Bad GPS accuracy when charging action camera

I’ve been doing some cycling Mapillary sequences the past couple of days, and I’ve been using my new Xiaomi Yi, which I have not used on a bike much yet. I’ve been carrying a spare power bank and cable so that when the camera battery gets low I can charge on the go.

However, I’ve been noticing that when I have the camera plugged in, my GPS accuracy drops precipitously. It’s most obvious in areas which are already difficult for a GPS, like under trees or going around fast corners (takes a second for the route to ‘catch up’, so to speak).

So a couple questions: is this confirmation bias? Am I seeing a drop in accuracy because I think it should be there? Or has anyone else had this same problem? To clarify, my GPS sits right above the camera on a K-Edge mount. When I remove the GPS and hold it away from the camera, the route jumps back into position.

Secondly, does anyone have ideas of how to mitigate this? I was thinking a Faraday style barrier between the two devices might help… I cannot reasonably move the GPS, and I’m not sure where I could move the camera.

Unrelated, but I also noticed the effect that high-power overhead cables can have on the GPS.

EDIT: I did some quick and dirty testing. Signal strength drops when the camera is on, let alone charging or taking pictures, although those seem to be a little worse. Shielding with many layers of aluminum foil between the GPS and camera seemed to help a little bit, but I’m not sure if it’s shielding the GPS to some extent as well, which defeats the purpose.

I have never had GPS and camera close together, så I don’t know if it can cause problems, but I have noticed several times that GPS doesnät work near electrified railroads or high voltage power lines. I have even had problems in certain places there I suspect underground power lines. So elektormagnetism can have an effect on the GPS quality.
So I think it is plausable that a small transformator in a camera can generate enought elektromagnetism to disturb a GPS at close range.

Better to keep the GPS somewhere else.

Reviewing past sequences, I think that yesterday may have just been a culmination of issues. It was cloudy, there were trees, the camera was charging. Together, accuracy plummeted in some places.

What’s really weird and I still can’t figure out is that at some random point, my GPS seems to have lost 17 seconds (or something). So when I correlated the photos to the GPX with a 3-second offset, the first third was right, but the back 2/3rds was consistently off. What I don’t understand is that I took some reference photos with the camera of the GPS time during the back 2/3rds that showed a 3 second difference.

I tested many times driving along a tramway and even along a sparkling tram, this has NO impact on the GPS signal.

AC or DC, do you know the voltage?

It was in Brussels…

600 V DC (or maybe 750 V) if you compare that with 15 kV AC in Sweden it is easy to understand, that you have no problems in Brussels and I have problems in Sweden

But then again, I have a good reception when I hang my GPS receiver outside the windows of a train.
But on the other hand, this is not under the pulling carriage.

Trains in Belgium run on 3kV DC not on 15 or 25kV AC. I think that is a big difference. But you could try it in the high speed train, that should be 25 kV AC. But there are som ways even for AC systems to minimize the elektro magnetic radiation. So I think it is reasonable to blame bad connectionion to High power gridlines, but that doesn’t mean that all gridlines cause bad connection

I did find an academic article investigating whether power lines affect GPS. It appears it does, if the GPS is moving. Link here:

Played around with positioning of GPS and Camera, and putting the GPS on the stem of my handlebars, with the camera on the out-front mount, seems to clear up the issue. The stem, being metal, probably reduces interference. I verified this by moving the GPS to the mount while the camera was running, and signal strength instantly dipped.

Interesting! I was researching this recently because I had a similar issue using the SJ4000 action camera, and then I bumped into this video: