I have a question for those of you (Pauluk) that have a background in the higher frequency (Pauluk) side of electricity. (I am hinting to Paul but I am sure there are others of you that have a great deal of knowledge in this area!) The situation is this. A broadcast studio. We are installing a ground reference designed for high frequency ground rejection in the MegaHZ range produced by digital television broadcast equipment. The installation requirements are not an issue as I have done this on a previous job for the same client in an even larger scale. The grounding scheme goes like this. Around the exterior of the building we sink 6 Linconite XIT 2" diameter chemical copper ground rods with bentonite and "Gem" back fill. Four at the corners and two mid span of the longest sides. We connect all the rods together with .02 thick 12" wide copper strap SILVER SOLDERED and bolted at the rods and at each splice. Then we t tap the 12" strap and run it into the rack room where it runs to all the rows of equipment racks. Each rack contains a copper ground bus that a 2" wide .02" thick strap is SILVER soldered and bolted to and spliced to the 12" strap. All the equipment in the racks is fed with an equipment grounding conductor to the power supply per article 250. The ground bus is simply a reference for high frequency noise. The intent of the copper strap verses copper wire is to obtain more surface area for high frequencies to flow. We bond this grid at one point per the NEC.(Although I am not sure if that is required anymore due to the nature of the system?) My question is, Does this sound like a ridiculous over kill or is this something that is truly needed for this type of equipment? I don't know a lot about the nature of the equipment so I really cant tell if this is good engineering or just a waste of time and money.
Nick, It's only overkill if you miss the dance of little sparklers over your expensive equipment. It's also deadly necessary to bond it to your equipment, otherwise you have 2 separate systems with two separate impedances, and that we call a lighting trap. There may be one or two rods overkill there, but I doubt it.
A heads up-- Practices like that are par for physics/pulsed-power facilities. The only caution is that they must always remain "supplemental" and never a substitute for NEC-based methods. Datacomm facilities are sometimes noted for asking for "alternate" methods.
[This message has been edited by Bjarney (edited 06-02-2002).]
I grew up in Florida and as being into radio equipment and have studied lighting. something that I have noticed about lighting. It can strike a 30' tree right next to a 120' tower. This phenomenon perplex me to the hilt. so I set out to try to find out why, as I installed several towers at the same height and with these towers I installed different grounding systems. some were simple ground rods at varying lengths and another was chemical grounds and using. the old flash bulb lighting detector, to see which ground attracted the most strikes. and after a few years. I discovered that lighting will strike almost anything that at the time of the strike will match the impedance of the frequency of the stroke. yes I did say frequency. remember that you don't need straight ac to get a impulse through a transformer or a sine wave. just a pulse of a given duration will work the same way. like in a ignition system for a car the points just pulse the coil to get flux for it to work. with knowing this then the pulses of lighting can be recognized as ac or dc pulses, rather than just dc. and if you understand that at certain frequencies what would be a short to dc ( or a low impedance path) will be an open to this ac. and lighting will go to something else that has a lower impedance path for the frequency of the stroke at that time. and this frequency is different with every strike. this is why we see strikes hitting deferent things that don't make any sense. especially when we have grounded something to the hilt and to only have a strike hit something that it shouldn't have. like the cable system or TV antenna that only had a little 3' ground rod with a #12 aluminum wire to it that's 100' long.
But I did show that if a ground has a low enough impedance like in a chemical ground, the lighting struck it the most. so I too use it the most when I have to ground a tower. I have at my house five 10' 5/8 ground rods driven down to where the top of the ground rod is 2' below the grade and have a five gallon bucket with holes drilled all through the bottom down over the each ground rod with round salt blocks in the buckets and as it rain's the water from the rain will filter down through the lid over the salt blocks and into the ground around the rod. and I have seen lighting hit my antenna and not even bother anything inside like my computer or TV's or even the CB that was hooked up to this antenna. and they were even up and running at the time of the strike. of corse I had to go and change my underware as when this happend I was trying to get the key into the lock to get into my house when it hit lol. :Wayne
[This message has been edited by hurk27 (edited 06-01-2002).]
Be Fair, Be Safe Just don't be Fairly Safe
Re: Ground Reference#10210 06/02/0205:39 AM06/02/0205:39 AM
Nick, I'm not familiar with the references to Linconite and Bentonite (U.S. trade names?), but I get the general idea.
Is it overkill? Probably not.
As the frequency goes up grounding becomes increasingly critical in HF applications. At upper UHF frequencies and above, even a right-angle bend in a regular conductor can introduce enough inductance to render a traditional ground ineffective.
Although a normal analog TV video signal is limited to a bandwidth of around 4.2MHz (U.S. system), digital processing is a whole new ballgame. The short rise-time and square-wave nature of digital pulses means that the signals are extremely rich in harmonics. Add that to the high sampling rates necessary for digital TV, and you could easily be dealing with frequencies up into the GHz range.
At the satellite station where I worked we had C-Band up/down links operating at 3 to 6GHz (3000 to 6000MHz). Grounding on that equipment was also by a complex array of rods, mats, and copper straps. There's the point you made about using flat straps to increase surface area to help deal with skin effect, and also the point that it's necessary to minimize any inductance in the ground path. In many cases this would mean parallel straps. (As an aside, normal coaxial feeders start to become very inefficient at these frequencies, hence the use of waveguides.)
At Ku-band frequencies (12 to 14GHz) used extensively for satellite TV these days, the problems become even more pronounced, so much so that it almost becomes black magic!
As has already been mentioned, yes, you should most definitely bond this ground system to the electrical supply ground.
Re: Ground Reference#10211 06/02/0211:31 AM06/02/0211:31 AM
Thank you everyone. Let me add a little more to the pot. On this job we are in a relatively high lightning zone. For that reason a lightning protection grid is going to be installed as well. My feeling on this (and please tell me if I am off base here) is that we need a separate counterpoise for the lightning protection system. The equipment reference is silver soldered together and I would think a lighting strike running through this system would melt the solder connections rendering it useless from that point on. My plan is to run a lightning protection grid as designed by the supplier around the building. Shade the trench about a foot and run the copper strap. The lightning protection grid, copper strap reference and building service will all be bonded at one point right outside the electric room. I think this will keep all the gremlins going where they are supposed to.
Nick, Very similar to airport grounding. The lightning system was separate from the electrical system.....kinda. The lightning counterpoise ringed the buildings (often a mile or a shade over in length) with 2/0 cu., ground rods fairly often in the system, but at least bonded to the building steel every 50', and the down leads were bonded to the building steel as they penetrated the roof.
The electric service was pretty standard grounded, to the water services (don't forget the sprinkler system) and the building steel, which is how it tied to the lightning system, plus all the extra electrical equipment ties to the bldg steel, xfrmrs etc..
Now don't forget the ductwork, and you can call the FAA for a flight path