3D printing

A 3D printer builds an object by depositing or curing material one thin layer at a time, following instructions generated from a digital model. That description covers two quite different domestic technologies which share almost nothing except the principle, and choosing between them is the first real decision.

The hobby divides into three activities that people weight very differently. There is making models, which is CAD and is a design skill. There is printing other people’s models, which is the majority of what happens and is a process skill. And there is maintaining and tuning the machine, which some people regard as an annoying tax and others regard as the actual hobby. It is worth being honest with yourself about which of those you want, because a machine that needs weekly attention is a delight to the third group and a disappointment to the second.

What has changed recently is reliability. A £200 printer bought today arrives largely calibrated, levels its own bed and will produce a good print out of the box, which was not true five years ago. The tuning knowledge below has moved from mandatory to useful — but it is still what you reach for the first time a print fails, and prints still fail.

First decision

Filament or resinReally a decision about where the machine lives and how you feel about solvents.

Most common beginner failure

The first layerAlmost everything that goes wrong goes wrong in the first two minutes.

Cheapest useful upgrade

A flexible build plate

What it is genuinely good at

Objects that do not exist to buyBrackets, jigs, spare parts, adapters — not replacing things you could order.

1. 1

   Step 1: Decide where the printer will live before buying it

   It will run for hours, it is audible, and it wants stable temperature and no draughts. A draught across a printing bed causes warping. Resin additionally needs ventilation and a surface you do not mind contaminating. This constraint eliminates more options than the specification does.

   You will know it worked when: A machine you can actually run without annoying the household.

2. 2

   Step 2: Buy a current mainstream filament printer with automatic bed levelling

   The market moves fast and the sensible advice is generic: buy something current, popular and well supported rather than the cheapest. Automatic levelling and a flexible build plate remove the two most common beginner frustrations. A large user community matters more than a specification sheet, because your first failure will be answered by somebody who has had it.

   You will know it worked when: A machine whose problems are already documented.

3. 3

   Step 3: Print the supplied test model first, unchanged

   It has been sliced by the manufacturer for that machine on that filament. If it prints well, the machine is fine and everything after this is your settings. If it prints badly, you have a hardware problem to solve before you introduce any other variables.

   You will know it worked when: A known-good baseline.

4. 4

   Step 4: Learn the slicer, not the printer

   The slicer converts a model into machine instructions and it is where nearly all of your decisions live: layer height, infill, supports, orientation, temperature. Start from the built-in profile for your machine and material and change one thing at a time.

   You will know it worked when: You can explain why a print came out the way it did.

5. 5

   Step 5: Print other people’s models for a month

   Model repositories carry an enormous amount of pre-tested work. Printing them teaches orientation, supports and adhesion without also asking you to learn CAD. Check whether a model is licensed for what you intend to do with it, particularly if you plan to sell anything.

   You will know it worked when: Process skill, before design skill.

6. 6

   Step 6: Then learn enough CAD to make a bracket

   The hobby changes completely when you can make the thing that does not exist. Parametric CAD — where you specify dimensions and constraints rather than pushing shapes around — is the right family for functional parts, and free browser-based options are entirely adequate.

   You will know it worked when: The printer becomes a way of solving problems rather than a way of printing files.

7. 7

   Step 7: Keep filament dry

   Most filaments absorb moisture from the air, and wet filament prints badly in ways that look like machine faults: popping and hissing at the nozzle, stringing, weak layers, a rough surface. Sealed boxes with desiccant cost very little and solve a problem most beginners misdiagnose for months.

   You will know it worked when: Fewer inexplicable failures.

The printer is most of it, and the accessories that matter are unglamorous. This list assumes filament; the resin additions are called out separately because they are not optional.

What you actually need:

• The printer — Current, mainstream, well supported, with automatic bed levelling and a flexible build plate. A large user base matters more than a specification sheet, because your first failure will already have been answered by somebody who had it.
• A flexible spring-steel build plate — Prints pop off by flexing the sheet rather than being prised off with a blade. It removes the most common actual injury in the hobby and the most common source of damaged plates. Often included; retrofit it if not.
• Filament, and somewhere dry to keep it — Start with PLA from a brand with consistent diameter. A sealed box with desiccant costs a few pounds and prevents a whole family of faults that look like machine problems.
• Spare nozzles — Brass nozzles are consumables and cost a pound or two. A hardened steel one is essential the moment you print anything filled with wood, metal or carbon fibre.
• Digital calipers — £12, and the difference between designing parts that fit and designing parts that nearly fit. The single most useful tool for anybody doing CAD.
• Deburring tool and side cutters — For removing supports and cleaning up edges. Flush cutters get into places a knife should not go.
• A slicer — Free, and where all your decisions live. Use the profile shipped for your machine as a baseline rather than starting from defaults.
• CAD software — Free browser-based parametric CAD is entirely adequate and is the right family for functional parts. Mesh sculpting tools are a different discipline suited to organic shapes.

Resin adds a non-negotiable set: nitrile gloves in quantity, isopropyl alcohol for washing, a wash and cure station or a working equivalent, a way to filter and store part-used resin, and somewhere ventilated to do all of it. Budget £80–£150 beyond the printer, and treat none of it as optional — the handling is the difference between a resin printer being a good purchase and a bad one.

A slicer exposes hundreds of parameters and about six of them account for nearly all of the difference between a good print and a bad one.

• Layer height — How thick each layer is, typically 0.1–0.3mm on a 0.4mm nozzle. Halving it roughly doubles the print time and makes layer lines less visible; it does not make the part stronger. Use fine layers for display models and coarse layers for brackets.
• First layer settings — The most consequential group, because most failures happen here. A slightly slower first layer, a slightly hotter bed and a correct nozzle height account for the overwhelming majority of adhesion problems. If a print is going to fail, it usually fails in the first two minutes — watch them.
• Infill — The internal lattice, expressed as a percentage. The intuition that more is stronger is mostly wrong: strength comes far more from the number of perimeter walls than from infill density. 15–20% is plenty for most things, and going from 20% to 50% adds hours and little strength.
• Perimeters and walls — The solid outer shells. Increasing walls from two to four is a much better way to make a part stronger than raising infill, and it costs less time.
• Temperature — Nozzle and bed. Every filament, and every brand of the same filament, differs. A temperature tower — a test model printed at descending temperatures — settles it empirically in one print, which is more reliable than any published table.
• Supports — Scaffolding for overhangs. The rule of thumb is that overhangs steeper than about 45° from vertical need support, but the better move is usually to reorient the model so it needs less. Supports leave marks, take time and waste material.
• Print orientation — The most underrated setting, and free. Layer adhesion is weaker than the material itself, so a part is strongest across layers and weakest along them. Orient so that forces act across layer lines rather than trying to peel them apart, and a plastic bracket becomes several times stronger at no cost.

• Clean the build surface — Fingerprints are grease and grease prevents adhesion. Warm water and washing-up liquid, dried properly, beats isopropyl alcohol on textured plates.
• Change the nozzle — A brass nozzle is a consumable. Abrasive filaments — anything filled with wood, metal or carbon fibre — will wear one out within a spool or two, and a worn nozzle produces inconsistent extrusion that looks like a hundred other problems.
• Keep the belts tight and the rails clean — Loose belts cause ghosting and dimensional error. Dust on linear rails causes both noise and layer shifts.
• Store filament sealed with desiccant — Not optional for PETG, nylon or TPU, and beneficial for PLA. A sealed box costs a few pounds and prevents a whole family of misdiagnosed faults.
• Check the resin vat film — On a resin printer the FEP or nFEP film is a consumable that clouds and stretches. A failed film during a print is the messiest thing that happens in this hobby.
• Update firmware deliberately, not automatically — Firmware updates can reset calibration and occasionally change safety-relevant behaviour. Note your settings first.

FDM — fused deposition modelling, also sold as FFF — melts a plastic filament and draws each layer with a moving nozzle. It is the technology most people mean by "a 3D printer".

Resin printing — MSLA in most consumer machines — cures liquid photopolymer with ultraviolet light through an LCD mask, a whole layer at a time. Far finer detail, and a genuinely different set of practical demands.

For a first machine the honest answer is almost always filament, unless your entire reason for buying a printer is miniatures — in which case resin is so much better at that one job that the extra handling is worth it. Filament is more forgiving, cheaper to run, and does not require you to think about ventilation and waste disposal before you have printed anything.

The filament types you will actually meet:

• PLA — The default and the right first choice. Prints at low temperature, barely warps, needs no heated chamber, and is derived from plant starch. Its limitation is heat: a PLA part left in a car in summer will sag, so it is not the material for anything that lives outdoors or near an engine.
• PETG — Tougher, more temperature-resistant and more flexible than PLA, and the usual second material. Slightly stringier to print and more prone to sticking to the build plate harder than you want.
• ABS and ASA — Genuinely heat-resistant and impact-tough, and much harder to print — they warp badly without an enclosure and they smell strongly. ASA additionally tolerates sunlight, which ABS does not. Both want proper ventilation.
• TPU — Flexible, rubbery, excellent for phone cases and gaskets. Needs slow printing and a direct-drive extruder; a Bowden setup will struggle with it.
• Filled and specialist filaments — Wood-fill, carbon-fibre-filled, glow-in-the-dark. Most are abrasive and will chew a standard brass nozzle within a spool or two — a hardened steel nozzle is £8 and mandatory for these.

The rest of it, honestly weighted:

• Hot surfaces — A nozzle runs at 200–260°C and a heated bed at 50–100°C. Neither is dramatic, both will burn you, and the nozzle is at hand height on most open-frame machines.
• Fumes and particulates — Printing releases ultrafine particles and volatile compounds, and how much depends heavily on the material — ABS considerably more than PLA. The sensible response is proportionate rather than alarmed: ventilate the room, do not sleep next to a running printer, and use an enclosure with filtration if you print ABS or ASA regularly. We are not going to quantify a risk here, because the published measurements vary widely by material and machine and we have not assessed that literature.
• Fire risk — Low but not zero: a printer is a heater running unattended for hours. Use the machine on a non-flammable surface, keep it clear of anything that could fall against it, fit a smoke alarm in the room, and treat thermal runaway protection in the firmware as a feature you check for rather than assume.
• Scrapers and knives — The most common actual injury in the hobby is a cut hand from prising a print off a rigid build plate with a blade. A flexible plate removes the tool and the injury together.

The hobby is built on shared model repositories, and the sharing culture is genuinely one of its best features — an enormous amount of useful, tested, free work exists because people uploaded it. The corollary is that licences matter and are widely ignored. Many models are published under terms that permit personal printing but not commercial sale, and "I printed it myself" is not a defence. If you intend to sell prints, read the licence.

There is a persistent tension between the maker culture the hobby grew out of — open hardware, self-repair, modification — and the increasingly closed, appliance-like machines that made it accessible. Cheap reliable printers with proprietary firmware and cloud slicing brought in a great many people who would never have tuned a bed, and they annoy a great many people who did. Both positions are reasonable and it is worth knowing the argument exists before you wander into it.

One norm worth adopting: publish the settings alongside anything you share. A model posted with its layer height, material and orientation saves everybody who downloads it an hour.

1. 1

   Step 1: A benchmark print that succeeds

   Bed levelling, first-layer adhesion, and one filament you understand. Almost all early failures are the first layer.

   You will know it worked when: Prints stick, finish, and come off cleanly three times running.

2. 2

   Step 2: Tuning rather than guessing

   Temperature towers, retraction tests, flow calibration. This is the step that converts a printer from a lottery into a tool.

   You will know it worked when: You can predict how a model will come out before you start it.

3. 3

   Step 3: Designing your own parts

   CAD, tolerances, and designing for how the machine actually builds. Printing somebody else's model is consuming; designing is the hobby.

   You will know it worked when: A part you designed fits something in the real world first time.

4. 4

   Step 4: The plateau

   Materials. PLA is forgiving and everything else is not, and people stall when a first attempt at PETG or ABS goes badly rather than treating it as a new calibration problem.

5. 5

   Step 5: Where it branches

   Resin printing for detail, functional engineering parts, or printing as one step in a larger making process such as prop or model work.

6. 6

   Step 6: When to upgrade

   When a specific limitation blocks a specific project — build volume, temperature, or accuracy. Not before, and not because a newer machine exists.