Have any of you encountered a situation where changing the ground (between utility ground and grounding rod, for example) source resulted in a loss of about 7 dB on both RX and TX?

For those of you who do solar powered installations, do you connect the negative (-) from the battery to ground? If so, do you use the same ground rod or a different ground rod for the antenna ground?

No one uses solar or wind power for remote sites?

I use solar sites quite often and always run my own earth cable to a ground spike or mulitple spikes depending on the size of the installation. So far I haven't noticed any signal loss doing this. I don't connect the negative lead to the spike. However I use steel housing for the radio equipment and lightning arrestors for the RF cabling. All of those are earthed to a 16sq mm cable or larger.

Utility ground probably has a better ground due to the fact that they have more conductive material in the ground, which travels horizontally through the ground as well as vertically, creating a better ground plain.
Soil conditions can effect your ground plain.
At the lakes where there is alot of rock, we throw the ground spool out into the lake to ensure a proper ground.

One should avoid using the utility ground, if and when avoidable.
In some locals it is against electrical code to run a ground from outdoors to an indoor panel.

I don't connect the negative lead to the spike.

You might be interested to know that the mounting holes on the Routerboard are connected to the negative lead, so if you use conductive material in those, your negative lead is connected to the spike.

As for me, my signal returned to normal by itself. I'm starting to suspect it might have been temperature differences over 12km.

Possible the Fresnel Zone got ya...

http://en.wikipedia.org/wiki/Fresnel_zone