For test purposes, don't worry about the grounding issues.
Running a single piece of coax (preferably new, but known good for sure) is an excellent first step in trouble shooting. You eliminate the question of whether an accessory such as a splitter, connector, amplifier, etc. could effect the signal from the antenna. Reliable reception after all, starts at (and often fails at) the antenna and it's mounting (location & aim). If you can't get reliable reception with a direct connection to a single TV, adding accessories will not improve the results, if anything you may only add more problems.
Use the shortest cable that will allow you to safely and practically make the connection from the antenna to the 'test' TV. This reduces the signal loss between the antenna and TV, further reducing the question of 'do I have a good signal from the antenna or not?'
Hopefully, your test will provide a solid, reliable signal on all channels of interest. If not, transition to adjusting the aim of the antenna. If needed, don't hesitate to check a full 360° of aim points in 5° steps. You're looking for reliable reception more than 'maxing-out' a signal meter reading. Take notes, with reference to compass heading or landmarks so you can return the antenna to the best aim point once your aim survey is done.
If you can't find a single aim point that provides reliable reception, you'll need to decide if the antenna can be located in a more favorable location, one that's not obstructed by buildings or trees. Higher is nearly always a better location. It's possible to clamp the antenna to a broom handle, then test alternative locations and aim points.
If all that fails, we can talk about antennas with greater gain...
Hopefully, you've already found that loose or corroded connection or optimum aim-point and all is well...
Post #20 of
http://forum.tvfool.com/showthread.php?t=901 is my stock answer regarding grounding the permanent installation.