[schiffern]

Commercial uses layered surge protectors (Type I, II, and III), which is also recommended for other users but rarely followed.

In surge prone areas, at a minimum I would have good quality whole-house surge protector (eg Siemens 140 or Eaton 108), and a good quality surge protector strip for any computer/TV/phone charger.

I also put surge protectors in front of expensive white goods like the fridge, washer/dryer, dishwasher, and garage door opener. Besides being costly to replace these can contain "sparky" motors and this provides protection in the other direction too. Over time smaller surges can degrade the main surge protector for your computer.

Nothing (reasonable) can protect against direct lightning strikes, but for anything less it should provide decent protection.

[tempaccount5050]

Are you in an area with a bad electrical grid or something? In 40+ years I've never had a single device get fried from a surge/storm. My "surge protector" power strips are from the 90s and probably don't even work.

[tharkun__]

This. Same timeframe and I've lived through both lots of lightning storms and in areas with lots of power failures. Some of them intermittent and essentially caused by transformers blowing up. Like earlier this winter, we had multiple storms where you'd hear a transformer blow up, in many cases even seeing the sky light up as well from it, power going out, couple seconds, power coming back, next transformer blowing out, rinse, repeat.

On the other hand I've read about plenty of stories of the "cheap" UPSs you'd usually buy as a consumer (not to name any brands coz I've never had any) actually causing such issues in the first place. Without any actual surges from the grid.

That said, being totally not superstitious (for real, but someone's gonna "kill me" if they find out I wrote this and something dies from a surge...), now I guess I need to knock on wood like seventeen times ...

I do use surge protectors when we're on generator power temporarily.

[ssl-3]

The things people often call "transformers blowing up" are usually not transformers blowing up.

Instead, it's usually just overhead wires that are too close or literally touching, often from influences like wind and ice. The electricity arcs between the wires, creating bright blue-white flashes that can be seen from far away, sometimes with instantaneous heat that makes hunks of metal wire evaporate explosively. It can be violent and loud, and repetitious as different parts of even a single run fail.

Transformers can certainly blow up, but that's less common. They're (generally) filled with oil for cooling purposes, and they're massive things that tend to take time to get hot. A failed transformer can produce arcing and blue-white light, but if things are that hot then the oil is also ready to burn.

And when the oil burns it isn't blue-white -- it burns with about the same yellow-orange color we saw the last time we accidentally flambéed dinner on the kitchen stove, or a Hollywood fireball.

A bright flash without a fire is probably not a transformer.

Here's a video of a transformer actually-exploding (note the prominent fireball): https://www.youtube.com/watch?v=oFkfd31Wpng

And here's a video of what someone describes as a transformer exploding, even though there are no transformers in the video: https://www.youtube.com/watch?v=rHVh0KwG_0k

[tharkun__]

Haha, I hear you. But yes, it really is transformers blowing up sometimes. Sometimes it really is just branches blowing up the line, sure.

A branch hitting a wire, happenes all the time here too. Lots of trees in this community. The video of a transformer you shared: that's not the transformer I'm talking about. That's at a transformer station.

I'm talking transformer on a street pole. The kind that hangs right across the street from me. This kind: https://www.youtube.com/shorts/y3E7avUvj6I

See it's the kind in your second video. It's a transformer. You just chose a narrower definition I suppose. It's a https://en.wikipedia.org/wiki/Distribution_transformer ;)

And yes, I know it's transformers and not just wires (but also wires do happen definitely) coz I do walk the neighborhood regularly and I can tell when a transformer is new vs. old up there. Ours is old. The ones a few streets over sometimes are very new and I see the Hydro trucks go by the next day(s) to make them new ;)

Again, like seventeen times knock on wood but the ones next to us have not actually blown up. But three streets over, seen the new ones. Literally last weekend, we had an ice storm come through and while no blowouts we could see or hear, the outage map showed plenty of failure.

[freeone3000]

Residental-scale transformers can and do explode. Shorts happen not-infrequently with freezing rain and ice storms especially causing issues - the internal oil gets displaced by the water, and the dirty water causes an internal short. It wipes out power to a few blocks here when it happens, but we get an outage due to it every year or two.

[ssl-3]

They can. They do.

But when the wind is whipping along on a warm day and there are bright flashes and audible bangs, that's (usually!) not signs of transformers blowing up... even though the popular vernacular often erroneously describes it that way.

[LorenPechtel]

It happens. The power company was very unhappy with my boss for destroying one of their transformers. The thing is while circuit breakers react very quickly to extreme overcurrent situations (shorts) they're much slower to react to loads which are only a bit over the limit, and if short enough won't react at all. Very common with heavy motors.

And that's exactly what the problem was--we had a whole bunch of really heavy motors. Getting ready to start for the day you flip on the switches and the big machines start to spin. The transformer on the pole was rated higher than the main breaker for the plant--but the transformer apparently was more sensitive to the temporary loads. Once the problem was identified it was resolved by staging it, instead of flipping them all on they were flipped on over 5 minutes.

[pseudohadamard]

It's not just cheap UPSes, it's cheap surge protectors as well. They exist because the vendor can throw in a MOV costing a few cents and increase the price of the power strip by 50%, not because they're any good. MOVs are sacrificial components which have either degraded to uselessness by the time they're actually needed or, if they're still working, can explode or catch fire from the energy dissipated. Even if they don't, all they're doing is converting an x-kV spike on active into an around-x-kV spike on neutral or ground. If you want to do it properly, use a series tracking filter, not a "surge protector".

[tharkun__]

No offense, but can you tell me how my 4.5 kW generator is gonna generate that kind of power surge?

[tlb]

One scenario: there's a short circuit somewhere, say rats chewing through insulation. This can cause a very high current through the short. A non-inverter 4500 watt 120 volt generator might have 0.2 ohms coil resistance, so the short circuit current can hit 170 volts / 0.2 ohm = 850 amps. When the shorted branch's circuit breaker trips, the inductance in the generating windings wants to keep that 850 amps flowing for at least a few microseconds, and it gets distributed across everything else that's still connected. Depending on what else is connected (hopefully including some surge protectors) the peak voltage can get into many kilovolts.

The circuit is something like this:

  voltage source -- parasitic inductor --+- circuit breaker -- short
                                         |
                                         +- circuit breaker -- your PC

[pseudohadamard]

More generally, for the previous poster, look at what happens when a magnetic field collapses suddenly, you can get kilovolt spikes. There's probably a ton of YouTube videos demonstrating this in various ways, it sounds like the sort of thing that Electroboom would do. Normally this is handled via snubber circuits which dissipate the energy before it can do anything, but in exceptional cases it could end up going where it shouldn't.

[burnt-resistor]

Definitely use quality surge protectors on expensive equipment connected to generators.

PSA: UPSes and GFCI/GFI extension cords won't work properly when connected to a stand-alone generator with a bonded neutral. I've tried using enterprise UPSes on such generators, but they absolutely won't work. In such scenarios, debond the generator's ground from neutral, apply a very large warning label to it being debonded, and drive a massive ground rod electrode into the ground as close to the generator as possible and ground the neutral there. This does work and is much safer because there's a stable voltage reference source. It's more of a hassle but can be necessary for some off grid and temporary scenarios.

[ssl-3]

GFCI works correctly either way. Their operating mode doesn't care at all about ground: Whether bonded, not bonded, or not even present (look, ma! only two wires!), they still perform the same way.

They respond to an imbalance in current flow betwixt line and neutral. What goes out must return; if it doesn't, then switch off.

Ground is not part of the equation at all.

[jcalvinowens]

That extra unbonded ground rod is the worst thing you can possibly do to make your generator vulnerable to lightning strikes.

[burnt-resistor]

That's an extreme edge-case and a strawman. Anyone operating temporary equipment on a generator during a severe storm will obviously unplug sensitive stuff to not take unnecessary chances regardless of safety precautions already in place.

Ground rods are required in certain situations according to the NEC.

Ground rods are for lightning protection, transient surges (over voltage), and induced surges; not for short protection, ground faults, or making ordinary extension cord use of bonded generators "safer".

Typically, they're required whenever it's a system that powers a building on its own, i.e., off-grid setup or with a floating neutral generator connected via a switched neutral transfer switch.

[jcalvinowens]

You can unplug everything and open all the switches, but a nearby lightning strike will still fry your generator through that unbounded ground rod. Lightning ground potential is very eager to take the shortcut to your other ground rods through a few millimeters of insulation and open switches on the path through your generator and house wiring, when the alternative might be tens of meters of dirt :)

I don't care what the NEC doesn't say, NFPA 780 says you have to bond all ground rods.

[mschuster91]

> In such scenarios, debond the generator's ground from neutral

eeeeep. Please for the love of all that is holy, CONTACT AN ELECTRICIAN before messing around with that - or before creating a ground bond where none should be (i.e. TT grid [1]). You may end up endangering yourself if you do not exactly know what you are doing - in the case of TT, you get ground potential difference current from other parts of the grid flowing to ground via your generator's bond. Best case you're getting problems with electrochemical corrosion (including in your foundation), worst case enough current flows to turn your bond wire into a thermal fuse.

Also, take great care if your grounding is provided via municipal water service, or if your original grounding rod has dried out to the point it's ineffective.

Let me repeat: LET ELECTRICIANS DEAL WITH GROUNDING AND SURGE PROTECTION. Floating grounds and improper ground connections CAN BE LETHAL OR POSE A SERIOUS FIRE RISK.

AND YES THAT INCLUDES "ISLAND" SCENARIOS OR EMERGENCY POWER INPUTS (e.g. via CEE plugs and transfer switches).

[1] https://de.wikipedia.org/wiki/TT-System

[dboreham]

I'm not sure I'd leave something like this to an electrician. Or if so at least make that electrician be experienced in this field. I think you'd want an electrical engineer to be involved with the plan to some degree.

[Kirby64]

Electrical engineers don’t know code requirements and wiring guidelines for household electrical wiring. They’re absolutely not the correct default. Electricians with specialization in generator setups, sure, but an electrician engineer on average is likely going to be more uninformed on code requirements than an electrician.

[mschuster91]

Electrical engineers know the theory but lack the practical knowledge which grid form is used at your specific address (yes, here in Germany we have a few towns where one half side of a street runs TT and the other one is already migrated to TN-C or TN-C-S).

An electrician specializing in lightning protection, uninterruptible power installation or in radio installations can sort out all of that far better than an engineer can.

[noahbp]

When I lived in Costa Rica, I lost three surge protectors in a year to power surges. During one such power surge, I didn't notice that the red light indicating surge protection was already out, and a power surge fried my (knockoff) Macbook power adapter, leaving me without a way to work for a day.

[LorenPechtel]

There's no real indicator on those things.

It's a 10 cent component that would require a $10k machine (this is old data) to do a non-destructive test.

[DiabloD3]

Not only should you get rid of them, but also they are a fire hazard.

Also, do not accidentally plug surge protectors into each other, metal oxide varistors can star fires _without_ meaningful surge conditions when you do so.

I prefer to buy products without MOVs entirely due to the risk, with the exception of one, Tripp Lite Isobars; but I prefer to use series mode protectors such as Brickwall or SurgeX.

[ssl-3]

> Not only should you get rid of them, but also they are a fire hazard.

Are they not a fire hazard even when new? MOVs do tend to degrade with use (especially after they've gone conductive to snuff one or more surges). But AFAICT we can't really know, without potentially-destructive testing, whether a given MOV is in good shape -- whether installed last week, last year, or 30 years ago.

> Also, do not accidentally plug surge protectors into each other, metal oxide varistors can star fires _without_ meaningful surge conditions when you do so.

What is the mechanism that increases risk for MOV-sourced fires in this arrangement?

I've also noticed that many of the power supplies I've taken apart (for very pedestrian consumer goods) have internal MOVs on their line input. Whatever the mechanism is that increases risk, isn't using one external surge protector already doing that in these instances?

> I prefer to buy products without MOVs entirely due to the risk, with the exception of one, Tripp Lite Isobars; but I prefer to use series mode protectors such as Brickwall or SurgeX.

I prefer to avoid MOVs, too. Broadly-speaking, diodes seem like a better way to do it. (Transtector is another reputable brand that uses diodes.)

---

That all said, I've noticed over the years that problems with dead (presumed-to-be-hit-by-a-power-surge) electronics tend to follow particular structures. And the reason for this seems related to grounding more than it is anything else.

So when I find someone (a friend, a client, maybe someone online that I'm trying to help) complaining about repeated damage, I often ask about grounding. Almost always, it turns out that they've got multiple grounding points for the electronics: The electric service has one ground rod, and the telephone/cable feet/satellite/whatever is connected to some other ground.

This might be a dedicated rod, maybe a metal pipe; whatever it is, it is distinct from the main service ground. It happens all the time. (It is worth noting that the NEC prohibits this kind of configuration unless extraordinary effort is put forth. See 800.100(d), for example.)

The way that MOVs -- and avalanche diodes alike -- behave combines with the fact that the earth is an imperfect conductor, such that having multiple ground points promotes dynamic ground loops that can provide quite large potential -through- the electronics that we seek to protect.

The problem appears suddenly, and repetitiously. Everything is fine, and then ZANG: The cable modem gets smoked along with the router it is connected to. So the modem goes back to Spectrum or wherever to get swapped, and the router gets replaced again, until the next time: ZANG.

TV connected to satellite receiver, with coax incorrectly grounded? ZANG. Over and over again.

I'd see it all the time when I was a kid back in the BBS days: The phone line was grounded improperly, and computer was the only thing that connected to both electricity and the telephone line. Some folks would go through several modems over the course of a summer, which was very expensive -- while most people had no problems at all. Next-door neighbors would have completely different failure rates.

Structures with correct grounding tend to do very well at avoiding these issues, and I've fixed these conditions in subsequent years more times than I can count.

(A coworker installed a phone system at a business once, wherein he made extensive use of Ditek surge suppressors -- on the incoming POTS lines, and on the power inputs. It blew up one day. So he called Ditek to try to get at least the cost of the phone system hardware covered. They asked him to draw up a map of how the building was grounded and send that over, so that's exactly what he did. When they saw his map, they very quickly identified a ground loop and denied the claim.)

[matthiasl]

"What is the mechanism that increases risk for MOV-sourced fires in this arrangement?"

I wondered the same thing, and failed to find a satisfying explanation.

I can find plenty of reports of MOV fires, especially in situations where there's a persistent over-voltage, e.g. a 120 V site actually having closer to 240 V due to a floating neutral. But I don't see how chained MOVs make that worse in general. This blog post has some nice photos:

https://www.electrical-forensics.com/SurgeSuppressors/SurgeS...

[consp]

No clue about the actual reliability of this[1] article but the mechanism mentioned (new pathways due to changes in crystalline structure due to uneven heating) sounds possible.

1. https://incompliancemag.com/how-and-why-varistor-failure-occ...

Reread your wondering and now conclude its about chained situations which this also does not answer.

[schiffern]

Not too bad, just rural. We used to lose stuff every 10 years or so.

One day The Big One came along and fried nearly everything. "Once burned, twice shy."

Hopefully someone can learn from my mistake and not have to do it post-mortem.

[MrDresden]

Heavy industry can also cause these kinds of power surges to happen.

Last year an aluminum smelter in Iceland had a transformer blow which caused a big power surge on parts of the very well developed national power grid. The surge caused damage to electronics in some households and companies near to the smelter.

[SmirkingRevenge]

I had exactly one device fried in my life. This was when I lived in FL in 1999. Lightning took out my sweet new 56k modem. My PC was fine.

[dboreham]

EE living in a rural location here: transient related failures do happen in my experience. Rare but they happen. And I've known of people who had everything in their house fried. For me it's just been a couple of Ethernet ports. Power strips don't provide much protection fwiw. Always worthwhile checking that your electrical service is properly grounded.

[acjohnson55]

I lost an audio mixer to a bad surge last year. I don't know whether it was additional load or just really bad fluctuations that damaged the device. Nothing else bit the dust, but the, digital board in this mixer got bricked.

[lucb1e]

How did you know it was a power surge? Not doubting your comment, just interested in knowing if this ever happened to me

[acjohnson55]

Just saw this days later - I'm not sure if it was a surge specifically, but it got bricked after a really wacky power swing. So maybe not a surge, exactly.

[radley]

I live in Oakland, and the Easy Bay has had its share of random outages without incident. The last outage was just a split-second blip, but it broke a $1k computer monitor.

[JanisErdmanis]

If your close neighbours have surge protectors then you benefit little from installing your own.

[sgc]

Another perspective: we should install whole house surge protectors if we can afford them, not only for ourselves, but to help our neighbors - even if in reality the help is minimal and they need their own as well. In the best case scenario, if everybody in a neighborhood has them, each individual house will be more resistant to surges than if they were the only house with one (five houses with surge protectors nearby is a lot better than one) - everybody wins.

[lucb1e]

Why? If the voltage spikes on the grid (that's what I understand a power surge to mean), wouldn't even more of it end up in your house (that is: the grid voltage spike even higher) if the neighbors have equipment that doesn't let their devices consume some of that energy?

Edit: wait, maybe I figured it out: those devices must be consuming the excess rather than blocking it. Is that it?

[JanisErdmanis]

Yes, energy dissipates. Although one still needs to look out for the distance from the neighbours as your ground can be different from your neighbours ground potential.

[dotancohen]

You might as well phrase that as "If your close neighbours have gotten vaccines then you benefit little from getting your own."

We live in a society. Everybody chips in. And each surge protector adds to the robustness of the grid.

[JanisErdmanis]

Yes, that puts it down perfectly. That’s why some don’t ever see the benefit of installing their surge protector whereas others install one way too small for their situation and find them useless anyway.

[margalabargala]

Eh. Most nice power strips are also surge protectors.

[bigfatkitten]

You don’t need a storm. A fault like an open neutral can very easily fry things.

[burnt-resistor]

Not completely correct, nuanced, or comprehensive.

Direct lighting strikes cannot be defended against without extreme costs. This type of risk is generally extremely unlikely except for certain niche use-cases like equipment or facilities on tall peaks.

Transients from lightning (E2) nearby and distant nuclear detonations can be defended against, and often require additional protection of telco and internet entry points. Whole house type 1 SPD devices exist for residential applications. This is much more likely than direct lightning strikes, especially in certain areas and can be defended against for reasonable cost. The main issue of lacking it is the unseen, cumulative degradation of semiconductor components that lead to instantaneous or eventual failure, especially in high value devices like electrically-communicated motors in HVAC systems. There is no reasonable expectation of defense against a direct lightning strike even with type 1 SPD, and there are different types of lightning with vastly different amounts of energy. A positive strike direct hit will totally fry anything and everything.

What generally isn't defended against at all in any infrastructure or system except some military equipment is H/NEMP E1 (short duration impulses) or E3 (E3a or E3b; long duration surges larger than lightning) such as from unusual space weather events or nuclear blasts.

[david_allison]

> Nothing (reasonable) can protect against direct lightning strikes

Belkin make a number of surge protectors which offer a connected equipment warranty in the UK. Admittedly: financial protection, not data protection, but I felt it was worthwhile for the peace of mind.

https://www.belkin.com/id/p/6-outlet-surge-protection-strip-...

[pseudohadamard]

Have they ever paid out on one of those, or is it like CAs who offer liability protection for their certificates carefully set up in such a way that they never have to pay out.

[gruez]

>Admittedly: financial protection, not data protection

You should have data backups regardless, because there are plenty of ways to lose data that don't involve power surges.

[gruez]

>In surge prone areas

What areas are surge prone?

[toss1]

Areas with lots of thunderstorms. Also more rural areas with long power lines with few taps off for customers — the long runs are both exposed to many nearby strikes and accept induction well, and the few customers are fewer power sinks to dissipate the spike. So, you're more likely to get hit, and hit harder.

[gruez]

Sounds like if you're in an urban area with buried lines, you don't have to worry?

[schiffern]

In urban areas you probably can just have the whole-house surge protector and skip the rest, since that protects all costly electronics not just a single device. With just a surge strip on the PC I'd say you're a tad under-protected, yeah.

Incidentally whole-house surge protection is now required by code in new houses. Existing buildings aren't required to upgrade, but by my reasoning what's good for the goose is good for the gander.

[marcosdumay]

I would recommend a circuit surge protector in urban areas.

Lightning getting through some structure and hitting the electric lines happens. Even when they are buried. It's less of a problem when the ground absorbs a lot of the power before it even get into copper, but it's even less of a problem if there's some cheap device that will burn and protect you from it.

[toss1]

No, just worry less about big lightning, but urban areas have many more sources of different types of spikes and line noise

[hedora]

The California bay area, at least all the sides of it I’ve lived on. We currently have a whole house battery, whole house surge protector, a second surge protector, and a UPS between the router/nas/etc and PG&E.

It’s not good enough. At least the power stays on once the grid stops bouncing (or once I manage to log into the rebooting battery gateway computer to have it flip the “off grid” breaker, or go outside and flip the manual one by the meter).

[jcalvinowens]

I've lived all over the peninsula, your experience is not normal.

I had far more power outages during my late teenage years in suburban Dallas than I've ever had in the bay. That was due to a bad transformer in the neighborhood which took years to replace, but once it was replaced everything was perfect. The moral of the story being: if your power is bad, it's probably because some piece of mains infrastructure near you is broken.

I had a string of annoying outages in 2023-2024, but it was all due to main service upgrades on my street, can't really complain about that.

[kelnos]

Huh, interesting. I've lived in the bay area for 22 years (first 6 of them in various places from SJ up the peninsula, remainder in SF), and I've never experienced damage to anything due to a power surge.

Not saying you're lying, but I do wonder if your experience is typical.

[nradov]

Color me skeptical. I've lived in several different Bay Area cities for decades. There are usually a couple power outages per year but I've never experienced a surge strong enough to cause equipment damage.

[frrlpp]

Open aerial wiring can shortcircuit two phases, bringing a low impedance surge that can damage most electric and electronic equipment.

[andylynch]

> Nothing (reasonable) can protect against direct lightning strikes

This is one reason why you bury power cables.

[Kirby64]

Burying power cables doesn’t help against direct lightning strikes. If lightning hits your house, underground power cables will do nothing.