the_simple_computer

Ode to Those Unsung Heroes of Surge Protection

Updated June 21, 2014.

This site is no longer being maintained so anything below could still be accurate, or very outdated.


I always thought this site may begin dabbling in hardware but I had no idea it would start like this. Summer is around the corner for you northern hemisphere people and that means thunderstorms. Fun stuff, but if a destructive power surge makes its way into your house and kills your computer or worse, you're gonna have a bad time.

Surge protection usually makes people think of lightning but that's actually one of the rarer causes of power quality problems for electronic equipment. More often would be an overcurrent swell resulting from an electrical fault somewhere outside your house, for example, due to digging or debris across power lines. Faults can happen indoors as well. Some reasons are old wiring that's falling apart at the connections and damage from rodents, earthquakes, fire and water.

But the most common source of power trouble is load switching at any point in the local power system, including inside your house. You've seen this when a high-current motor or transformer turns on and momentarily dims the lights or distorts audio or video signals. These surges can come from your air conditioning, vacuum, washing machine or refrigerator, or power source switching such as when a backup generator kicks in. Load switching from nearby buildings can even affect your power quality. Also when the power goes out and resumes again, whether a flicker or a 9 hour outage, surges are part of the package.


Cool Story

If power surges are of concern to you, proper surge protection is the only real solution and it's a quick & easy one if you just want a "buy this, install that" directive. As for the why, well that's not quick. Nor easy. So on the menu for today, we'll dive deep into the topic of surge protection and how to best apply it to protect your building's tender electronic insides. By building, I mean something residential in size and power draw like a house, workshop, pool shed, solar panel tree, etc. This is the not so quick, not so easy route but later I give some general recommendations for different price points and protection levels.

One thing though. This site has many readers around the world and while the concepts and best practices mentioned would be valid anywhere, this article is intended specifically for North America. There are some fundamental differences in power delivery, wiring and testing standards between there and other countries so, just know that not everything below may be applicable to you.


What Is a Surge Protector Really?

Surge protection for AC power
is easy. Adding signal lines
is where things get complicated.

A "surge protector" is a general term used for one of three main types of surge protective devices (SPD) as defined by the American National Standards Institute and Underwriters Laboratories standard 1449, 3rd edition. UL is one of several American Nationally Recognized Testing Labs (NRTL) which conducts safety tests for electrical devices. Some other NRTLs you may encounter are Intertek Testing Services, the Canadian Standards Association and TÜV Rheinland.

Relevant to us here are ANSI/UL1449 type 1, 2 and 3 devices. Type 3s, also called point-of-use or plugin devices, are the iconic surge protector strips we've all seen before—not to be confused with plain power strips which look identical. Other forms of type 3s are wall-mount multi-outlet plugs and outlet receptacles themselves can contain surge protection.

Type 1 and 2 SPDs resemble type 3s in function but aren't confined to points of use. They're often referred to as service entry or whole house protectors and just as your router's firewall guards your entire local network, a service entry SPD is the perimeter device for your house's electrical system. They come in various forms and many are UL listed jointly as type 1 and 2, meaning they underwent stricter testing and can be used outdoors. Some type 2 devices are an integrated circuit breaker/SPD combination but more common is a small box mounted next to, on to, or inside your main electrical panel.

Some service entry SPDs attach directly to the load center, others
through a conduit. Always you want the shortest possible wire length between them.

The active ingredient in most applications of surge protection is a metal oxide varistor (MOV) which, in an SPD, very quickly redirects incoming surges to the building's safety ground which then exits through the building's Earth grounding. This protects everything downstream of the surge protector, be it your server rack or your house's entire electrical system.

The bigger the MOV, the higher current it can shunt and the more smaller surges it can withstand before failure. Yes, MOVs do have a service life but a well made protector with properly sized MOVs can last decades. A dead type 2 SPD will be nothing more than a passive piece of eye candy but sometimes a failed type 3 can still be used as a power strip if the power disconnect fuse isn't blown...but you don't necessarily want to do that.

Of course there can be more than just MOVs inside these devices. Silicon avalanche diodes (SADs) and gas discharge tubes (GDTs) are other common surge suppression technologies. They're mainly used in signal and data line protection (television, ethernet, etc.) but can also be found in hybrid devices in combination with MOVs.

Then there are 'surge filters' which are basically line filters. They use transformers, resistors, inductors and capacitors to reduce and absorb surges and redistribute the remaining energy into the building's wiring at harmless levels, or into the air simply as heat. These are mainly point-of-use protectors but a few branch panel versions do exist. The actual filtering implementations are usually patented and differ between manufacturers, but probably the most well known is called Series Mode technology.


Types of Power and Surges

Phases and modes are two words you'll frequently see in SPD specs. Residential AC power is usually called split phase power, meaning there are two feeder wires (2 phases of the voltage cycle) coming from the pole, L1 & L2, along with a neutral wire (N). Confusingly, split phase is technically a form of single phase power but can also be called two phase or single phase 3-wire. Three phase power (L1, L2 & L3) is normally for big loads needed in commercial and industrial facilities but it can be found in residential too. Needless to say, a type 1 or 2 SPD must be matched to the building's power delivery.

There are two modes which describe current pathways between line, neutral and ground. Normal mode describes a voltage difference across any or all hot lines and neutral (L-L & L-N), also called differential, symmetrical or metallic mode. Then there is common mode, also called longitudinal or asymmetrical mode. This describes voltage across each hot line or neutral to ground (L-G, N-G). Some SPD specs say "all mode" and this means both normal and common.

Normal and common mode power fluctuations are constant occurrences but at such low voltages they amount to noise interference. Here's an example: see figures 1 and 2 in this test paper by Juice Goose where they record 48 hours of noise across both modes at their office in Houston, Texas. This can be tamed by SPDs with built in EMI/RFI filtering but that still leaves you vulnerable to surges in the passthrough frequencies.

Surges originating upstream in the power system will arrive to you as normal mode surges. Some scenarios are a lightning strike on a pole, recloser operation and external load switching; all SPDs for AC power have at least normal mode protection. Common mode surges are usually localized to your building's electrical system and while seldom damaging, they can cause signal disruptions in electronics using the building's safety ground (a 3 prong, grounded plug). Destructive common mode surges can be an electrical fault inside the building, or induced onto wiring or even entire grounding systems by large motors and lightning strikes.

North American ungrounded plugs aren't
physically capable of using the building's
safety ground.

It can get controversial as to whether a surge protector (type 3s especially) should have both normal and common mode protection. Manufacturers of Series Mode SPDs are the main proponents against common mode but it's rare to find a non-SM protector with only L-N and L-L protection. While not diverting surge energy to the safety ground has benefits, proper application of all mode and signal line SPDs surpasses that in total protection because SM SPDs aren't built to deal with any incoming surges on that safety ground.

It's not a case of asking which mode is better or harmful—they're two different sides of the same coin. It's rather a question of: In what scenario would you either not need or not want common mode surge protection in your setup? In North America, if you're in low risk area for lightning strikes and your building has a sufficient single-point Earth ground, then you could be fine with only normal mode protection in a service entry SPD since neutral and ground are bonded (connected) in the panel. Yet even these devices are uncommon. Of course anything with a 2-blade NEMA 1-15 plug, from modems to toasters to mobile device chargers, does not need common mode surge protection. Add additional complexity and consideration if you have interconnected equipment on different branch circuits and you can see there is a lot of room for confusion and implementation error.


Proper Surge Protection

In an ideal world, AC power and data signals (phone, television, ethernet, camera, etc.) enter the building at the same location. Power is guarded by a beefy surge protector at the load center (the main electrical panel) and signal lines each have their own SPD connected directly to the ground bar in the panel. This safely keeps surges out of the building from source lines while ensuring that power and signal have the same Earth ground reference.

In one room, the sensitive electronics
are confined to one block.

The next step is the building's interior at the points of use. All outlets in the building were tested for correct wiring so we're OK to move forward. Think about where your electronics are located. Are they spread across multiple rooms or taking up several areas in one room? Do like your main electrical panel does but on a smaller scale; whereas it divides different rooms and appliances onto different branch circuits all in one panel, you take the different electronics in a single room and confine them to their own SPD circuit, block or whatever term you prefer.

In another room is a second block
for a different computer. Note the
protector for the ethernet wire.

Assign a quality type 3 SPD to each block of devices you want to protect. For example, in room 1 are a modem, router, computer and printer while in room 2 are your TV, satellite and stereo receivers (or another computer like in the diagram). The coaxial lines for the satellite signal and cable modem get their own SPDs, as would the phone lines going to the satellite receiver or DSL modem. When possible, these signal SPDs are grounded to the metal chassis of the device itself, for modems and other things with plastic enclosures, the SPD's own ground pin is used.

Last, the ethernet lines must be referenced to their end point devices with the computer/printer/etc. on one side and router or switch on the other. Ground referencing is especially important if you have interconnected devices on different protection blocks.


...and How to Choose it

Whether you're looking for service entry or point-of-use protection, specs for AC power SPDs are similar. Start by choosing a service entry SPD and then supplement it with point-of-use protectors for sensitive electronics. This is known as cascading, basically similar to a network's security layers but don't forget the often unguarded back doors like electric pet fences, proximity lighting and power for irrigation, well or septic pumps. In this section are the specs to pay attention to when selecting a type 1, 2 or 3 SPD. Some manufacturers of service entry protectors to consider are Cooper Bussmann, Ditek, Eaton, Emerson, ERICO, Hubbell, Intermatic, Leviton, Richard Gray's Power Company, Siemens, Square D, Thomas & Betts Power Solutions, Total Protection Solutions and Transtector. Of course there are many more.

Maximum Surge Current

The max surge current rating is the most defining number of an SPD (NOT the joule rating!!) and will be listed in kiloamperes (kA). Generally speaking, a kiloamp value over 100kA per is the threshold between residential territory and light commercial or industrial use. Yet the real value in a higher max current rating is a longer service life because frequent smaller surges won't degrade the protector as quickly. To quote from Eaton's Guide to Surge Suppression [p.22]:


A properly constructed suppressor having a 250 kA/phase surge current rating will have a life expectancy greater than 25 years in high-exposure locations.

Unfortunately there are no industry standards to determine, test or verify a protector's max surge current rating. Often the maximum current capacities across normal and common mode pathways are combined which gives an artificially high number. So a 30kA surge capacity across L-N and L-G & N-G would be sold as a 60kA max current rating per phase, possibly even 90kA (though I've only seen that once).

Thus, it's important to always look for the max current per mode which serious SPD manufacturers list in their datasheets. Also look for the waveform and line voltage used to determine that current rating so you're comparing numbers across manufacturers on even ground.

Voltage Protection Rating

VPR is the ANSI/UL1449 3rd ed. standard for ranking let-through voltages, the remnant of the surge which get past the protector. Figuratively speaking, look at VPR as an SPD taking 10,000V and reducing it to 600V. Do not confuse or synonymize VPR with clamping voltage which has no derivation standard and can mean multiple things. See this 2009 whitepaper by Emerson Network Power for how VPR is determined.

VPR doesn't directly equate to the real-world voltage getting past the SPD, but it's worth mentioning that an ultra low let-through is not necessarily better and according to the ITIC Power Acceptability Curve, modern electronics can already withstand microsecond pulses in excess of 500% AC line voltage, so about 600V. LEA International adds that an effective SPD must lower lightning surges to within 678 and 848 volts for protection [p.2]. The ITIC Curve isn't a guarantee but it does indicate that modern electronics are not as fragile as we often think, as well as reason to not to fixate on the lowest VPR possible.

Test results from different certifications show
how an SPD handles multiple energy levels.

One more thing to keep is mind about VPR is that it's a UL-specific test and UL1449 is a test for safety, not performance. Other NRTLs do their own testing with different energy levels. For example, the IEEE uses a a C3 impulse wave at 20kV,10kA and a 10x1000 waveform which results in much more energy than the 6kV, 3kA 8x20 test used for UL's VPR. The more test results provided by a manufacturer against the more waveforms, the clearer picture you'll have of an SPD's capability.

Thermal Induced Disconnect

This takes several forms and you want an SPD with a combination of them to prevent thermal runaway. ASNI/UL1449 listed type 3 protectors will have a built-in circuit breaker, and unless your service entry SPD is the circuit breaker design, a type 2 protector will attach directly to its own double-pole breaker in your electrical panel. This is protection against overload and short circuits, fairly standard protection these days.

Equally important are thermal fuses which begin breaking down as they heat up due to excessive current. This opens the circuit and immediately stops power delivery. Some type 3 SPDs have one thermal fuse for the entire device so if the fuse blows, it won't power anything. Type 1, 2 and 3 protectors can also have one fuse per MOV so if the fuse dies, only that one varistor is left incapacitated which then lowers the total protection capability of the device. That's when the SPD's indicator light goes out and it should be serviced or replaced. An alternative to fuses is a thermally sensitive soldering compound. Only consider SPDs with some form of thermal cutoff and for type 3s, an internal circuit breaker too.

Parallel Varistor Configuration

This is for protectors which use multiple MOVs per mode pathway. You want those varistors arranged in parallel in the circuitry as opposed to being in series. This evenly distributes surge current to each MOV instead of only the first MOV in series constantly taking all the hits until failure. If MOVs can't act in unison, the protector's capability and longevity are diminished.

Electromagnetic Noise Filtering

Also referred to as sine wave tracking, EMI/RFI filtering is fairly widespread (to various capabilities) among even midrange SPDs. Most filter-type protectors operate entirely on this basis. To a point, noise in a power system is normal if for no reason other than wire impedance. Good noise filtering improves the lifespan of MOV and SAD based protectors while reducing the low energy, high frequency impulse and ringwave transients [p.2] which don't harm equipment, but can upset data signals (remember the recordings from Juice Goose). An example of this would be TV picture quality degradation by a running vacuum cleaner from an outlet on the same circuit.

Eaton's Guide to Surge Suppression discusses EM filtering and specifies a signal attenuation of 50 dB @ 100kHz "for premium performance" [p.15] based on the MIL-STD-220 filter performance testing standard, but expect to find around 40dB, even for quality protectors. SPDs with noise filtering are often listed as both a UL1449 Surge Protective Device and a UL1283 Electromagnetic Interference Filter. Nearly all point-of-use surge filters are strictly UL1283 listed.

Enclosure Rating

For a service entry SPD, the enclosure should be NEMA or IEC rated. That will indicate its overall build strength and if the protector can be used outdoors. Should something crazy happen, plastic or fiberglass don't dissipate heat and contain fireworks as well as steel. I personally prefer metal enclosures, even for type 3s and even though they're more expensive. If a load center's SPD cannot double as a SWAT battering ram, it has no business protecting my Ninja Blender. There's my bias.

If you decide on a protector with a composite body, you can buy a NEMA or IEC rated steel enclosure, powder coated with knockouts, and mount the SPD inside. On the other hand, if you choose a protector that isn't weatherproof and want to use it outdoors, there are both composite and steel enclosures for that as well. These are available from electrical supply stores for about $20 in the size you'd need.

Warranty

Normally I place little confidence in warranties but an exception must be made for service entry SPDs. Some manufacturers include full replacement warranties with their products, meaning that if the SPD or its modules were to die, be it due to a destructive surge—including lightning—or old age, they replace it and the cost to you is only shipping. The warranty life varies by manufacturer and this is a non-transferrable warranty so it applies only to the homeowner at the time of installation. A free replacement warranty for an SPD is a hallmark of high-end service entry protectors.

On the other hand, connected equipment replacement warranties are mostly bullshit, intentionally crippled with loopholes and contingencies. You won't find these from any service entry SPD manufacturer worth considering but they do plague nearly all type 3 protectors. Sure this isn't universal among manufacturers, but you're wasting your time if shopping for plugin protectors on the premise that you'll get $30k for a new home theater because the $30 surge protector supposedly didn't do its job. Do as you will, but I suggest that resources are more wisely spent on quality surge protection and grounding from the beginning.

Signal and data line protection

Lines for tv, phone, ethernet and such things are less straightforward than power because of the need to account for different frequencies, and operating and let-through voltages. Often comes even more complexity because now you have two points of termination per line (computer to router, camera to controller, etc.) usually in different rooms but sometimes in different buildings. Remember that ideally you want signals and data entering the building grounded at the electrical panel. Here's a diagram by MCG Electronics which shows a best case scenario.

An increase in voltage potential means an increase in current.
Image from Surge protection for Local Area Networks by
MTL Surge Technologies.

That leaves the remaining threat to signal and data ports being voltage differences between the two ends within the building. Say a near lightning strike induces a surge onto an ethernet wire (and the longer the wire, the stronger the induced surge). One end of the data line will nearly always be closer in wiring distance to the building's Earth ground than the other which means a lower voltage potential on the closer side of the ethernet line. Different voltages are bad, so surge current rushes to and through the device with a lower potential.

This could be a single computer or an interconnected stack of switches and other appliances but either way, this damages network controllers and usually other components as the current seeks out a ground. Surges in the power wiring can find a better path to ground through signal/data ports too, and this is one result of insufficient Earth grounding. This happens despite having an AC power protector for that computer or not, and if you unplug these devices from power entirely, they could still shock you from surges on those signal lines. Protection or unplugging must be on both signal AND power, and this is why equal reference to the building's safety ground is so important.

Many brands of plugin SPDs offer models with phone, coaxial and even ethernet inputs but I suggest you avoid them. Too often the (reported) let-through voltages to these lines are way too high for equipment preservation. UL1449 doesn't test signal or data ports so how these let-through voltages were determined is either unknown or ambiguous to the entire device, in which case the UL1449 VPR is the let-through. That will destroy the signal inputs of modems, motherboards, routers and digital tuners. Buying standalone SPDs for signal and data is more expensive and more complex but you actually get equipment protection rather than just human protection.

A Tii Technologies coaxial
GDT surge protector
with grounding block.

A Bourns ethernet
GDT/TBU module
SPD with grounding wire.

When choosing signal and data line protection, you'll be selecting from UL497 listed devices which use gas discharge tubes (GDTs), silicon avalanche diodes (SADs) and/or inductive technologies. Still important are the maximum current capacity from the perspective of service lifetime, and the let-through voltage as there's no VPR for UL497 devices. See pages 24-30 of this Littlefuse reference document for more info on UL497 testing.

Coaxial SPDs for television, cable broadband and ham radio differ mainly by signal frequency. Phone and ethernet protectors come in many configurations for connector shielding, power over ethernet, gigabit speed...when it comes to signal and data protectors, it's best just to call several manufacturers, tell them your application and compare.

UL listed, optimally spec'd ethernet protectors from the industrial & commercial SPD manufacturers range from about $40 to $120 varying from basic CAT5 models to those with all the options. Lower that range by about 25% for phone and coaxial protectors. Bourns, CITEL, Ditek, Emerson, L-Com, MCG Electronics, Novaris, ONEAC, Polyphaser, Tii Technologies, Total Protection Solutions and Transtector are some manufacturers worth considering for signal and data line protectors.


What About Lightning?

Telco towers, rocket launch facilities and high rise buildings handle direct strikes just fine but their protection infrastructure is far beyond what's found residentially. For homeowners and lightning, it all comes down to your Earth grounding but more on that later. The average ground lightning stroke is around 20kA but recorded strokes range anywhere in the area of 5kA to 220kA and higher in extreme record cases.

Along with fault current, lightning is the most abuse your local utility line or antenna could see, but your surge protectors won't ever meet such high energy, only the remaining transient. Here are two quotes by Schneider Electric (the parent company of APC and Square D) from a whitepaper titled Choosing the Right Protection. Peak Ampere Surge Current Ratings of SPDS [p.2-3].


Annex A of IEEE C62.41.1-2002 also provides information demonstrating the physical impossibilities for high surge currents making it to the point of the SPD or into a building. Since the basic insulation level (BIL) of a typical electrical system is approximately 6,000 volts, the high voltages that accompany high surge currents will flash over prior to reaching the SPD.

From these scenarios it is clear that higher-rated surge current (kA) devices offer no real benefit over lower-rated devices when comparing their abilities to shunt large amounts of current. This is due to limiting factors within the distribution system that keep the SPD from experiencing extreme currents. The advantage of the higher-rated device is the ability to handle more transient surge events over time, not necessarily larger individual transient surge events.

Here is another quote from Eaton's Guide to Surge Suppression. [p.11]


It is physically impossible to have the energy associated with a lightning stroke travel down the AC power conductors.

A building's electrical system can't conduct more than 6,000 volts at 3,000 amps. The full energy of a lightning stroke won't enter the building on the AC feeder lines because it will arc from panel busbars to the panel enclosure itself, both of which are connected to Earth ground. While the outcome wouldn't be pretty, if such a high current tries to enter elsewhere (like a hit to your satellite TV dish) that results in heat, melting, maybe fire, explosion and/or arcing to more convenient grounds than air. Such damage arises from a vastly improper Earth grounding (or in the case of most satellite mini-dish installations, zero Earth grounding).

The basic insulation level is one rationale for using the IEEE based C1 impulse wave (6kV, 3kA 8x20) in UL1449 3rd ed. testing as designation for low exposure conditions. But lightning greatly magnifies differences in ground potential. Even two different NEC compliant Earth electrodes, when unbonded, can create 120,000 volts with just a 5kA lightning stroke (Ohm's law). You can get away with an insufficient Earth ground when fielding minor surges but the thunderbolt will be the great equalizer.


When Decisions Are No Fun

This section is for the people who would rather just know a model number, buy and move on with life. However, treat these recommendations like you would default settings—suitable for most uses but may or may not be your exact fit. SPDs at the service entries for power, signal and data are the most cost-effective solution to surge protection. Even if you do nothing more than this, you're still massively improving your situation. If you're unable to install at the service entry because you don't own the building or have very old wiring, then type 3 SPDs (and possibly even normal-mode-only type 3s) are your only hope so make your money count.

I recommend these devices not based on personal use, but by datasheet specifications and having spoken with the companies about the SPDs' internals. Do not equate these recommendations to any kind of partnership or means of sales generation. These are not sponsored listings, these are protectors I would comfortably spend my own money on but it's entirely possible these change in the future. I offer four different AC power service entry SPDs to choose from with all-mode protection and exceptional warranties. To order one of these, you must call the manufacturer directly or search their website and they'll refer you a distributor. These models are also available outside North America.

Service Entry SPD Recommendations

First is the TransTrack TTLP by Total Protection Solutions. The TK-TTLP-1S240-FL costs about $1000 in split phase form and is rated at 50kA per mode (8x20 μs waveform). It's both UL1449 (as a type 1) and UL1283 listed with a 600V L-N VPR but its IEEE C1 impulse wave let-through is 587V between a 120V phase and neutral. It has individually fused MOVs, low pass filtering of -41 dB @ 106 kHz, a steel NEMA 4 weatherproof enclosure and optional Form C contacts for remote monitoring. It's pre-wired with 10 AWG stranded wiring and comes with a flush mount panel, a $20-$30 extra for other protectors. For residential use, the TTLP has a lifetime unlimited free replacement warranty (see p.10) and it's made in Virginia.

Match a service entry protector to your
level of risk for surges.

Moving down the line in price is the PT80 by MCG Electronics. It's UL1449 listed as a type 2 and costs about $700 for split phase with a 40kA max current per mode (8x20), a VPR of 800V L-N and a C1 let-through of 630V. It has individually fused MOVs, EMI filtering ranging from -30 to -60 dB @ 50-230 kHz, a steel NEMA 1 indoors enclosure but has an optional 4x enclosure kit and Form C contacts. The wiring is 10 AWG stranded, the PT80 has a 20 year free replacement warranty and is made in New York.

A hard to beat best value protector for low to medium exposure areas is the Eaton Innovative Technology XT-50 rated at 50kA per mode (8x20). The split phase model costs around $420, each MOV uses thermally sensitive soldering instead of fuses and the XT-50 carries testing certifications from UL (a type 1), CE, CSA and even IEC for vibration testing. It has a 700V L-N VPR, a NEMA 4x polycarbonate weatherproof enclosure and 10 AWG stranded wiring but does not have EMI/RFI filtering nor optional Form C contacts. The free replacement warranty is for 5 years (see p.2) and it's made in North Carolina.

An even less expensive choice for lighter exposure areas but by no means an econodbox model is the Joslyn Surgitron III for about $285. UL1449 listed as a type 1, the split phase model with both common and normal mode protection is the 1265-49-C rated at 40kA per mode (8x20). Its listed VPR is 800V but L-N C1 let-through is only 480V. Again we find individually fused MOVs and a NEMA 1 composite indoors enclosure which is small enough to mount inside your main load center. The Surgitron III has 14 AWG stranded wiring and optional Form C contacts but it doesn't have EMI/RFI filtering. The free replacement warranty is for 3 years and it's made in Virginia.

Point-of-Use SPD Recommendations

A good SPD at the AC service entry will leave a let-through voltage of about 500-800 volts for a type 3 device to deal with. With surges originating inside residential buildings mostly being line noise, major events may or may not be a concern depending on the condition of your appliances and wiring, geographic location, proximity to other buildings and what you have to protect. The least expensive option for type 3s is to stick with MOV-based plugin protectors. They're all largely similar and many I'd not waste money on but a few do stand out to me.

Here's the Ditek DTK-8FF to start. It costs about $25 and each of 8 outlets has its own SAD and MOV array. When any of the 8 MOVs and/or SADs dies, the indication light goes out but the other outlets still conduct power. It has a two-line RJ11 for phone, fax or dial-up internet, not DSL. The Ditek's VPR is 330V on L-N and "around 200V" on the phone line which is fine. It's UL1449 listed, has a 15 amp load capacity, circuit breaker, wiring fault detection, thermal fusing for the device, all-mode protection and noise filtering (though Ditek didn't have the EMI specs available). The case is plastic and it's made in China.

Type 3s with only 1 or 2 outlets
can be expanded with a power strip.

If you have the budget, surge filters are more expensive but offer more stable power to downstream devices. Smart Power Systems stands out because they make several UL1449 and UL1283 listed surge filters which do both common and normal mode protection. They have a circuit breaker, thermal fusing and wiring fault detection but here's the focal point of filter protectors compared to MOVs—the L-N let-through voltage is a mere 0.5V, L-G is 10V.

The lighter use Smart Cord model sells for about $130. It has only one outlet so you'll want to use it with a good power strip, an extra $30 or so. It only supports up to a 10 amp load (there's a 7 amp model too for ~$100) but 10A will still accommodate a 1000W computer power supply, IPS monitor and some extras. Their Computer Guardian is a heaver duty 2-outlet protector with a 15 amp load capacity, phone/fax/DSL and RJ45 (non-gigabit) jacks for about $200. Both are made in Houston, Texas.

For data and signal, save yourself a lot of hassle by calling CITEL, L-Com or any of the other companies mentioned earlier. Tell them what you want to protect, tally up their suggestions and decide.


Earth Ground

I've saved the most important part for last because this is the Achilles' Heel of surge protection. Earth ground is the contact your facility has with the the actual planet (the dirt, the soil) and into the Earth is where electrical current wants to go. Though related, Earth ground is a separate concept from a building's safety ground or the chassis ground of electronics enclosures. With high-end SPDs and code-compliant building wiring, Earth ground becomes the bottleneck. If a surge can't quickly dissipate into the Earth, it looks for other places to leave the building and at that point, you should expect damaged equipment.

So what's an effective earth ground? United States National Electric Code specifies a maximum Earth ground resistance of 25 ohms. That can be achieved in different ways but single point Earthing is the overall goal. A code-compliant ufer ground, ring ground or driven rods are all effective Earth grounding systems. Many buildings also incorporate an underground copper water pipe into Earthing and these can all be combined as shown in the image.

A wind turbine with a combination ufer/ring Earthing system
and a building with a ring/electrode system. Both (in red) are bonded together.

Yet sufficient Earth grounding is just the start because the Earth itself ultimately dictates how effective your installation is. Soil composition, moisture level and temperature, even the location's geology all play a part in the soil resistivity and the capability of your finished product. Don't let that get you down though, because you do have the majority of control over the real-world effectiveness of an Earthing system.

Now, what is not effective Earth grounding? A single ground of only a water pipe was common practice decades ago but is now against NEC code due to the increased use of PVC and PEX in water delivery systems. Rebar used as driven rods will corrode, loosing conductivity and increasing resistance. Shallow ground rings. Multiple unbonded Earth grounds cause voltage potential differences. Decrepit electrodes should be replaced. Basically any Earth ground installation with a resistance above 25 ohms (or even near it if you want to be strict) can be considered an ineffective Earth ground. To change that, the question then becomes: How much do you want to spend?

Determining your building's Earthing configuration could mean a phone call to the contracting company who did the original installation, or may be more complicated for older or communal buildings. Determining the resistivity of your Earth ground or soil is not trivial but not prohibitively difficult or expensive. But in the real world of preexisting structures, you must abide by practicality and cost so while it's against NEC code and best practices to have "insufficient" Earth grounding, that doesn't necessarily mean you're in a dire position. It's all about the finished product in context of your surge exposure level. The bigggest complication for most people will be multiple service entry points resulting in unbonded electrodes.

And Now the Disclaimers

Thus concludes tSc's hardware debut. I think it went rather well, yes? But we all know datasheets and specs don't tell the full story so I admit, without any hands-on time opening up SPDs and testing them (blowing them up), this is a bit of a watered down hardware debut. Eh, that's how it goes sometimes.

I am not an electrician or electrical engineer. I did have auto and home electrical experience prior but 90% of what's above is the result of A LOT of reading and rereading, research, studying and phone calls—all to only scratch the surface. I assume absolutely zero liability and/or responsibility for any and all resulting scenarios of reading this page and/or acting on the information within it. Hopefully this page illustrates that you don't need to invest a lot to protect what's worth exponentially more money, not to mention preventing the stress and hassle of dealing with dead equipment, data loss, insurance claims and other unnecessary consequences.

Share this page.


Resources