This is cool. But now I want a power bank with hot swappable batteries/modules. Or better yet a connector that you can attach multiple power banks to and it gives power if at least one non-empty powerbank is connected.
> But now I want a power bank with hot swappable batteries/modules
I think it’s possible to build one. Get a bunch of 18650 cells from a reputable brand (or just an old laptop battery, if you’re brave enough), then look for a 18650 powerbank kit on AliExpress – some of these would be with slots you can just plug the cells into, without welding/soldering.
I'll elaborate. You'd think that by putting N powerbanks daisy chained together (except i imagine at either end) you get ~N times the mAh or runtime but i think you get probably .25N or less. the "last" powerbank in the chain, the one you'd charge to charge them all, would run out of power first, and about 15% of seconds more afterward, the second one, then the next, with the "first" powerbank, the one you're using for its USB ports to power a load lasting about 2x as long, no matter how many powerbanks you put in a row.
this is a supposition, but i don't think the numbers are very far off. Most powerbanks are 18650 or 26650 inside - flatter or "better" ones are lifepo or lipoly or whatever, instead of cylindrical Li-Ion. anyhow those are 4.2V nominal and USB wants 5V so all single cell or parallel'd 18650/26650 power banks are going to use a boost converter to get the voltage to 5.1V nominal at the USB terminal (assuming a dumb power bank, a power bank with power delivery will boost that even higher but also probably has multiple batteries in series, but that doesn't matter, it makes efficiency worse for our daisy chain regardless!). So these boost circuits "charge" something and store it until there's the correct number of electrons to equal whatever charge/joules is required to run the load. There is >15% wasted as heat either in the capacitor or the inductor (depending on the style of boost converter). There's another ~15% or so lost in the charging circuit, as it has to take 5.0VDC in and run a charging circuit (similar to mppt) on the battery at different voltages and/or amperages which again waste is generated as heat. Cable losses in the daisy chain probably account for a percent or two each, the indicator lights across all the power banks, plus all the microprocessors/etc inside of them probably waste another couple percent.
You'd lose a lot in efficiency. A power bank designed for this would have 2 cables and 1 or 2 wires, power transfer cables and sense/control wires. My engineering brain says "don't do this it's stupid and I think OP was making a joke I don't get"
There are a good amount of lower power ones (e.g. with an N100 CPU) that draw ~15W usually and not that much more at full bore, and some of them are starting to come with USB-PD power inputs (even if they come with a DC power adapter some will accept USB-PD on another port).
i have a passively cooled Quieter 4C [1] with N100 and a 4TB [dram-less] nvme [2] and it draws 4.5W total at idle running EndeavourOS w/ KDE Plasma. at full load (Handbrake 1440p transcode) it draws ~10W.
i left it at stock bios settings and did not put it into higher TDP mode though, since i use it as an htpc and it mostly idles, i'd rather it stay at more comfortable operating temps.
the nvme cost me more than the pc, which for $207 includes 16GB ram, 512GB nvme, and a Win11 Pro license. insane.
i'm mostly converting 4k from a phone to 1440p and cutting [backup] storage by 6x. most of these vids are < 3min, and i can do quite a few in a batch job overnight. im in no rush. it does about 4fps with the 1080p30 HQ preset with the only additional change to up the res to 1440p. i would not use it for batching feature length films, but you can do one in a pinch :D
I’d much rather do that than purchase a NUC with a built-in battery. Keeping the two pieces separate makes them easier to repurpose later.
Regardless, the market for such a device is relatively small.