Saw a teardown/repair of a big home flatscreen TV and while the label on the outside said “120VAC”, the power supply inside said 120/240V, so it was the same internals worldwide, but labelled differently to discourage gray-market imports.
The technical side should continue to work until you reach the lower limit of what the bootstrap circuitry can handle.
The problem will be the amperages required... say, a fridge. Its motor will run at ~200 watts, which means 1A at 230V. No biggie. But if you're going down to 48V, the high-voltage side will now run 4A, and more the lower you go on voltage - and that's without taking peak inrush currents into account which can easily go to 10A@230V (or 40A@48V) for a few split seconds. If your current source or the wiring can't hold that, you'll likely run into serious issues.
https://library.e.abb.com/public/71d5526a71b34360bb3a258c950...
But now that I think about it, it’s possible they’re fitted with different replaceable fuses (half the amperage in 240v areas).
There are two different problems:
1) What conditions equipment works under
2) What conditions equipment went through testing under, as per regulatory regimes
Equipment I worked on needed to be tested for a variety of failure modes (e.g. a power supply doesn't cause a house fire), EMI/EMC, and so on. I would guess that also varies internationally, as well as by use (UL-listing is not legally required, but it is required if you sell to customers who want to have fire insurance, for example).
Some of this is wonky administrative overhead, but some of it is not. For example:
- A DC-DC converter will have a different switching frequency if running from a different input voltage. This might work fine, but it will have a very different profile for EMI/EMC. Testing that might be $40k. If I know I have a 12V wall wart powering it, there's no reason to spend $40k to be able to label the box "5V-40V."
- If a PCB has creepage / clearance for 120V, even if the power supply works fine for 240V, I can't sell it as 120V equipment. The reverse holds true the other way for wire gauge.
These things wouldn't cause problems very often (indeed, it'd be exceptionally rare), but if we have a house fire 0.001% of the time people bought a TV set, we'd have hundreds of thousands of preventable house fires. That's why there's testing before products are sold.
Potentially it could work with a DC input though (since DC will pass a regular bridge rectifier), but it could screw with things like the power factor correction… I’d try it if it was a unit I didn’t mind potentially destroying.
If they only assumed AC, and if they used AC coupling for the sine/peak/phase detection, then DC would probably look closer to 0V, and just shut the switching off, right?
I can't think of an "easy" way to do it that wouldn't require extra smarts, to handle DC, and everyone knows smarts costs money!
This person has a thesis that it can go up to 20 volts (which is 80% of the rating of one of the capacitors)
It’s just a smoothening capacitor, so higher voltage rating is the critical part and can probably get away with non-exact muF capacity.
Possibly removable if you have a lead acid or lithium-ion battery capable of providing lots of current when needed and cleanly.