Nuno

Update!
With only the bad string connected, I took out one of the panels and I think I got lucky right on the first one! I tried to connect the string with 2 panels to the inverter and it began charging! So i guess I found out the culprit.

Is it worth to open the panel's "box" and look into some possible issues I can fix? Or is it good to replace? (its only 2 years old )

Ok guys, mystery solved! Kind of ....
I was experimenting connecting just one battery at a time to check if maybe one of them was fault. Then I noticed something odd. Sometimes when I turned on the battery switch it went to 7V, other times it stayed off. I then tried the same thing with the 3 of them together and got the same result... no consistent behavior.
Then it ocurred to me that maybe the fact that I'm switching them on one by one is somehow confusing the inverter<->battery, and by safety it shuts down.
So I unplugged the inverter cables, turn them all on (getting 50.4V), and then plugged in the cables (getting the expected spark in the last connected terminal). The charge held up at 50.4. So I switched on the panels and finally the inverter came alive
My bad for not having a proper DC breaker, this whole thing might have been avoided!
Phew! thanks everyone
 

I can put some load, but I wanted to get the stats of how much power can be captured regardless of the load. How do all the solar monitors work? I had the idea  that the stats you usually see online are the max power output of the panels, regardless of consumption. It never occurred to me that might not be possible if power is not being used (facepalm)

Yes, I know It was just a quick way to represent it in the drawing
For what I was told, this kind of (cheap) inverters generate AC on the live and the neutral sides, half in each, and thats why I got 110v live-ground and 110v neutral-ground...