In the last couple of months, I've built two UAV's based on the Bixler airframe and the ATMega 2560 autopilot with XBee 900 Mhz telemetry. UAV1 has been flown a number of times and seems to be performing reasonably well. I started testing UAV2 on Friday and am seeing some significant random deviations (+/- 30 meters) in altitude measurement while the UAV is sitting immobile on the ground. The VSI deviates as well (+/- 100m following the altitude deviation). For benchmarking purposes, I fired up UAV1 (which has the components placed a bit differently) and am seeing +/- 3m altitude and +/- 4m VSI with that configuration.
After studying the forums a bit, I see that having the XBee module close to the IMU shield can cause this type of problem, so I moved the XBee module as far as the cable would allow (5 cm or so) and, while this did improve the situation, I'm still seeing a +/- 10m jump in altitude with every reading.
The next thing I'm going to try is moving the antennae for the receiver. Anyone have any wisdom to offer on this problem? I've got the new 3DR telemetry system I could swap in, but I'd rather figure out what's going on with the current configuration.
Thanks,
- Mark
Replies
Just to close the loop, I definitively determined that the source of the altitude deviations was due to the XBee by removing it and hooking up to the autopilot via USB. The altitude readings were quite steady w/o the telemetry module.
Based on tips from other posts here, I made an 18 inch XBee cable and was able to find various places on the Bixler that I could place the XBee module with minimal interference with the altitude readings. Interestingly enough, the altitude interference seems to be affected by the XBee antennae orientation, the XBee antennae shape, and the routing of the cable from the autopilot to the XBee module. I was able to come up with a combination that works with my set up. (Module next to the engine mount, antenna pointed along the wing). So now I'm down to an altitude error of around +/- 1 meter.
It would be interesting to experiment with EMF shielding for the autopilot....