I spent a good chunk of the last month getting reliable multi-material prints off a single-extruder machine, and I want to write down the honest version before the novelty wears off and I start remembering it as easier than it was. Short answer: it works, the results can be genuinely lovely, and the cost in filament and time is higher than anyone selling you the idea tends to mention.
By multi-material here I mean the affordable hobbyist route: one hotend, one nozzle, and a mechanism that swaps which filament gets fed into it. Not a true dual-extruder, not an industrial machine with soluble support and a heated chamber. The kind of setup a person can have on a desk in a spare room.
how the swap actually works
The core problem is that you have one melt zone and you want more than one colour or material passing through it. So between every colour change you have to get the old plastic out and the new plastic in, and you have to do it cleanly, because any leftover of the previous colour bleeds into the next.
The way the cheap end solves this is the purge. Before printing with the new filament, the machine extrudes a slug of it somewhere off to the side until the colour runs true. On my setup that "somewhere" is a purge tower built next to the model, or a wipe into infill, or a dedicated waste bucket depending on how I've sliced it.
And this is the bit that surprised me. The purge is not a rounding error. On a print with frequent colour changes, I have watched the purge tower end up using more plastic than the actual object. A small two-colour keycap that masses a couple of grams can drag a purge tower behind it that masses five or six. You are not printing a model with some waste. You are sometimes printing the waste with a model attached.
tuning the purge volume
The single most useful thing I learned was that the default purge volumes in the slicer are conservative, deliberately, because they're set to cope with the worst case: dark filament bleeding into a pale one. Going from black to white needs a long purge. Going from white to black needs almost none, because a tinge of white in black is invisible.
Most slicers let you set a purge volume per transition rather than one global number. Filling in that matrix by hand, with bigger volumes for light-after-dark and tiny ones for dark-after-light, cut my waste by roughly a third without any visible bleed. It's tedious to set up once and then it just sits there saving you filament on every print.
A rough table of what I settled on, in mm³ of purge:
to white to red to black
white - 70 40
red 180 - 40
black 220 140 -
The asymmetry is the whole point. The expensive transitions are the ones into a lighter colour, and you pay for those whether you like it or not.
the failure modes are different
A single-colour print fails in familiar ways: bed adhesion, a clog, a layer shift. Multi-material adds a new category, the swap failure. The most common one for me was the new filament not loading cleanly, so a layer or two prints with a thin or missing extrusion right after a colour change, because the previous filament's tip ground itself into a blob that didn't feed.
The fix was mostly mechanical and mostly patience: making sure the cut on the filament tip was clean, slowing the load, and accepting that the first prints after any maintenance would be test pieces. The thing nobody tells you is how much of multi-material is really filament-handling: tip shape, grinding, moisture, the geometry of the path the filament takes to reach the hotend. A drop of moisture that you'd never notice on a single-colour print becomes a swap failure when the timing is tight.
so, is it worth it
It depends entirely on what you're making, and I've landed somewhere unromantic.
For functional parts, mostly no. If I'm printing a bracket, a jig, a replacement clip for something that broke, the colour is irrelevant and the part is stronger and cheaper in one material. Multi-material there is pure cost with no benefit.
For text and embedded labels, yes, and this is where it genuinely earns its place. A part with a contrasting recessed label, printed in two colours, looks properly made in a way a painted-on or engraved one doesn't. No second operation, no paint to chip off. That alone justifies the kit for me.
For decorative multi-colour models, it's a wash that comes down to mood. The results can be beautiful. They also take twice as long, waste a pile of filament, and fail in ways a single print wouldn't. Some weekends that trade is exactly what I want from the hobby. Other weekends I want a thing to come off the bed and be done, and then I load one spool and don't think about purge towers at all.
What I'd tell anyone considering it: don't buy in expecting it to be free colour. Buy in expecting to pay for every transition in plastic and time, tune your purge matrix early, and treat the first prints after any change as sacrificial. Do that and the good results stop feeling like luck. Skip it and you'll spend your filament budget on towers and wonder where it all went.
One last practical note. Keep the purged filament. It's clean plastic, it just happens to be the wrong colours fused into a sad little brick. Some of it is recyclable, some of it makes perfectly good calibration test pieces where the colour is irrelevant, and watching the bucket fill up is the most honest reminder of what multi-material actually costs. The model on the bed is the part you wanted. The bucket is the part you paid for.