Subject: Re: lightning protection of roof towers In article rec.radio.amatuer.misc the question is asked: >I have recently come into posession of a roof mounted tower (about 10 ft. >thrust bearing, uses guy wires attached to roof, etc.) and want to >mount it on the roof of my two story house. My problem is that the tower >will be one of the highest structures around, and is almost certain to be >struck by lightning sooner or later (at least I want to be prepared). >I would like to know the best way to ground the tower. > >In the ARRL handbook (and also the NEC - National Electrical Code) there >is a lot of discussion about protecting the feedline by grounding the shield >of the coax, using arrestors, switching them to ground when not in use etc. >There is also some discussion of grounding a REAL tower. But I haven't been >able to find much on roof towers. What I want to be sure (or at least as >sure as possible) of is that if the antenna/tower gets hit, the lightning >strike doesn`t arc through the house, etc. >Some specific questions I have are: >1. What size of ground lead(s) to use for the tower? I realize that >bigger is better but the cost of enormous cable is prohibitive. What >is reasonable if you EXPECT or PLAN on a direct hit? The NEC says number 8 solid is the minimum size for the downlead of a lightning rod. Consider this the smallest wire you can use. Since lightning is a pulse that has RF components through VHF, skin effect matters. A better downlead is 5 inch wide copper flashing run as straight as possible to ground, but in no case having a 90 degree or sharper bend in the path. >2. How many? If you use the correct size, and do the bonding correctly, one is sufficient. If you aren't going to be able to inspect the run frequently, you might consider paralleling a number 8 solid wire with the 5 inch strap so that if corrosion opens one connection, you will still have a better path to ground than through your roof. >3. What paths do I run them? The obvious answer would be the shortest, except that the shortest path is on the opposite side of the house from the shack and the feedlines. Would the resulting 'cage' effect be good or bad? Normally, you want the shortest and *straightest* path to ground. However, ground loops can be disasterous. Make up your mind to establish a single ground point somewhere on your property and connect all ground runs to this one point. Make sure the utility grounds are also bonded to this point. Don't ever depend on earth conductivity to close a ground circuit and don't ever use a feedline shield as a ground conductor. It's ok to have multiple grounds, but they must tie to the single point ground via low inductance, high conductivity cabling in a *star* configuration, not in a daisy chain. Daisy chaining grounds is a recipe for disaster. Setting up a ground cage is an extreme measure of protection. To do it right is not simple. Doing it wrong can be dangerous. If you can't analyse all the current paths properly for ground loop effects, don't do it. Bring *all* cables into your shack via a "ground window", that includes power and telephone. A ground window is a single small area, usually a rack panel, copper plate, or the like, that each wire is attached to via an arrestor of the proper type for that kind of cable. The ground window is then connected to the master single point ground by a heavy strap or cable. This assures you that all cabling entering the shack will be at near the same potential during a strike (plus or minus the breakover voltage of the arrestors). With quality arrestors, you shouldn't ever have a voltage differential of more than 200-400 volts across your equipment. Most equipment will tolerate this kind of voltage for the few milliseconds required. Note that your entire shack may be elevated to several thousand volts above ground due to the resistive drop of your single ground lead, but as long as *every* part of the shack is elevated the same amount, no net current can flow. >3. Do I need to stand them off from the roof? It's generally not necessary, but it won't hurt. Remember that a direct lightning stroke will typically flow 4,000 amperes or more for several milliseconds. The wire can get quite hot. I've seen number 12 solid *vaporized* by a lightning stroke. That's why you want to use at least number 8 wire or heavy strap. >4. Do I need to attach to each leg of the tower? That depends on the tower construction. If the tower is welded, connecting to one leg is sufficient. If it is bolted or riveted, inspect it carefully for loose connections and use bond straps to make it a single conductor. >5. What about the guy wires? Generally, *don't* ground the guy wires. If the tower is going to carry an HF antenna, you'll likely want to break guys with insulators anyway to avoid pattern disturbances. >6. Anything else I should know? A single ground rod is not an effective ground in most soils. Use at least three no closer than 6 feet apart in a triangular arrangement. Bond them all together with your heavy strap. If you have *metal* underground water piping, tie to that too. Remember, *star* configuration. No ground should connect to another except at the *single* point connection. And no downlead should connect anywhere but at the single point. Always *mechanically* bond ground wires and straps. Don't depend on *any* soldered connection. Solder *will* melt when lightning strikes. The NEC says there should be *no* splices in a ground run. That's perhaps a bit extreme. A *good* mechanical and electrical splice should be ok, but rather safe than have the insurance company void your claim. >I know that there's no guarantee when it comes to lightning strikes, >but it is my family in the house and I want to be sure I have done >everything I can to protect them. That's a good attitude Dave. Lightning kills more people each year than any other weather related thing. Besides, the better ground field can help your HF signals *every* day. The members and HQ staff would like to thank the following people for their contributions to this information file: Gary Coffman, KE4ZV gary@ke4zv.uucp Send any additional information or changes to mtracy@arrl.org. 73 from ARRL HQ.