https://longnow.org/ideas/very-long-term-backup/
There was a near miss in 2012. These things are not than uncommon, on a century-scale:
https://en.wikipedia.org/wiki/Carrington_Event
We had a miss in 2012:
finally my backup strategy of tens of thousands of pages of dot matrix printing on continuous feed paper has been validated
He had me come in each morning and go over the logs for the servers on our VOIP product - which fed to a line-printer (dot matrix) in real-time and would spit out a ton of paper. (as much as I hated this, I love the sound of dot-matrix printers)
I had to go through the logs with a highlighter and highlight out anything suspicious...
This was the mid-late 90s and we were one of the few companies to have already adoped Linux to production at that time.
Some of our employees were the authors of H.323 and some other protocols - and flourished out to all sorts of your streaming techs today.
(At this point, I've no idea what eventually became of those boxes, though...)
> Today, any information stored only on a floppy disk is essentially gone.
That's totally untrue. You can buy a USB floppy drive on Amazon for $20, and a pack of disks for another $20. I have two 1990s-era Power Macintosh machines with floppy drives sitting on my desk right now, they both work, as do the floppy drives.
And I was able to access some ~15-year-old floppies just a couple of years ago with no trouble.
Now, do I put any faith in the reality of being able access any of it the following week? Not really.
There isn't really a way to replace them.
So.. less than Earth's ambient magnetic field?
While low power, long distance lines (like telegraph or POTS) would be in trouble, and maaaybe some network cards would burn our due to long Ethernet wires, storage devices and many computers would be fine. If you are super worried, seal some hard drives in food tins - the frequency is pretty low so it doesn't even have to be fully airtight.
Am I thinking about this the wrong way? Low frequency will penetrate more easily, high frequency will be blocked. Did you mean high?
Of course I didn't actually look up what frequencies we're talking about- low and high can mean a lot of different things
Apparently the electrical fires were from GIE (https://en.wikipedia.org/wiki/Geomagnetically_induced_curren... ), which are "measured in V/km".. but how big were they?
After some searching I found this PDF presentation: "Geomagnetic Storms and the US Power Grid" https://www.swpc.noaa.gov/sites/default/files/images/u33/fin... , which says: "GIC are Quasi DC currents" and "Moderate Storm - Electric Field at Earth’s Surface 1-5 V/km / Largest recorded 20 V/km" and
Let's assume 10x worse than largest recorded, 200 V / km.. not sure if this is physically possible but probably a worst case.
A high-voltage transmission lines would be in danger, as presentation said. Not only because of overvoltage, but also because this is injecting DC into lines designed for AC only.
A 10 km old-style telephone or telegraph line would get 2000 volt on it... this would definitely burn the equipment or set it on fire.
What about my house? The AC comes in from nearest distribution transformer ("pole pig"). The nominal output is 110V RMS (155V peak), and let's say there is ~0.6 miles (1 km) from it to my house. This will be 355V peak arriving, equivalent to 251V RMS. My appliances and lights (which were turned on) are likely gone... but many computers might survive, as modern power supplies are often rated "90-240V" for european usage (and they don't care about AC/DC mix either). And of course if you had any surge protectors in your extension cords, they'd protect your devices -- they were designed exactly for situations like those.
What about my local network? There are less then 100 meters of distance, so 20V induced voltage.. and common Ethernet is rated for 2000V, so this is going to be fine.
What about USB cords, low-voltage power cords, etc...? Well, those are shielded / have two wires right next to each other, so they would normally be unaffected. But let's say you had a bad ground connector or incompletely plugged in device, so you have a loop. In this case, 2 meter (6ft) cable would have 0.4 volts. That's within tolerances of most ICs, they'd be fine.
What about unplugged device? Your 18" laptop will develop 0.08 volts in the worst case. Most of the modern ICs run in 0-1V range, so they won't burn up (but might have some glitches if device is working). Your antennas (WIFI and 3G) would also develop extremely low voltage so they would be fine as well.
TL/DR: country-wide systems (power grid, old telegraph) is in trouble. City-wide might or might not survive. Anything smaller than a house is fine. If you are worried about second Carrington, get water, food and generator, and get some surge protectors.
Other tests done during electromagnetic compatibility testing (EMC) include bombarding things with radio waves, testing susceptibility to surge voltages and injecting voltages into power cables.
https://emcfastpass.com/emc-testing-beginners-guide/emc-immu...
The key for all of these is duration and total energy. A TVS diode can deal with high voltages - but if they're sustained for too long the diode will burn up. The testing regimes all use quite low total energies.
The values you give for geomagnetically induced currents talk about volts per km. EMC testing of industrial equipment is typically done at volts per meter - a factor of 1000x greater. The industrial standards require 10 V/m, military stuff is tested at 200 V/m.
So all that EMC stuff is very important in everyday life (and it killed more than one FET :) ), but seems inapplicable for Carrington-style event. There, it will be all about huge currents and low voltages. If the wire is long enough, TVSs may burn out, capacitors may overvolt and explode, any ESD protection diodes will also get destroyed. Luckily, we don't have a lot of long-distance copper wires connected to sensitive electronics.
The latest example is 1989 Quebec power blackout: https://en.wikipedia.org/wiki/March_1989_geomagnetic_storm . The storm tripped some protections, that overloaded other circuits which tripped as well, and the whole province went dark. It took 90 seconds for everything to shut down, and 9 hours to restore everything. Judging by that timeline, no major equipment was damaged.
I keep reading about government mandating solar storm protection for power grids.. hopefully next time a big storm comes, we'll be more protected than Quebec in 1989.