The MK3S+ has a claim no newer printer can make: it has been metered by more people, for longer, than anything else on this site. When a figure of 80 W appears for this machine it is not extrapolated from bed area — it is the settled result of a decade of owners plugging watt meters into a design that barely changed.
That makes it the right machine on which to look at what electricity actually contributes somewhere it is expensive.
The same print, at the UK rate
A 130 g PLA part, 8 hours, $22/kg filament, priced at 34 cents/kWh rather than the US average:
- Filament: $2.86
- Electricity: $0.22, from 0.64 kWh
- Machine time at 25 cents an hour: $2.00
- A 4% failure allowance: $0.21
- Total: $5.29
Doubling the electricity price against the US average moved the total by about a tenth of a dollar. That is the honest scale of the electricity term on a small open-frame machine printing PLA, and it is why this site treats the rate as an input worth getting right rather than as a headline. Change the region in the cost calculator and watch how little moves; then change the machine rate and watch how much does.
The 25-cent machine rate above reflects a printer bought secondhand and expected to run for years. On a new flagship it would be double that or more, which is the real reason two people printing the same file quote different prices.
What a design this old is still good at
Serviceability. Every part is documented, printable or purchasable, and there are step-by-step repair guides for failures that on newer machines end with a support ticket. A printer you can fix at midnight has a different effective lifetime from one you cannot.
Predictability. The community has already found every quirk of this machine. Whatever is going wrong with yours has a thread with photographs.
A direct-drive extruder with a gearbox. TPU and TPU 85A work here without heroics, which is more than can be said for the bowden machines of the same era.
What it gives up to anything modern
Speed, mostly, and the reason is worth understanding rather than accepting. The limit is not the motion system's rigidity but melt: an older hotend can only turn so many cubic millimetres of solid filament into liquid per second, and past that point the extruder skips. Newer machines print faster largely because their melt zones are longer, not because their frames are stiffer.
The practical consequence is that this machine will not benefit much from the aggressive profiles shipped with newer slicers. Pushed too hard, it produces under-extrusion and filament grinding rather than fast parts. Print it at the speeds it was designed for and it produces very good parts indefinitely.
It also has no enclosure, so ABS and ASA remain a project rather than a feature, and its 8-bit heritage means firmware capabilities stop well short of what a modern board offers.
Where it sits now
Superseded twice over by later revisions in the same family, and no longer the machine to buy new. Its position today is as the strongest used purchase in desktop printing: a well-maintained example with a fresh nozzle and a flat sheet will produce parts indistinguishable from a new machine's, more slowly, for a fraction of the money.
Check three things on a used one: whether the sheet is warped or scored, whether the extruder has been apart and reassembled correctly, and whether the bed cables show cracking where they flex. The first is cheap to replace, the second is free to fix, and the third is the one that ends printers.
Who should not buy a used one
Y is the binding constraint at 210 mm, not the 250 on X, so large parts are the first reason to look elsewhere. Warping materials are the second, since no accessory turns an open frame into a chamber. The third is anyone who wants to press a button and receive a part, for whom this machine's greatest strength, its openness to intervention, is simply a longer list of things to think about. The Bambu Lab A1 is the honest recommendation in that case.