Grounding an Outdoor Antenna

I'm wrapping up plans for installing a 22 dbi parabolic antenna on my roof to ensure a robust cell signal from T-mobile. I know I need to ground the coaxial cable itself (because it will be connected to a battery powered device, and will not function as an RF shield unless grounded).

I've read that I should "ground the antenna" but this makes no sense to me. It seems that if I actually grounded the antenna, I'd be draining away all the signal it so carefully collects.

I also need to surge protect what I can, but none of the hardware is very expensive, unless I use Ethernet cables instead of Wifi. (Then a strike could zap my computers.)

To "ground the antenna," should I simply ground the antenna mast?

Can such a ground wire be connected to the ground wire used for my coaxial cable?

Would I be better off driving my own ground rods?

Do I need two ground rods or one?

Should I instead run the ground wire over to Georgia Power's ground?

Or would that set up conditions for some sort of 60 cycle hum that would drown out my feeble cell signal?

Is grounding the antenna mast the same as adding surge protection?

I know that I have to match "impedance" in all the components, and that this is a trivial matter of selecting components that have matching Ohm numbers, but I wonder if anyone can explain what impedance is and why it causes problems when there is an "impedance mismatch."

You can ground the antenna mast for lightening protection. Grounding the antenna won't affect the signal. The parabolic dish is just a reflector. It works that way whether grounded or not.

One ground rod, close the the antenna mast would be best. You don't want the current to have to travel a long distance to ground. Don't connect to the power ground since some of the lightening surge could enter you house and cause damage.

You can protect the coax by running it through an inline coaxial surge protector. The case of the protector connects to the antanna ground.

High frequency lines and components must be impedance matched, otherwise you will have reflections and standing waves in the line, which can seriously degrade the signal.

If you do have an existing ground rod for your house electrical system, bond the the antenna mast ground to it (a buried #6 gauge wire will do). Don't the run the wire inside the house.

Any grounding electrode added for communications systems is required to be bonded with a 6 AWG or larger conductor to the building or structure grounding electrode system in accordance with 800.40(D), 810.21(J), 820.40(D), and 830.40(D).

The grounding of power and communications systems to the same single point ground helps in reducing the difference in voltage potentials between the systems. This becomes very important where these different systems are integrated together in sensitive and expensive electronic equipment.

Click to expand...

http://www.mikeholt.com/mojonewsarchive/GB-HTML/HTML/2002NECGroundingCommunicationsSystems~20020701.htm

http://www.reeve.com/Documents/Articles Papers/AntennaSystemGroundingRequirements_Reeve.pdf

GregTuve said:

I've read that I should "ground the antenna" but this makes no sense to me. It seems that if I actually grounded the antenna, I'd be draining away all the signal it so carefully collects.

Click to expand...

First some basic electrical concepts. A 200 watt transmitter is connected to a long wire antenna. Touch one part of that antenna to be shocked by over 100 volts. Touch another part to feel zero volts. Why two completely different voltages on the same wire? Because even wire is an electronic component. Electricity is not same at both ends of a wire. A concept that must be understood to also understand why that antenna must be "earth" grounded (not safety ground). And why a ground wire also is as short as possible, no sharp bends, etc.

Code requires an antenna to be earthed. A wire from that antenna to an earthing electrode is required. Code also requires the antenna's coax be earthed where it enters a building. To the same earth ground used by AC electric and telephone. Not safety ground (ie wall receptacle ground). To earth ground.

Why are safety ground and earth ground so different? Read the first paragraph.

What determines all protection? Not a protector. Protection is defined by the quality of what does all protection - earth ground. Especially critical is that service entrance earth ground. Better earthing means better protection. Both for humans and transistors. "Better" does not necessary mean, for example, that a water pipe is better than earth ground rods.

Important is to route a coax cable via the service entrance so that it connects low impedance (ie 'less than 10 feet') to 'single point earth ground'. The ground (and not a coax protector) mostly determines protection.

Coax is often 'protected' by a ground block. Maybe $4 in Lowes. An example:
http://www.dbsinstall.com/N-images/Whatis/Driploop-1.jpg
But unfortunately, that ground wire violates good installation practice. Therefore compromises protection. The ground wire is curved; is even bundled with the incoming cable. Two mistakes.

A shielded cable works fine if the shield is not earthed.

Earthing the dish is not for protection from a direct lightning strike. Far more elaborate protection is required for that. Earthing is for static charge buildup, near lightning strikes, etc. You are not going to have large earth currents from the dish. Installation is a lot less critical than for a municipal cable system.

The requirements of the National Electrical Code would be the minimum. You need a ground block where the coax enters the house, which allows the shield to be earthed. Plus the metal parts of the mast need to be earthed. Both MUST be connected to the power earthing system. The connection from the point of entry to the house should be "as short as practicable". That would be, in most cases, through the house. (But a drip loop, for instance, is entirely reasonable.) Best is if the cable enters near the power panel. (There is a separate provision if the wire from the ground block to the earthing system is over 20 feet.)

You do not want an isolated ground rod. In the event of a near lightning strike, or a strong surge earthed by the power system earthing system, separate ground rods can be thousands of volts different.

--------------------
Coax has a "characteristic impedance" If you feed an RF signal into an infinite length of coax, the ratio of voltage to current is determined by the coax impedance. The impedance is determined by the diameter of the shield and center conductor plus the characteristics of the insulation.

The best power transfer from an antenna to the coax and from the coax to a receiver is when the impedances are matched. If not matched, you do not get maximum power transfer, and some signal is reflected back down the coax.

For different types of coax with the same impedance, some types have higher losses per foot of coax than others. Losses depend on the frequency being used - higher frequencies have higher losses. Lower loss coax is more expensive.

Hey everybody,

Thanks for all the responses. My "service entrance" (the point where power lines enter my house, I think) will be on the opposite side of the antenna, so I do plan on driving a ground rod right under the antenna mast. (I've already bought the rod.) I'll make sure the ground wire for the mast runs straight to the rod without bends.

Of note: The device to which the broadband signal will be routed is battery powered much of the time. I think some people (not necessarily you guys) have argued that the coax shielding should not be grounded in that situation, because it invites lighting to travel from the antenna down the coax... but other people have told me that coax shielding doesn't shield at all unless it is grounded. I'm unclear on what a "safety ground" is... the third hole in the outlet?

There seems to be debate among your members as to whether or not I should run a second ground wire from my new grounding rod into the service box to join with my house's grounding system.

I view a direct lightning strike as improbable because there are many tall trees near my house. Maybe that's relevant to something...

The point about the grounding of the reflector portion of the antenna being immaterial is well taken. It certainly will reflect signal the same way whether grounded or not, but there is a collector of some kind at the focus of the parabola. I guess grounding the coax from that won't ground the center lead so that shouldn't "drain away the signal." The reflector itself could wind up electrically isolated.

One nice thing about running the antenna signal directly into a battery-powered Wifi access point is that there is basically no way for a surge to leap from the access point to any other electronics (disregarding direct strikes, which can tunnel 20 feet through sand).

I invite additional advice and comment. Thanks again.

GregTuve said:

There seems to be debate among your members as to whether or not I should run a second ground wire from my new grounding rod into the service box to join with my house's grounding system.

Click to expand...

Code says that second ground rod must be connected to the service ground. Many ignore that requirement.

Grounding does not attract lightning. Lightning finds a path from a cloud three miles up to earthborne charges four miles distant. A shortest electrical path is down to earth at your house. And four miles through earth. Lightning will strike no matter what you do. Best is to make that path via conductors that cause no damage. Even wood might conduct lightning. Provide lightning with a better path so that is does not use wood or anything else inside the building.

Safety ground is also defined by code as equipment ground. It performs human safety functions. That receptacle third prong is part of safety ground. And is electrically different from another ground - earth ground. Your concern here is earth ground.

Earth ground must be installed to meet safety code requirements. And to exceed those requirements to protect electronics. For example, code says nothing about low impedance (ie 'less than 10 feet', no sharp wire bends, ground wire not inside conduit, etc). Those requirements upgrade earthing to also perform electronics protection.

So, will lightning find earth via a battery powered WiFi? Apparently nothing recharges its battery (otherwise it is not isolated). A path to earth might be via that Wifi unit to earth via wood. Or via other more conductive materials such as linoleum or concrete. But if the dish and coax is earthed, chances are that current will take that better earthed path; not via the Wifi box.

Apparently you believe a direct lightning strike is improbable (ie no strikes to any neighbors in over a decade). And any freak strike would only damage a battery powered box. Any freak strike might best be made irrelevant by the earthed dish. The dish would act light a lightning rod - maybe even protecting the building. And may not result in damage elsewhere inside the building. Well, that judgement (the risk) is best made by you.

I don't think there really is any debate about having to bond the rods together, it's required by code and reduces possible damage from ground potential rise from a near strike or utility line fault.

http://www.esgroundingsolutions.com/about-electrical-grounding/what-is-ground-potential-rise.php

GregTuve said:

Hey everybody,

Thanks for all the responses. My "service entrance" (the point where power lines enter my house, I think) will be on the opposite side of the antenna, so I do plan on driving a ground rod right under the antenna mast. (I've already bought the rod.) I'll make sure the ground wire for the mast runs straight to the rod without bends.

Click to expand...

In the event of a close lightning strike, with your dish ground rod in line toward (or away from) the strike point and the power earthing, the dish rod and power earthing system can be many thousands of volts different. It is the same as the dead cow in nsaspook's link (where the distance is between the cow's legs).

Similarly, with a 1,000A surge current earthed on the power service earthing system, if there is a single ground rod with a near miraculous resistance to earth of 10 ohms, the building "ground" is 10,000V above 'absolute' earth potential. In general, 70% of the voltage drop away from the rod is in the first 3 feet. Your dish rod will be at least 7,000V different from the building ground. It is why separate isolated rods are a bad idea.

The NEC wants the ground wire from the coax entry ground block to the power earthing system to be 20 feet long or less. If not, you can add a ground rod at the coax entry point and bond the new rod to the power earthing system.

Ground rods are one of the worst earthing electrodes.

GregTuve said:

Of note: The device to which the broadband signal will be routed is battery powered much of the time. I think some people (not necessarily you guys) have argued that the coax shielding should not be grounded in that situation, because it invites lighting to travel from the antenna down the coax... but other people have told me that coax shielding doesn't shield at all unless it is grounded.

Click to expand...

Lightning will hit the dish whether it is earthed or not.

You will not have lightning protection for the dish. That is not what the minimal earthing required by the NEC (and discussed here) is for.

The shield prevents outside fields from reaching the center conductor and imposing noise. Doesn't matter if the shield is eartheded (unless you create "ground loops").

GregTuve said:

There seems to be debate among your members as to whether or not I should run a second ground wire from my new grounding rod into the service box to join with my house's grounding system.

Click to expand...

The NEC requires all earthing systems be bonded. The minimum bonding wire size is #6, which is much larger than the wire required from the dish supports and entry ground block.

It would seem like the bonding wire would eliminate the voltage difference that can be created by an 'event' (above). But surge currents have relatively high frequency current components. That means the inductance of the wire is more important than the resistance. You can have thousands of volts between the ends of the bond wire.

GregTuve said:

I view a direct lightning strike as improbable because there are many tall trees near my house. Maybe that's relevant to something...

Click to expand...

The minimal dish earthing will NOT provide protection from a direct lighting strike. Ask ham radio operators with high antennas what is required.

Suppose you have an "average" lighting strike of 20,000A and the rod has the near miraculous resistance to earth of 10 ohms. The rod will be at 200,000V relative to 'absolute' earth potintial. There will likely be arcing across the surface away from the rod. Will a rod that is not bonded protect the house? Not with 200,000V.

You must bond the dish ground rod.

Grounding the braided jacket of the coaxial wire is one thing; grounding the parabolic reflector is another.
The first is to preserve the signal received, the second to provide a path for lighting strike.

If the antenna location takes a long cable to the cell phone, losses will not allow much improvement. If inside your attic, lighting discharge protection will not be that much effective.

I had to use low loss ethernet adapters for my ~40dB setup, which takes 25 feet from the antenna to the USB port here :

http://s588.beta.photobucket.com/user/Innernet/media/HFuWave.jpg.html?sort=3&o=15

[Posted via airborne internet]

The cows provided a particularly memorable example of how voltage varies with distance from the strike. Thanks, nsaspook; it confirmed that I understood what I was reading elsewhere. Bud's detailed reply was also quite helpful as was the final confirmation that coaxial shielding works whether or not it's grounded.

I was interested to note that the NEC requires the coaxial cable's braided jacket is to be grounded with an insulated wire... "The grounding conductor shall be insulated... and it shall be as short... and as straight as practicable [800.40]." --"practicable" who talks like that?

I think that truly robust lightning protection may be too expensive for this project. Now that I think about it, we did have a near miss about 12 years ago that fried our home security system, so I may have underestimated the probability of direct strikes and near misses. I am careful to disconnect electronics when storms approach, but staying on top of the computer/wiring configuration well enough to be sure there isn't an overlooked copper back door to your computer is a worrisome challenge.

Another concern is the ground at the service entrance. When this house was constructed 60 years ago, the standards were almost certainly lower. For example, one of the outlets in this room was not grounded until I personally ran a ground wire from it to a pipe of some kind to stop complaints from my UPS. If the safety ground is not "earthed," at my house, I suppose it must travel some considerable distance back to Georgia Power's grounding network. I may need to do a better job of earthing the service entrance, assuming this has been done at all. Should I call an electrician? Seems that way since I'm also not sure exactly where to attach the wire that bonds my grounding rod to the house's ground, assuming there is one. A few years ago, some professionals (Mr. Sparky) put in a new circuit breaker box and charged 1500.00 for half a day's work, too. There are probably other outlets that aren't safety grounded.

One person (not here) said I should create 2-3 loops (each a foot in diameter) in the coax before it enters the house. This had something to do with "inductance." How does this help protect against surges from nearby lightning strikes? I was guessing that the loops generated a magnetic field, causing resistance that encouraged most of the electricity to take some other path. If my guess is correct, the loops would even help discourage electricity from traveling down the inner conductor of the coax. Is that close to right?

Finally, it looks like adding an MOV (metal oxide variable resistor) to the coax line before it reaches the Wifi axccess point will be affordable and of some benefit. A recommendation would certainly be appreciated.

GregTuve said:

Another concern is the ground at the service entrance. When this house was constructed 60 years ago, the standards were almost certainly lower.

Click to expand...

Best protection also costs less. Earth ground to provide better protection AND meet code is one or two ground rods in earth just outside from the first breaker box. A ten foot ground rod and less than 10 feet of #6 copper wire is significant earthing. Note the 10 foot number. 20 feet is for human safety. Less than 10 feet is required for transistor safety. Code defines human safety. Transistor safety means exceeding code requirements - a shorter wire for lower impedance.

Grounds discussed here are irrelvant to Georgia Power grounds. Especially since their grounds are more than 10 feet distant. Your main breaker box (one closest to the meter) must be earthed as required by safety codes. A 'less than 10 foot' connection from its bus bar to the earthing electrode becomes your building's single point earth ground.

You had damage 12 years ago? A direct lightning strike to wires far down the street is a direct strike incoming to every household appliance. An event that may occur once every seven years. Therefore one 'whole house' protector on AC mains is essential protection for everything. The IEEE says it is not perfect. A properly earthed solution will do maybe 99.5% to 99.9% protection. Cost is about $1 per protected appliance.

Unplugging is a least reliable solution since surges often occur when unexpected, when not home, when asleep, etc. Dishwasher, GFCIs, furnace, and even smoke detectors also need protection. How are those unplugged? Many reasons why earthing one 'whole house' protector is a superior and least expensive solution. So simple that all parts are even sold in Home Depot and Lowes.

A one foot loop, at best, might add another 0.2% protection. If it does serious surge blocking, a loop completely blocks even higher frequency internet and TV signals. Obviously not. If a loop does sometime useful, then tying coax into tight knots is even better. And also blocks weaker and higher frequency TV signals. Knots do not even block radio frequencies.

If the loop has sufficient inductance, then the recommendation also gave a number in Henries. Ball park inductance in microhenries:
(4dN^2)/3
where d is feet and N is number of turns.

What stops current flowing down the inner conductor? A best explaination from Richard Harrison was entitled "Lightning Arrester" in the newsgroup rec.radio.amateur.antenna.

Coax, inside, rejects common-mode propagation of lightning energy. Coax, outside, needs good grounding to make a good path around (bypass for) protected equipment.

Click to expand...

An MOV between two wires might avert damage from transverse currents. But destructive surges are, as Richard Harrison notes, common-mode currents. That means protection is defined by connections to earth.

A MOV only does something useful when connected low impedance to earth. Replace the MOV with a wire. That connection from shield to earth is even better protection. And also required by code for human safety.

The cow is another example of why single point earth ground is necessary; to minimize or eliminate GPR problems.

The original discussion was limited to a battery powered Wifi repeater that has no other connections. Apparently you also want protection for everything else. Backdoor connections will exist if any wire enters a building without first connecting to single point earth ground. That solution is not only the best. It is also easiest and least expensive.

For surge protection - excellent information is at:
http://www.lightningsafety.com/nlsi_lhm/IEEE_Guide.pdf
- "How to protect your house and its contents from lightning: IEEE guide for surge protection of equipment connected to AC power and communication circuits" published by the IEEE in 2005 (the IEEE is a major organization of electrical and electronic engineers).
And also:
http://www.eeel.nist.gov/817/pubs/spd-anthology/files/Surges%20happen!.pdf
- "NIST recommended practice guide: Surges Happen!: how to protect the appliances in your home" published by the US National Institute of Standards and Technology in 2001

The IEEE surge guide is aimed at people with some technical background.

westom said:

Best protection also costs less. Earth ground to provide better protection AND meet code is one or two ground rods in earth just outside from the first breaker box. A ten foot ground rod and less than 10 feet of #6 copper wire is significant earthing. Note the 10 foot number. 20 feet is for human safety. Less than 10 feet is required for transistor safety. Code defines human safety. Transistor safety means exceeding code requirements - a shorter wire for lower impedance.

Click to expand...

It is extremely likely earthing electrode(s) were installed when the house was built. Even more likely they were present when the service was upgraded.

Ground rods are about the worst earthing electrode. But they are easy to install.

The author of the NIST surge guide has written "the impedance of the grounding system to `true earth' is far less important than the integrity of the bonding of the various parts of the grounding system." The minimum distance you want, in particular, is from entry protectors (phone, cable TV, ...) and a common connection point on the power earthing system. You also want a short distance from the service panel and the common connection point. That minimizes the voltage between power and phone/cable/TV/... wires. An example of a connection that is far too long is in the IEEE surge guide starting page 30.

In the event of ground potential rise from earthing a strong surge or a near lightning strike, much of the protection is that all wires rise together. (That is why I don't like separate rods, even if bonded.)

A coax entry ground block does not limit the voltage between shield and center conductor. A MOV would not be used on coax, but other components can be. IMHO protection on coax at the point of entry is more important for cable, which has much higher surge exposure. But a ground block with surge protection should be pretty cheap. It needs to be rated for the frequencies you are using.

westom said:

Therefore one 'whole house' protector on AC mains is essential protection for everything.

Click to expand...

Service panel protectors are a real good idea.
But from the NIST guide:
"Q - Will a surge protector installed at the service entrance be sufficient for the whole house?
A - There are two answers to than question: Yes for one-link appliances [electronic equipment], No for two-link appliances [equipment connected to power AND phone or cable or....]. Since most homes today have some kind of two-link appliances, the prudent answer to the question would be NO - but that does not mean that a surge protector installed at the service entrance is useless."

Service panel protectors do not, by themselves, prevent high voltages from developing between power and phone/cable/... wires. The NIST surge guide suggests most equipment damage is from high voltage between power and signal wires. An example is in the IEEE surge guide starting page 30 (where a service panel protector would also provide no protection).

westom said:

The IEEE says it is not perfect. A properly earthed solution will do maybe 99.5% to 99.9% protection.

Click to expand...

The 99+% figures are from the IEEE "Green" book and are for lightning rods. They have nothing to do with service panel protectors.

westom said:

A MOV only does something useful when connected low impedance to earth. Replace the MOV with a wire.

Click to expand...

Plug-in protectors are effective without a low impedance to earth - they do not work primarily by earthing surges. As explained in the IEEE surge guide starting page 30, they work by limiting the voltage from each wire (power and signal) to the ground at the protector. The voltage between wires going to the protected equipment is safe for the protected equipment. The guide explains earthing occurs elsewhere.

If using a plug-in protector all interconnected equipment needs to be connected to the same protector. External connections, like coax also must go through the protector.

Whether or not you use one or two ground rods depends a lot on the type of soil you have. Here in S. Florida a ground rod is almost useless unless you can get it down to the water table.
So my best advice is to check with a local Electrician for the standard practice in your area.

John

thewhitedogdad said:

Here in S. Florida a ground rod is almost useless unless you can get it down to the water table.

Click to expand...

A FL couple suffered numerous direct lightning strikes to a wall. So lightning rods were installed. Lightning again struck that wall; not lightning rods. That wall contained plumbing that connected to pipes buried deeper in a more conductive soil. Lightning rod had only been connected to a single 8 foot ground rod in sand.

Solution was simple. More and deeper rods that connected to that more conductive soil. Lightning stopped using that walll and its plumbing to connect to a better ground.

Same applies to all protection; lightning rods or protectors. Lightning stikes to connect to earth. Protection is defined by a connection to and quality of earth ground. Water table was irrelevant. More rods connected to more conductive soil meant lightning need not use a house or its contents to make a destructive electrical connection.

Anyone building a new home can install best protection if installed when footings are poured. Ufer grounds were pioneered in muntions dumps because direct lightning strikes must cause no damage. Earthing works everywhere. Earthing that exceeds what is required by code (and electricians). Every protector or lightning rod is only as effective as its earth ground. Protectors and lightning rods (ie ESE devices) a dedicated earth ground are ineffective.

Every professional organization bluntly makes that point. IEEE and NIST citations say same. For example, from the NIST on Page 19 (of 24) says

A very important point to keep in mind is that your surge protector will work by diverting the surges to ground. The best surge protection in the world can be useless if grounding is not done properly.

Click to expand...

Useless? Because protection is always about connecting to a best earth ground.

Thanks for all the new replies. You guys don't have to sell me on binding my ground rods; that's already done.

Westom: Your comment about transistor safety and the ten foot rule is helpful. You seem to be saying I need a ground rod every ten feet or so attached to the #6 gauge wire that extends between my antenna mast ground and the service entrance ground. I unplug my computer equipment whenever I leave the house or go to bed, so it's kind of reliable but--as you say--that won't protect my furnace or my flat screen.

Westom: As for the MOV, I was assuming I would be able to find a ground lug on the device and that it would shunt surges to my earthing grid and ultimately to ground. I don't think you disagree with that, but I've been having trouble finding a MOV device designed for coaxial cable. Maybe there isn't one. Mostly they're embedded in power strips. The current plan (from a ham radio guy) involves using a gas discharge device as the first line of defense (because they handle large loads well) and a MOV inside the house (because it has a very short "clamp time.") I'm clear that an MOV needs to be properly grounded to work well, but if you're saying MOV's aren't suitable for coax even when directly and properly earthed, that's new to me and I'll need clarification. Bud seems to be saying the same thing.

Westom: No, no. I'm sticking with my plan for grounding a battery powered wifi device and all the antenna stuff that feeds it signal. At this point, it looks like I might learn enough to protect everything in my house from lightning, but that's not the immediate goal. I'm not sure what "Common mode" and "GPR" mean.

Bud: You say a MOV would not be used on COAX but other components can be. What are those components? Quarter wave thingies? Also, I have identified a "ground block" (a gas discharge tube in this case) and have verified it's suitable for the carrier frequencies I plan to pass through it. Your quote from the NIST guide is most enligtening: Yes [a surge protector at the service entrance is adequate...] for one-link appliances (some electronic equipment), but no, [it's not adequate protection for] for two-link appliances (equipment connected to power AND phone or cable or....)--typically the most expensive ones. As I see it, the two links attached to a powered telecom device function as the front and back legs of a cow, forcing the device to play the role of a grounding wire. (If that doesn't make sense, don't worry about it; it's too long to explain what I read about the cows.)

All: It was a ham-radio operator who suggested putting a MOV on the coax. He also recommended the loops in the wire, but Tom knows the math that says this measure is basically pointless. Regulations regarding the binding of ground rods say the heavy wire should not have any sharp bends in it. Apparently, lightning surges don't like going around sharp bends. I figured the extra loops had something to do with this but the recommended curve radious was about six inches, too large from what I've seen so far--because this curve radius is described as the minimum for an exterior link between grounding rods. People have suggested several articles on whole house lightning protection--but I'll have to read those after this post.

All: Thanks in part to the feedback I've gotten here at allaboutcircuits.com, I've been able to come up with a rather detailed drawing of what's needed. Maybe that should be posted here somehow when it's done. One outstanding question has to do with impedance in signal meters. I need to hook my parabolic antenna (50 Ohms) to a signal meter so that I can be sure it's pointed dead at the signal source (a cell tower). If there is an impedance mismatch between the signal meter and the antenna, will that cause a constant level of signal attenuation (a situation that wouldn't interfer with aiming the antenna) or would the mismatch cause a fluctuating level of signal attenuation? In other words, when I select the signal meter for aiming my antenna, do I have to make sure it's 50 Ohms?

Can anybody recommend an MOV or similar rapid-clamping surge protector for coax? One provisioned for grounding? One that passes 1700 to 2100 MHz signals? The "spark gap"/Gas Discharge Tube device takes a while to ionize the gas that leads to the ground.

GregTuve said:

Westom: Your comment about transistor safety and the ten foot rule is helpful. You seem to be saying I need a ground rod every ten feet or so attached to the #6 gauge wire that extends between my antenna mast ground and the service entrance ground.

Click to expand...

First, an application note may make things more obvious. The antenna and building are treated as separate structures. Each must have its own earth ground. Any wire leaving either structure must first connect to that earth ground either driectly (ie ground block on a coax cable) or via a protector (how to earth the center wire in a coax):
http://www.erico.com/public/library/fep/technotes/tncr002.pdf

Note a ground wire interconnecting each structure's earth ground. That increases protection as well as mildly improving earth ground for both structures. But most important is every incoming wire connecting to each structure's earth ground.

Assume a coax, instead, interconnects two buildings. A direct lightning strike to one building is a direct strike to electronics in the other building. If the coax is not earthed at both ends where it enters each building. Same concept.

Second, do not earth equipment. That would make equipment a potentially destructive path to earth. Instead, connect a surge (not equipment) to earth ground. Earth the incoming wire distant from equipment.

Third, a coiled wire does supplement protection only if that wire first has a low impedance connection to earth. For example, microhenries does not stop any surge. But it slightly increases protection WHEN a surged wire is first connected to earth. Then microhenries encourage lightning currents to take the better path to earth. Microhenries in a coil do nothing if the earth ground does not exist. The Ham may have forgotten to mention the required earth connection.

Telcos also prefer their electronics to be up to 50 meters distant from an earthed protector. That separation alone does no protection. But that separation (increased impedance) makes a surge more want to find earth outside; less want to find earth destructively inside. But only if the wire is earthed outside - before entering a building.

We want earthing wires to have no sharp bends, no splices, and not be inside metallic conduit. So that impedance to earth is as low as possible. Then any minor impedance increase in the wire to appliances (longer wire, sharp bends or coils) can increase protection. Again, only if that low impedance connection to earth exists.

Fourth, MOVs have high capacitance. Therefore would be harmful to signal propagation on coax cables. Protector for coax cables must be a low capacitance device. Gas Discharge Tubes (GDTs) or equivalent devices (semiconductor) are typically used. But at gigahertz, only GDT are best utilized. See Polyphaser.com for examples of such devices. (Also view Polyphaser's application notes; highly regarded in the industry.) As with all protectors, that GDT must be earthed to perform protection.

Fifth, what is the difference between one-link and two-link appliances? If a 'whole house' protector on AC mains is the only protection, then all protection has been compromised. However, if all incoming wires are earthed to a common ground (either driectly or via 'whole house' protectors), then all appliances (one, two or many links) are fully protected.

For example, your phone line already has a 'whole house' protector. Therefore an answering machine (a two link device) is not protected. Why? Because AC electric and other wires also enter without the short connection to earth. A surge on AC mains would use the answering machine to find earth destructively via the telephone line. Once every incoming utility has proper earthing, only then is the answering machine and everything else protected.

In simpler terms, everything has best protection as long as every wire inside every incoming cable is properly earthed. Then no surge current is inside hunting for earth destructively. The "one-link" statement plays games with your reasoning. If only a 'whole house' protector on telephone (no AC mains protector), then many one-link appliances (ie dishwasher) remain at risk.

Sixth, common mode is a current that seeks earth ground. AC powers an appliance by current entering on one wire while leaving on the other. Common mode is different; a current enters on both wires (in the same direction). AC power cord has two wires. A common mode current is traveling in the same direction (incoming) any or all wires. And that current must also be outgoing from the appliance on some other path. Maybe via phone wires. Maybe the wooden desk. Maybe wall paint. Or a nearby air duct. Voltage will increase as necessary so that current incoming on both AC wires finds some other destructive outgoing path to earth.

Incoming from the parabolic dish. Outgoing via a battery to earth via, well, what is that wifi device sitting on? A desk? A common mode current might use that desk as an outgoing connection from battery to earth. Common mode - the outgoing path is to earth.

GregTuve said:

Westom: Your comment about transistor safety and the ten foot rule is helpful. You seem to be saying I need a ground rod every ten feet or so attached to the #6 gauge wire that extends between my antenna mast ground and the service entrance ground. I unplug my computer equipment whenever I leave the house or go to bed, so it's kind of reliable but--as you say--that won't protect my furnace or my flat screen.

Click to expand...

Westom is saying that surge protection is not effective if it is not connected with a 10 ft or shorter wire to an earthing electrode.

There are at least a couple of problems with that. In the case of cable/phone/... entry protectors you want a short connection to a common connection point on the power earthing system. You want to minimize the voltage between power and signal wires.

And it is not relevant for plug-in protectors which do not work primarily by earthing surges.

GregTuve said:

Westom: As for the MOV, I was assuming I would be able to find a ground lug on the device and that it would shunt surges to my earthing grid and ultimately to ground. I don't think you disagree with that, but I've been having trouble finding a MOV device designed for coaxial cable. Maybe there isn't one. Mostly they're embedded in power strips. The current plan (from a ham radio guy) involves using a gas discharge device as the first line of defense (because they handle large loads well) and a MOV inside the house (because it has a very short "clamp time.") I'm clear that an MOV needs to be properly grounded to work well, but if you're saying MOV's aren't suitable for coax even when directly and properly earthed, that's new to me and I'll need clarification. Bud seems to be saying the same thing.

Westom: No, no. I'm sticking with my plan for grounding a battery powered wifi device and all the antenna stuff that feeds it signal. At this point, it looks like I might learn enough to protect everything in my house from lightning, but that's not the immediate goal. I'm not sure what "Common mode" and "GPR" mean.

Click to expand...

Common mode is if the cable shield and center conductor (for example) rise together with reference to something else, perhaps the building "ground". If a surge is between the center conductor and the shield it is not common mode (it is differential or transverse mode). Power line surges are differential mode (between H and N) at the panel because of the N-G bond at the service.

GPR is ground potential rise. GPR is what kills the cow - the potential of the ground rises, and rises higher at one leg than the other. GPR is in the IEEE surge guide starting page 30. GPR is also what I described when I said "Your dish rod will be at least 7,000V different from the building ground." GPR is why I don't like separate ground rods, even if bonded.

GregTuve said:

Bud: You say a MOV would not be used on COAX but other components can be. What are those components? Quarter wave thingies? Also, I have identified a "ground block" (a gas discharge tube in this case) and have verified it's suitable for the carrier frequencies I plan to pass through it.

Click to expand...

I agree with westom that for signal circuits MOVs have too high a capacitance. Gas discharge tubes are one device that is used. It has a short time lag for the "discharge" to be established (as you note later in your post). There are much faster semiconductor devices. One is a Transzorb.

GregTuve said:

Your quote from the NIST guide is most enligtening: Yes [a surge protector at the service entrance is adequate...] for one-link appliances (some electronic equipment), but no, [it's not adequate protection for] for two-link appliances (equipment connected to power AND phone or cable or....)--typically the most expensive ones. As I see it, the two links attached to a powered telecom device function as the front and back legs of a cow, forcing the device to play the role of a grounding wire. (If that doesn't make sense, don't worry about it; it's too long to explain what I read about the cows.)

Click to expand...

Correct. Power wires are the front legs. Phone (or cable or ...) are the back legs.

A separate ground rod gives you widely separated back legs (even if bonded).

If all entry protectors - phone/cable/TV/dish/... - have short ground wires to a common connection point on the power earthing system, and there is a short distance from the N-G bond at the power service to the common connection point, then a service panel protector is more likely to protect "two-link appliances". The 20 ft limit for the ground wire in the NEC is too long.

GregTuve said:

All: It was a ham-radio operator who suggested putting a MOV on the coax. He also recommended the loops in the wire, but Tom knows the math that says this measure is basically pointless. Regulations regarding the binding of ground rods say the heavy wire should not have any sharp bends in it. Apparently, lightning surges don't like going around sharp bends. I figured the extra loops had something to do with this but the recommended curve radious was about six inches, too large from what I've seen so far--because this curve radius is described as the minimum for an exterior link between grounding rods. People have suggested several articles on whole house lightning protection--but I'll have to read those after this post.

Click to expand...

A surge is a very short duration event. That means the surge currents have relatively high frequency current components. That means the inductance of the wire is more important than the resistance. In the IEEE surge guide example, a 30 ft ground wire from a cable entrance ground block to the power service earthing has a 10,000V drop from end-to-end when earhting a surge coming in on cable. Sharp bends increase the inductance at that point. A 6" radius is good.

(Impedance of the wire is why bonding a separate ground rod is not as effective as you might think.)

GregTuve said:

All: Thanks in part to the feedback I've gotten here at allaboutcircuits.com, I've been able to come up with a rather detailed drawing of what's needed. Maybe that should be posted here somehow when it's done. One outstanding question has to do with impedance in signal meters. I need to hook my parabolic antenna (50 Ohms) to a signal meter so that I can be sure it's pointed dead at the signal source (a cell tower). If there is an impedance mismatch between the signal meter and the antenna, will that cause a constant level of signal attenuation (a situation that wouldn't interfer with aiming the antenna) or would the mismatch cause a fluctuating level of signal attenuation? In other words, when I select the signal meter for aiming my antenna, do I have to make sure it's 50 Ohms?

Click to expand...

IMHO an impedance mismatch would give you a constant error which would not be a problem in aiming the antenna.

Westom and Bud: Your posts have been invaluable. Particularly with the latest two, I think I'm really starting to get a handle on concepts like impedance. The "ham radio" operator I told you about is actually the son of a Ham Radio operator. The failure to mention the presence of a gas discharge tube before the loops he recommended is mine, not his. The most knowledgeable people agree that typical MOV's won't help in this situation because of high frequency of the carrier signal I'm monitoring and for other reasons I don't fully understand. Special thanks to Bud for taking the time to define GPR, "transverse," and "common mode" before I had time to look these things up. My design seems complete except for the possibility of inserting a Transzorb device and improving the grounding of my house. More's the pity: I have to take one more look, but I don't think my house is properly grounded at all. This place is an electrician's nightmare with a mix of fuses, circuit breakers, and flying splices. My puny water-pipe ground for the outlet this computer is plugged into may be the only one. I'll take Westom's advice not to earth the equipment itself to heart. That would be a big oops.

GregTuve said:

I think I'm really starting to get a handle on concepts like impedance.

Click to expand...

Just to be clear, the 50 ohms impedance of a coax cable is completely different from impedance created by a long wire or sharp wire bends. That 50 ohm cable impedance is more accurately called its characteristic impedance.

The other impedance is defined by other parameters such as that coil that creates a microhenry.

For RF, a transzorb (sometimes called an avalanche diode) typically has excessive capacitance. Another type semiconductor device that is similar but has less capacitance include products from this manufacturer:
http://www.semtech.com/circuit-protection/low-capacitance/