The A1 mini prints in one colour. Multi-material is an add-on, and Bambu sells one: the AMS Lite, ₹23,000, which arrives assembled and is guaranteed to work with the printer it was designed for. I bought a BMCU instead. It cost ₹7,000 and arrived as a bag of parts.

The easy way to tell that is as a saving, and it was one — ₹16,000 is not a rounding error when you are in school and it is your own money. But that framing gets the trade backwards. The two products are not the same object at two prices. The AMS is a sealed unit that does a job for you. The BMCU is the same job, unsealed, handed over with the expectation that you will finish building it. What the extra ₹16,000 buys is not really hardware. It is the right to never find out how any of this works.

That is a legitimate thing to buy. Most tools are sold this way and should be. I do not want to assemble my phone. But I had not thought carefully before about what the sealed version actually withholds, which is that the filament path — spool to buffer to hub to extruder — is supposed to be Bambu’s problem and not mine. Every part of that path is a mechanism with failure modes. When you buy the assembled unit, the mechanism stays behind a plastic shell and a support email address. When you assemble it yourself, there is no shell. There is a motor, a gear, a board, a wire, and the fact that if you route that wire badly it is going to be your fault in three weeks.

The first real cost was not the ₹7,000. It was half a spool of PLA. Most of the BMCU is a printed object — the body and the moving parts that carry filament come off the same machine they are eventually bolted to — and I printed the whole set in PLA, because PLA is what I had and what I default to. The parts did not fit. PLA was the wrong material for them, and nothing I could do at the assembly stage was going to rescue a set of components that were wrong before I picked them up. So I printed all of it again in PETG.

Half a spool of filament and a second full run of print time, to learn something I could have known by taking the material recommendation seriously the first time. That is the actual shape of the ₹16,000 saving. It is not ₹16,000 that stays in your pocket. It is ₹16,000 minus whatever you waste finding out what the assembled unit already knew.

The assembly itself then took considerably longer than the day I had budgeted for it, and very little of that was skill. It was slowness — checking an orientation twice, working out from a photograph which way a small part is meant to sit, stopping because I was no longer certain a step had gone in correctly. Nothing about it was intellectually hard. It was hard in the way that being responsible for something is hard, which is a different and less flattering kind of difficulty.

What I got at the end was not just a working multi-material unit. It was a mental model of one. I know what is inside it because I put it there, and when something goes wrong I have somewhere to start, because I have held every part of the thing that is failing. That is genuinely worth something, and it is the part people undersell when they talk about DIY as a budget decision.

It is also worth being honest about what I gave up, because there is a version of this post that is smug and it would be a lie. There is no support line. There is no warranty. If the unit misbehaves, the first question is not “what broke” but “whose mistake was this” — the design’s, or mine, or the specific way I seated one connector at eleven at night. That ambiguity is exhausting in a way that a fault code is not. Bambu’s ₹23,000 is, quite reasonably, the price of never having to ask that question.

The half spool also made something clearer that I had been treating as a novelty. The printer builds its own upgrades. The BMCU housing came off its plate, and so did the tilt kit that now raises its back so finished parts slide off, and so did the scraper on the print head. The machine’s only method of improving itself is the thing it already does — which means every modification I make is bounded by how well it currently prints, and by whether I chose the right material that morning. The PLA run was not a printing failure. The parts came out fine. They were just fine in a plastic that could not do the job, which is a failure one level up from the one I was watching for.

And then there is the cost I did not price in at all, which is that my printer is now frozen in time.

The BMCU presents itself to the printer as an AMS, and the printer has always noticed that something is not quite right about it — a firmware mismatch warning that sat there and did nothing. For a long time it was a notice. Recent A1 mini firmware turned that notice into a stop. Users on newer releases report prints being cancelled outright, along with errors that had never appeared before: motor overload, filament that will not push. The BMCU documentation is blunt about the remedy, which is to not update the printer.

So I do not. I am several versions behind, deliberately, and every update prompt is one I dismiss. Whether that is a company protecting users from unsupported hardware or a company protecting its ecosystem is an argument I do not think I can settle, and both readings have honest people on them. What is not in dispute is the position it leaves me in. The BMCU’s own maintainer has said plainly that compatibility keeps tightening and that the unit may eventually stop working on these printers at all, and has started looking at open-source Klipper machines instead — which is what preparing for eviction looks like.

That is the real thing the ₹16,000 was for, and I did not understand it when I was comparing prices. It was never only a warranty. It was the right to press update without wondering whether the machine still works afterwards. I own the printer. I am also, in a specific and slightly uncomfortable sense, a tenant in it.

The A1 mini has not become a better printer through any of this. It extrudes exactly as well as it did the day it arrived. What has changed is the boundary — where the manufacturer’s responsibility ends and mine begins. I moved that line, deliberately, and paid for the move in half a spool of PLA, a second print run, a lost warranty, considerably more time than I planned, and a machine I am no longer allowed to update. I would make the same trade again, but not for the reason I thought I was making it.