Suggested fixes to EM interference

[Jose_Alberto]

Hello,

I am installing an Emlid Reach in a UAV to work with a Reach RS on the ground.

All works very well, except that when the drone is switched on, the satellite reception drops by 10-15% due to noise and I lose fix.

The tallysman antenna is placed on top of the drone and has a 10cm aluminum ground plate. I thought this would be enough to eliminate all EM interference since EM waves don’t penetrate aluminum.

But still, even powering the reach separately from a USB power source, when the UAV electronics are turned on the reception drops and the solution is lost.

How would you go about solving the problem? Should I try to isolate the reach itself? The cable?

The biggest interference seemed to be coming from the on-board linux computer, or from its power supply

I found this @RickJ looking through the forums, but my issue doesn’t seem to travel through the cables

I would try usual EMI cures. Good low resistance grounds. High quality USB3 hub. Added shielding on USB cables &hub. Adding low pass filtering to USB data/power leads,Etc., etc.

Potential multiple causes. Lots of trial and error to solve.
(Isn’t the easiest thing in the world to do…)

[bide]

Where is the Reach module compared to your other electronics? Maybe the noise injection point is not the antenna or the ground plane, but between the antenna connector and the GNSS chip on the module.

Temporarily place a piece of tin foil between reach and your other wires/circuits and see if the sat reception jumps way up.

[Jose_Alberto]

Here is a picture of the setup. I will try covering different parts with tin foil to learn more about the problem, but I will need a more elegant solution since this is for a customer.

[Simon_Allen]

wrap your reach in tinfoil. I did on my Phantom 4 and it improved things massively. For me it was the P4 wifi that was killing me.

[Jose_Alberto]

Thank you for your suggestions. I did try adding tin foil to the reach but unfortunately it didn’t help much.

What ended up solving the issue was as simple as separating the antenna cable from the inference source (The black box left of the pixhawk below the reach antenna)