Hi there,

wait a minute. We're mixing two different aspects.
Here, Craise was talking about power distribution, not about surge protection. And from that point of view, it's pointless to just use one socket of a double or triple set.

Yes. But the so-called surge protectors that you can buy for about $5 at every electrics dealer for home and office use ... well, I mean those that you plug between the wall socket and the power cord, they're normally nothing more than two varistors connected from the active terminals to earth. Maybe an indicator lamp as an extra gadget. The same applies to most power strips with integrated surge protection.

Varistors, however, don't react to a surge of current. They react to the voltage across them. When the voltage across a varistor exceeds a certain threshold, it becomes conductive and shorts the over-voltage that way. As a consequence, the circuit breaker will disconnect. Well, hopefully.
As you explain below, it depends on the amount of energy "annihilated" in the varistor how it ends: The varistor survives a low-energy peak and still operates normally; it melts internally and produces a permanent short after a medium-energy peak; and it is blast to pieces on a high-energy peak, after which it effectively isn't there any more.

But still, you need high voltage across the varistor to get an effect. High current flowing past it doesn't do anything.

That's a different strategy, it turns single-ended voltage spikes into common-mode ones, which is usually less dangerous for the equipment.

Yes, of course. But I've seen that only for industry use yet, not available to the wide public.

No, rarely between the L and N wire, only from L and N to earth. I know because I dismantled a few of them already. Maybe there are a few with differential mode protection, but they're rare. The same is true for most power strips with integrated surge protectors.
Differential mode varistors are more commonly found inside the electronic equipment as their own local protection.

[X] Doc CPU

It depends ... on whether equipment powered from one protector is cross connected to equipment powered from the other. If they are NOT (ever) connected, then the 2 protectors can be plugged in separately, providing separate zones of protection. Then do not connect anything between those zones.

If you do have "whole house" protection, then violating the above won't have as much of a consequence. If you do NOT have "whole house" protection, violating the above has greater consequences.

Divide up your equipment into zones which will not be interconnected. The VCR connects to the TV, I'm sure, so those cannot be divided and must be in the same zone. If the computer connects to the TV, then it, too, must be in the same zone.

One zone is protected by ONE protector. But cascading protectors can be done in limited cases for devices with limited connections.

Safer is to have everything in an interconnected group protected by ONE protector, and wired as close together as possible. Whole house protection does this for everything in the house, but starts to lose some protection when long wires are involved. That's where point of use protection (a surge protector power strip) helps.

Destructive surges are a current. It has near zero voltage if properly connected to earth. Or it has a massive voltage if hunting for earth destructively inside. Anything that tries to stop that current source only causes voltage to increase.

A varistor never trips a circuit breaker for too many reasons. For example, breakers take milliseconds. A surge is done in microseconds. Second, if any breaker opens to stop a surge, then voltage increases as necessary to blow through that breaker. Protection is never about stopping a surge. A varistor with near zero joules does not absorb a surge.

If a varistor melts, then it creates a major human safety threat. Varistor must never fail catastrophically. A degraded varistor should have no visual indication. Even the 'failure light' cannot report a varistor's acceptable failure mode.

What causes most damage? Common mode transients. Differential mode is often trivial - mostly made irrelevant by what is already inside electronics. Destructive common mode transients (ie hundreds of thousands of joules) are either connected harmlessly to earth. Or that current is inside creating high voltages as necessary to get to earth destructively via appliances.

Meanwhile a 'whole house' protector addresses all modes.

Unfortunately protectors were disassembled that do not claim to protect from typically destructive surges (that were only L-N, N-G, L-G varistors). Assumed was that ineffective solution is THE solution. Protection is about where hundreds of thousands of joules (and less surges) are absorbed. Protection is always about where energy dissipates. How many joules will that L-N varistor absorb?

Critical to protection is a 'whole house' protector that connects 'low impedance' to earth. An effective protector makes a 'less than 10 foot' connection to earth. So that hundreds of thousands of joules are absorbed harmlessly outside. So that the current is not inside generating high voltages that will even blow through an open circuit breaker or a millimeters gap in a switch. Protection means that current is not inside the building. Therefore the least expensive solution is also the best.

If he does not care about having surge protection, then he can pretty much wire things up any way he wants, within the constraints of overloading circuits.

You have a misunderstanding of how surge protection works. Surges can happen without the level of current that would trip a circuit breaker. Even if surges did have sufficiently high current to trip a breaker (it can happen), by the time a breaker trips, the surge has already caused the damage.

Circuit breakers have two means to trip. One is a magnetic coil in series with the circuit that will trip the mechanism if a sufficient field strength exists for sufficient time to pull the mechanism. The energy level required to do this is much higher than the energy level needed to damage a computer. A sharp 7000 volt spike induced from a lightning strike nearby will rarely, if ever, trip a breaker, yet will damage many components in a computer if it reaches there.

The other circuit breaker trip mechanism is thermal, building up heat over time from slight overloads. A short circuit at the outlet can take a couple seconds for this to trip (which is why the magnetic one is there). Surges won't ever trip a thermal element.

This is not how they operate.

While a massive surge could melt it, it is also possible that it will end up as an open circuit this way, too. This is not the mechanism of protection.

The "short circuit" is a result of its resistance curve. At higher voltages it merely conducts current at a near zero resistance. As a result of the near zero resistance, it actually dissipates LESS energy and can survive many surges, even big surges.

Surges come in two forms: differential mode and common mode (or some combination of these)

Common mode surges have the same voltage on each wire. This results in a very wide magnetic field, and thus is impeded by very high self-inductance. The length of wiring in a typical home branch circuit will seriously impede the common mode surge, especially at the riskier high frequencies.

Differential mode surges have opposing voltages on pairs of wires. Normal power distribution is a differential mode voltage, and is why there are two wires. The magnetic field is small, generally confined to the space between the wires (and trying to push the wires apart mechanically). Differential mode can propagate long distances and poses the greatest risks. Protection from this mode is easy by just shorting the two opposing voltages together and providing a current path between them. That diverted current then flows down the other wire back to where it came from on the first wire. The end result is that you have a COMMON MODE surge propagating beyond. This means the voltage relative to the world will rise on the protected equipment, but damage is very rare from this as long as there is no other path for that voltage to produce a high current. If the protected equipment has other connections that are NOT on this same voltage during the surge, then a high current will flow through the equipment and you will have damage.

You will have a high voltage in differential mode.

If you do not have a high voltage, but do have current flowing, that is common mode. It would start with a common mode voltage. Then if that voltage can reach somewhere with a different voltage (for example a TV set with its own separate wire to ground), then you have the high current. A high voltage out to the end of a wire with no other connection does NOT have high current.

This is why the power company workers up in an insulated bucket lifter are able to safely touch those wires above 10000 volts. There's voltage, but no substantial current.

There are a few. Shop around. All of the strips I have on my computer shelves are the metal cased ones which I bought at Walmart (large consumer commodity discount store chain in the USA).

What I have not yet found is a surge protector that has a 8P8C (wrongly but commonly known as RG-45) connector and MOVs connected to it. I have found one in a plastic case. We used those in our offices since each office had both power AND a CAT5 drop. Each office became its own zone having exactly ONE point of protection covering both the power and CAT5.

Typical all-purpose home surge protectors have connectors for phone line (4-wire) and cable (F-type coax) in addition to power. Be sure your phone and coax wiring coming in are protected. A "whole house" protector especially needs these AND they need to come into the building at or near (within 3 meters) of where the power comes in, and all be connected to the ONE protection device.

You dismantled a cheap one. That's likely because most of them are cheap ones.

With those you get a little less protection, but you still get some. If the voltage is between L and N, then it will get applied across BOTH MOVs in series. That won't short until an even higher voltage is applied.

The surge protector in the computer room at work has 10 large MOVs:
A-B, B-C, C-A, A-N, B-N, C-N, A-G, B-G, C-G, and N-G (yes, even N-G can have surges). A,B,C are the three hots of the three phase power going in. FYI, the computer room itself is the protected zone since it is a multi-tenant building with "wrong wiring" (it has an entry point surge protector that was made useless because of a lot of other wiring that was done in the building by people that were not directed by someone with surge protection knowledge). So we had to shell out another $20K for a premium surge protector for the computers, which were then totally isolated from anything else by using fiber optics to the router and switches located in the telco closet which is NOT protected due to all the metal wiring involved there (so we had to treat that part as losable).

1 members found this post helpful.

One of the big advantages of whole house protection is that they are generally well designed and well manufactured (from companies like ABB, Eaton (Cutler-Hammer), Schneider (Square-D), Siemens, etc). Still, even there, cheap knockoffs are around, so beware.

Another advantage is that whole house protection is practical to connect to earth at one point and the right point (multiple earth connections decrease protection). This allows a substantial amount of the surge to "bleed off" to ground (because the ground generally has the base zero reference voltage), reducing the total amount that gets in. Shorter, heavier, wires to ground, and good earth electrodes (multiple here is OK when they are all bonded together at one point of protection) improve this earthing protection and leave even less of the surge to reach sensitive equipment.

In the rare cases where ground voltage rises due to a close lightning strike, or a nearby downed power company distribution wire, the surge coming in via the earthing electrodes is common mode, since only one wire point is involved. This would not be the case with multiple separate earthing points (this is why doing just one is correct).

"theres a wall socket- i can choose the bottom outlet or the top one, but can both be used for surge protectors?"

A common wall socket is for this issue the same exact thing. Top or bottom are same.

Either could be used to then plug in a surge protector. If you want, both could used to plug in two surge protectors. If they have a switch on them you could use one for 24/7 stuff and the other for limited time use stuff.

Hi,
Not if the straps have been cut to provide independent circuits. Note that the NEMA/NEC (USA) does allow a single switched socket within a duplex to be isolated. One is live(can be switched) and the other socket can be switched but both must be from the same source(feed) at the distribution/circuit box. Sure the normal circuit is that both are live with proper neutral & grounds.

Hi there,

yes, I agree. But a current needs voltage as a driving force. Without voltage, there's no current. That's Ohm's law - and it remains true even if capacitive and/or inductive elements are involved. Only the mathematical approach is different (differential equations or complex maths).

Right, that's why a varistor doesn't try to stop the surge current - instead, it shorts it, giving least possible resistance.

Yes it does. It does a few milliseconds after the surge occured, because then it might be permanently shorted (destroyed). From Wikipedia:
Typically, manufacturers don't consider real-life surges for selecting the right varistor type and size, but only the mandatory requirements of the test procedures. They pass the lab test, but usually fail sooner or later.

I agree to everything you say in this section. However, a varistor may be able to withstand the energy of the surge, but when there's a peak current of several 100 amps, it's very likely that the small little semiconductor disc is permanently damaged. And then it gets blown to pieces by the regular operation voltage that it shorts. Very often, you can see that the enclosure popped open, sometimes there's nothing left but the short terminal wires. That's not a myth, it's experience from countless safety tests I did.

Only to units that have a connection to ground. Fully isolated, ground-free equipment is almost immune to common-mode transients.

Hundreds of thousands of Joules? Now I see we're talking about very different phenomena. The kind of surge I had in mind is caused by a few hundred volts of over-voltage and produces a surge current of a few hundred, at most about a thousand amps. Together with the extremely short time in the range of microseconds, we get an energy of about 1..10 Joules.

[X] Doc CPU

The only reason the cut outs would be removed is that there is a live circuit and one with a switch. It is not to code to run two circuits to that jack without a warning of dual source. They both have to be powered by the same circuit and therefor the same for the OP's request.

A transient of hundreds of volts often will be ignored by a protector. A ballpark number: a protector remains inert - does nothing - until voltage exceeds 330 volts. If a surge is an even larger current, then voltage might rise even to 900 volts.

Appreciate the relevant concept. A utility is electricity from a voltage source. That means current increases as necessary to maintain that voltage. A surge is electricity from a current source. That means voltage increases as necessary to maintain that current.

Protection means that current creates a near zero voltage. Therefore earth ground and its connections are critical for protection. It a current is not connected low impedance (as short as possible) to earth, then the resulting hundreds or thousands of volts may be inside hunting for earth destructively via appliances.

Destructive surges (that typically cause damage) may occur once every seven years. A number that can vary even within a town. If having frequent surges, then dimmer switches, GFCIs, or clocks are replaced daily or monthly. Due to protection even in those devices, your concern is the infrequent transient. Earthing and a 'whole house' protector mean that rare transient does not overwhelm superior protection already inside appliances.

An effective protector must be sized to remains functional. Posted were responsible manufacturers who provide these (ie General Electric, Intermatic, etc). A 'whole house' protector should be rated at least 50,000 amps. A number that estimates a protector's life expectancy.

A protector that fails shorted is grossly undersized and a human safety threat. That varistor must not trip a circuit breaker. It trips a tiny thermal fuse. A circuit breaker leaves a surge connected to adjacent appliances. The thermal fuse disconnects only a varistor. And if the thermal fuse does not disconnect fast enough, then a fire (human safety) threat exists.

Just don't make metallic connections between two groups of equipment or your surge protection is reduced. Just how much depends on the circuit wiring to the duplex outlet. If they are separate circuits, or shared neutral, the risk goes up quite a bit. But if they are the same circuit connected together in the outlet box, then the risk goes up only a little bit.

For a big surge, yes breakers can be tripped. The MOV may or may not be damaged, anyway.

But ... the circuit breaker is NOT a part of the way the surge protection is performed for the equipment being protected. It is protected by the fact that the MOV shorts differential mode surges leaving only a common mode voltage rise (which is also much smaller when a properly wired whole house point of entry protector is used).

Very high currents will indeed destroy MOVs. They can even destroy plain copper wire. I've seen heavy duty conductors (that normally can handle 400 amps) vaporized by lightning (over 100kA).

In most cases, the destruction of the MOV will be replaced by an arc that continues to carry the current. Arcs like that will continue to be destructive to the protector enclosure, but ironically can still protect the sensitive equipment until the source energy is no longer available. Arcs tubes are another form of protective device because arcs are very low resistance.

Some great information in this thread

Would the knowledgeable contributors like to comment on (primarily in the context of lightening-induced transients):

1. The importance of low resistance earth (ground) paths.
2. The effectiveness of 1:1 ("isolating"?) transformers.
3. The effectiveness and desirability, within a junction box, of connecting L-N, L-E and N-E terminals with wire and cutting the wire -- with the intention of providing a normally open circuit spark gap that will be jumped by high voltage differences. EDIT: presumably an arc tube is a more evolved way to implement this.

Hi,
Never said anything about dual source;Whenever you cut the strap you are isolating the sockets for control purposes. The source remains the same if you follow the NEMA/NEC. The switching is just convenience.

Not low resistance. The term is low impedance. For example, a 20 meter AC wire inside the wall may be well less than 0.1 ohm resistance. That same wire can be 120 ohms impedance during a surge. A trivial 100 amp surge might create something less than 12,000 volts (100 amps times 120 ohms) between both ends of the wire due to 'impedance'.

Impedance is why the wire to earth must be as short as possible. No sharp bends. Not inside metallic conduit. Etc.

A best isolation transformer is already inside electronics. Rated typically more than 1000 volts. That galvanic isolation already exists for human safety.

Protection principles were demonstrated over 100 years ago using spark gaps. And later using GDTs. Today, we use avalanche diodes and varistors that have much better life expectancy. GDTs remain used in circuits that must have low capacitance (ie radio antenna). Spark gaps and GDTs typically are not used on power circuits.