on long leads has cured the problem,

RSS

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Drone Delivery Challenge, is here

© 2014   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service