GPS over 75 Ohm Satellite Cable

I needed to get a GPS signal for a couple of upcoming projects that use it for time synchronisation. The problem? Where I intend to place this equipment happens to be a windowless cupboard in the middle of the flat, and the metal roof blocks out the GPS signal. 

However! The original builders ran satellite coax cable to each of the bedrooms and the living room and terminates in said cupboard, along with two feeds from the communal Satellite distribution, allowing you to "patch" services as you require. I'm curious how many residents have discovered this as its all hidden behind a blank face plate, but it has previously came in handy when I wanted to mirror the Satellite reciever to the bedroom and ended up with a rather hacky, HDMI over Ethernet over Coax solution with cheap boxes from China that effectively ran powerline over the Coax cable.

But GPS receivers are expecting 50 ohm coax, not the 75 ohm cable used for television distribution. Will it work?

This is how things looked shortly after moving in to the flat. There's a face plate in the lounge doing all the things with satellite, TV and radio sockets, and additional TV socket which would have been for looping the analog output found on old satellite boxes to the bedrooms, and a telephone extension socket. You can ignore the socket below, its another telephone socket. Because you can never have too many.

In the photo below you can see what it looks like inside. The metal box is a diplexer, splitting the signals into Satellite / TV / Radio outputs depending on their frequency.

And this is what it looked like inside the cupboard. Someone at some point split the loop from the lounge to both the bedrooms with an amplifier. I don't think this setup would have been great, with the amplifier amplifying an already strong signal again. There's no gain control on this amplifier.

So with a handful of adapters, to convert the SMA found on the GPS antenna to F-Type in the living room, then from F type to TNC in the cupboard, I got everything hooked up, feeding the GPS antenna over the satellite coax into the cupboard, to a GPS receiver I've played with previously. Did it work? No.

I started troubleshooting and noticed that there was no DC voltage at the antenna side which is used for powering the internal LNA. I haven't managed to figure out why yet. I suspect the diplexer might have something to do with it, but in theory satellite boxes also send DC over the coax to switch the polarisation.

Fortunately, there's one socket without any diplexer or filtering, the Television loop socket. Its a Belling Lee / UHF / PAL socket, which is even worse than F Type but I had an adapter. And voila, I started to see GPS satellites appearing in gpsmon and got a solid GPS lock.

Below is an image I generated using gpsprof over 12 hours, showing plenty of satellites in view and a decent signal level. The poor signal to the North West is because the antenna is in the shadow of my building, but there's not much I can do about that unfortunately.

It's worth noting that this setup does run the risk of damaging the amplifier in the antenna. A mis-matched impedance will make the amplifier work harder cause reflections - power coming back towards the amplifier. This can damage them, although I suspect they're fairly resilient. I guess I'll find out over the coming months.

So, its not perfect, but it can work in a pinch!