Niel,
Yes, it's obvious you know your way around radios. Your advice was definitely of help, and I want thank you for it. Everyone sees things from a slightly different angle, and that's why it's good to hear from others like you. Everything is still speculation until we get back to the mountain.
To answer your previous question regarding our other power source. Yes, the radios operate year-round, 24/7. If they're not up 98% of the time, we take a financial hit on our maintenance contract. Needless to say, solar has been turned off at these sites until we resolve the issue.
Our other power source is redundant propane fired internal combustion generators. They are diesel engines retrofitted to burn gaseous propane. The engines spin 3 phase generators whose output is fed to a rectifier cabinet and converted to 24vdc. The system is rather unique in the aspect that the generator output is regulated by a control loop monitoring the dc, after it has been rectified. In addition, there are seperate controllers for each generator which monitor the battery charge level and call the generators to start only when batts are depleted. A charging profile is then completed with bulk and taper segments. A PLC acts as the master or "mother" controller, and allows only one engine at a time to respond to the low voltage "call to start", toggling between the two so they both get exercised. In addition, each engine's controller also controls environmental dampers/fans to regulate hut temperatures, and the PLC has the ability to override/control these elements as well, should the engine controllers fail to do so.
Extensive sensor wiring monitors voltages, current draw, rpm's, charge current, temperatures, oil pressure, smoke, fire, fuel levels, weather, as well as data from various site loads, and all is reported back to a master station over microwave links/telco circuitry where the data is recorded into a database for analysis. This circuitry also allows for manual engine control and modification of operating parameters in response to site specific conditions.
The "on demand"nature of the system works well with the solar. In summer, the engines rarely see a "call to start" condition, saving lots of $ in propane cost.

Since it sound like your having large costs shuting down your solar, I would suggest you purchase at least ONE WX-MPPT from xantrex before your next visit to debug. Since it has Class B FCC certification you know, worst case it will fix the issue, if all else should fail and by the sounds of it, even at 600.00 each, it cheaper than having to haul in fuel

At that frequency, a majority of the EMI could be common mode noise (current) coming out of the leads of the unit. some things to try:

Ferrite beads--They make lossy split ferrite beads (like those bumps on cords for computers/digital electronics) that you can clamp on each pair of signal wires (or entire groups of wires--especially +/- heavy current wire pairs, don't put individual wires through separate inductors, large currents will saturate the cores--making them useless, and the added differential inductance can cause signal/stability issues; doing pairs only blocks the common mode noise but keeps signal quality high and circuit stability unaffected). Don't put ferrite beads on your lightening ground connections--it will just make those ground runs more inductive and not pass lightening strikes to ground as well.

Look at the catalog for size and frequency range covered. You want "lossy" ferrite beads designed for emissions control so that they convert the RF energy into heat. There are also larger solid rings you can wrap an extra turn of cable through that can help too. It has been too many years for me to remember vendor names/part numbers but here is a search that may help.

http://clusty.com/search?query=ferrite+filters+beads

Shielded cable--using braided shielded cable (with the jacket grounded back at the noise source--MX60 metal case) will help too. Generally, only the first 10' or so (longer will not hurt) needs to be grounded. The capacitance of the braid will short out most of your RF to ground.

You can get various braided "socks" (some can be wrapped around existing cable) or get the wide flat braid used for grounding applications--it is hollow and you can pull your wires through the hollow center section and build your own braided cable harness.

In the old days, you could also get old braided conduit from ships/navy surplus too.

Regular solid conduit will work fine too (grounded at source of noise). I would not trust flex conduit (spiral type) unless you do your own testing first (needs solid ground connection entire 10'--spiral connections along length may not work as well).

-Bill

i'm not criticizing the second opinions you are getting, but just reafirming how it parallels my advice. yes, i can overlook something so no problem in more minds at work. that's why you initially posted as you're not dumb, but you can overlook something too.
that's quite a nice system you guys devised there. i do have to agree with solar guppy and the owner in that you should order at least one new xw controller if your next bout with tackling the rfi doesn't pan. time is essential and you want a good working system there. yes, there are costs, but it could be more costly otherwise as the system has been compromised by this rfi. at the very least you would have a backup controller and the cure now would allow you time to play with the causes/cures of the interference at a more relaxed pace.

Thanks for all the input. I've got some split ferrite chokes in hand, and just got approval to buy a Xantrex. Might get to the mountain by the end of this week.

AKremotes, have you found anything over the months that you could share with us if you're still reading the forum? i know that everything is frozen solid up there now as it's frigid here now too as we just had a record low the morning of the 6th of december 2007 of 7 degrees f setting a new record low for the date by 4 degrees as it was previously 11 degrees f.

Neil,
We ended up moving the controllers. Tried the split ferrite cores but no improvement. Never even tried shielding. By moving the controllers to the generator shack we got several benefits-
-less than 10' of cable to the batteries
-generator building is skinned in metal, so it's somewhat of a faraday cage
-further from the tower/antennas, which is where the noise was getting into the radios

The sites that have never given us a problem will probably not get moved.

glad to hear you had resolved the problem without too much trouble. keep warm up there.