So you're saying the tower is 50' away from your equipment?
Probably more like 30' horizontally.
This is BAD as the building where the equipment is probably has a DIFFERENT Earth Ground.
They are actually bonded, probably #2
Grounding radios to towers is NOT a grounding system. Steel is a much higher resistance than copper so any stray current, static charge, or ESD charge will take the nice copper Ethernet cable instead. Yes there should be a MAIN ground bus on insulators at the base of the tower that is connected to all the ground rods or halo system for the tower. As well the tower should be connected to the ground rods as well but but to your bus which is insulated. You always insulate your grounding system except where it bonds to the antennas as you do not want to drain the tower in the middle of charge but rather hope that charge follows the steel down as much as possible so as not to overwhelm your ground path. You only want to drain the steel at your antennas. You're hoping charge in the middle will go down the tower instead of up and then into your ground wire as the tower legs are also bonded to the ground rods.
Now I get it--two independent ground paths--one for the radios/antennas and one for the tower steel. I suppose digging down and tying in the antenna ground path as low as possible to the halo and rods (and enhancement material, etc.) is preferable?
There should be a BOND wire between the tower ground system from this main bus to the service ground rods and a ground bus in your cabinet but since this is 50' away it should almost be #1 gauge wire which is 1" thick.
There should be a #2 green running up the tower to an insulated ground bus near your antennas and a #6 from that bus to each antenna/radio.
If the radios are spaced over 30 or 40 vertical feet, where to locate the (single?) insulated ground bus? High, low, mid-range?
You want to make sure you have ample Ethernet Service loops so your Ethernet path is at least 10% longer than the intended ground path.
If they are run the additional 40-50ft to the cabinet (as opposed to the existing LA's at the base of the tower) they're a lot longer. And why the 10% rule of thumb--is this empirical in any way, i.e. corresponds to the conductivity difference between copper and steel?
You need understand that ground potential differences is just as dangerous to your equipment as ESD and Static discharges.
Follow the guidelines in these posts (Read these posts several times):
http://forum.netonix.com/viewtopic.php? ... =30#p13447
The idea is that all equipment is within 5 ohms of each other as far as ground potential and no matter where there is stray voltage or CURRENT that the intended path is shorter and less resistive than the Ethernet cables.
I don't suppose testing ground potential difference under ordinary conditions means that it would look like the same under the conditions of a nearby lightning strike. You just hope your grounding system is up to the task of distributing charge when something like that happens....
Thank you for your time and passion as the Grounding Guru and Evangelist.