Whatever takes the grid down, your house fails in a predictable sequence. Preparing for that sequence covers the exotic case for free.
What breaks first when the grid goes down →EMP protection is two different subjects wearing the same name. One is a genuine engineering discipline with military standards, measurable shielding in decibels, and a body of test data. The other is a marketing niche that sells fabric bags, panel boxes and PDF guides on the strength of a scary video. They overlap much less than you would think, and the second one has trained a lot of people to prepare for the wrong things.
What follows is the physics first, the evidence second, and the product claims last — including the ones that do not survive a close look.
Three pulses, three completely different problems
A high-altitude nuclear detonation produces three effects that get lumped together as “EMP” and behave nothing alike. A severe solar storm produces only the third.
| Pulse | Timescale | What it couples into | What helps |
|---|---|---|---|
| E1 | Nanoseconds; under 1 microsecond total | Short conductors, circuit traces, cables | Shielding, disconnection |
| E2 | Under 10 milliseconds | Similar to a nearby lightning strike | Ordinary surge protection |
| E3 | Seconds to minutes | Very long lines, transformer windings | Grid-side hardware only |
E1 is the one the marketing is about. CISA’s planning guidance uses a worst-case peak field on the order of 50 kilovolts per meter, with energy spread from about 1 MHz into the hundreds of MHz. It induces voltage on conductors — roughly 20 kV inside an unshielded building, around 300 kV on external power lines, by those planning figures. The key word is induces. The pulse does not reach into a device and cook it; it puts a voltage on whatever wire the device is attached to.
E3 is the slow geomagnetic surge, sub-hertz, lasting up to hundreds of seconds. It is the one that pushes DC through transmission lines and saturates the big transformers at substations. There is nothing you can buy for your house that addresses E3, because the vulnerable equipment is not in your house.
The part that is not hypothetical
On March 13, 1989, a geomagnetic storm collapsed the Hydro-Québec grid in about 90 seconds. Six million people lost power for roughly nine hours, in Quebec, in March. The same storm damaged a large transformer at a nuclear plant in New Jersey, well over a thousand kilometers from the Quebec grid.
That event is the honest center of this whole subject. It happened, it is documented, it is repeatable, and it argues for exactly the preparations that also cover ice storms and hurricanes: water, heat, food, and a way to make power for a week or two. It says nothing whatsoever about needing a bag for your phone.
What the vehicle testing actually found
“Every car in America stops” is the most repeated claim in the genre, and it is the one with the most direct evidence against it. The EMP Commission put 37 cars (model years 1986 to 2002) and 18 trucks (1991 to 2003) into an EMP simulator and ran them up in field strength.
- Three of the 37 cars stalled, at around 30 kV/m or above. One car needed a dashboard electronics repair.
- Twenty-five cars showed effects the Commission classified as nuisance only — blinking lights and similar.
- Eight cars showed no anomalous response at all.
- Of 18 running trucks, three motors stopped; two restarted immediately and one needed towing. Five showed nothing up to about 50 kV/m.
Worth stating the caveat the Commission’s own critics raise: testing generally stopped at the first sign of upset, because the vehicles were borrowed and had to go back intact. So these are floors, not ceilings. But the picture is nothing like “every vehicle is dead” — it is closer to “some vehicles are briefly annoying and a small minority need a mechanic.” That should recalibrate everything else you read on the subject.
Products that do not do what they promise
Panel-mounted “EMP shields.” These are surge protective devices — fast clamping components wired across your service entrance. As surge protectors against lightning and utility transients they are a legitimate product and worth having for that reason. As advertised protection against E1, the claim does not hold up: a device at the panel cannot intercept energy that couples directly into the branch wiring inside your walls on the far side of it. When you go looking for the supporting evidence, what you find is vendor-commissioned summaries, not third-party MIL-STD-188-125 test reports you can read. Buy one as a whole-house surge protector at whole-house surge protector expectations, and price it against other surge protectors.
Untested “Faraday fabric.” RF shielding fabric is a real material with real numbers. Reputable makers such as Mission Darkness publish shielding-effectiveness curves in dB across a frequency range. A great many listings publish nothing. Ask for the chart. A bag with no attenuation data is a bag, and silver-colored nylon is not evidence.
The microwave-oven trick. A microwave door screen is engineered to block one narrow band around 2.45 GHz. It leaks badly at the lower frequencies where E1 carries most of its energy, and there is a power cord passing through the shell, which defeats the concept entirely.
“EMP-hardened” portable power stations. Nobody publishes a shielding spec for these, and the enclosures are plastic. If you want a power station protected, the answer is a spare in a closed metal container, not a sticker.
The video sales letter genre. If a pitch opens with a countdown, an anonymous insider, and a claim that ninety percent of Americans will die, you are not being sold protection. You are being sold a PDF. The tell is that the physics is never specific: no frequencies, no field strengths, no decibels, no test standard.
What is plausible, in order of value per dollar
Everything below is worth doing on its own merits, which is the point — none of it is wasted if the exotic scenario never arrives.
- Two weeks of water, food, heat and sanitation. This is the entire ballgame. A grid event of any cause is a long outage, and long outages are survivable with supplies and miserable without them.
- A generator plus a manual connection method. An interlock kit has no logic board to fail and no utility sensing wires running out of the house.
- A small box of spares in a taped metal container. A spare AVR, an ignition module, a cheap radio, a multimeter, a spare CO alarm. Thirty dollars of shielding around two hundred dollars of parts. The generator version of this is the practical form of the whole Faraday cage idea.
- Unplugging things during a warning. A geomagnetic storm comes with hours of notice from NOAA’s Space Weather Prediction Center. A device unplugged from a 200-foot house circuit is dramatically less exposed than one plugged in.
- Ordinary surge protection at the panel and at sensitive loads. Cheap, effective against E2 and lightning, and useful roughly every summer.
Where I come out
The threat is real enough that Congress funded a commission and CISA publishes guidelines. It is also not the threat the ads describe. If you spend $500 on shielding products and skip the water storage, you have made your household less resilient, not more — because the thing that will actually happen to you is a nine-day outage after a storm, and no bag helps with that. Prepare for the long outage. Spend the last thirty dollars on a taped ammo can of spares if it makes you sleep better. Then go do something else.
Amazon affiliate links follow. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you. Details
AM/FM/shortwave emergency radio on Amazon →
Aluminium foil tape, 2 in x 65 ft, on Amazon →
As an Amazon Associate I earn from qualifying purchases.
Frequently Asked Questions
Would my car still run after an EMP?
Based on the EMP Commission’s simulator work, most likely yes, possibly with a restart. Of 37 cars tested, three stalled and one needed a repair. Testing was cut short at first upset, so treat that as a lower bound rather than a guarantee.
Does a solar storm do the same thing as a nuclear EMP?
No. A solar storm produces the E3 effect only — the slow geomagnetic surge that stresses transmission lines and large transformers. It does not produce the fast E1 pulse. Your electronics are not the target; the grid is.
Do I need to protect solar panels?
Panels themselves are large-area semiconductors and fairly robust; the exposed parts are the charge controller and inverter at the end of long DC runs. A spare charge controller in a metal container is a more sensible hedge than anything you can do to the array.
How long would the grid really be down?
Nobody honestly knows, and estimates range from days to years depending on how many large transformers are damaged and how badly. The 1989 Quebec event was nine hours. Planning for two weeks covers the overwhelming majority of realistic outcomes; planning for a year is a different lifestyle, not a purchase.