Jigs, fixtures, end-use components, low-volume production runs and replacements for parts nobody makes any more. Polymer and metal powder bed fusion, resin and extrusion — quoted against your drawing, machined where the drawing demands it, and measured before it ships.
Additive is not one technology. The same bracket costs, weighs and fails differently depending on how it is built, and the honest answer for a given part is often the least glamorous process. Each entry below states what it is genuinely good at and where it will disappoint you.
Nylon powder fused by laser in a heated bed. The unfused powder supports the part, so there are no support scars and no draft or pull direction to design around — internal channels, living hinges and thin captive features all build. The workhorse for functional plastic parts.
Surface is a uniform matte grain. Trapped volumes must have escape holes or the powder stays in the part.
Fusing and detailing agents jetted onto a nylon powder bed and driven by infrared. Denser and more repeatable part-to-part than laser sintering, with crisper small features and sharper edges. Good when you need the tenth part to measure like the first.
Native parts are grey and are usually dyed black. Same powder escape rule as SLS.
Resin cured layer by layer. The highest feature resolution and the smoothest as-built surface of anything here — the right choice for master patterns, optically clear parts, fine text and lettering, and form-and-fit checks where the surface is being judged.
Photopolymers creep under sustained load and yellow under UV. Not a long-service structural material.
Extruded engineering thermoplastic. The route to genuinely large parts and to polymers the powder processes do not offer — ABS, ASA, polycarbonate, PEI, PPS and the carbon-filled grades. Chemical and temperature resistance you can look up on a real datasheet.
Anisotropic: weaker across the layers than along them. Build orientation is a design decision and we will tell you what we chose.
Photopolymer droplets jetted and cured, with rigid and elastomeric materials in a single build. Used for multi-durometer parts, overmould and seal appearance models, and colour-accurate presentation pieces where the look of the assembly is the deliverable.
Same photopolymer caveats as SLA, plus support removal marks on down-facing surfaces.
DMLS-type metal printing in stainless, tool steel, aluminium, titanium and nickel superalloy. Conformal cooling channels, consolidated weldments, topology- optimised brackets, and spares in alloys that are no longer stocked in the section you need.
Always stress-relieved on the plate. Sealing faces, bores and threads are machined after — printed metal is a near-net blank, not a finished surface.
A kiln process rather than a beam process. The printer lays down build powder alongside a separate support powder chosen to be non-wettable, so nothing soaks into the support, and the filled crucible then goes into a kiln. From there the route splits. With infill, a second metal is added and the crucible is held above the infill's melting point but below the build powder's — the infill wicks through the build powder only, and the part comes out fully solid. Without infill, the bake simply sinters the build powder, which is quicker but leaves the part porous.
What you get out is set by how you cool it. Cool fast and the two powders stay distinct, giving a composite. Cool slowly and atoms diffuse across the boundary into a uniform alloy — copper and tin held around 790 °C mix into bronze in roughly two hours. The crucible atmosphere has to be reducing to stop the powder oxidising, which means carbon or a hydrocarbon packed in with it, or a controlled hydrogen–argon furnace for reactive metals. Out of the kiln, parts are cleaned back with a wire brush or a blast cabinet.
Two things make it worth having. Because the whole part heats and cools together there is no sharp thermal gradient, so it does not build in the locked-in stress and warping that layer-wise melting does — you get a solid metal part that is not fighting itself. And it is materially agnostic in a way laser processes are not: anything that flows through the pourer aperture can be placed, which puts glass, ceramics and composites in scope alongside metal. Cost per part is substantially below laser powder bed fusion.
Resolution, first and loudest. The minimum feature width is one pourer diameter, and pourers run roughly 0.9 mm to 9.8 mm. Minimum layer height is around 0.1 mm, so the vertical axis is much finer than the horizontal one. That makes SPD one to two orders of magnitude coarser in plan than laser powder bed fusion. Do not specify it for a part that needs fine detail.
The sintering route shrinks and warps, and reaching full density with pure metals is hard — glass and some alloys are considerably more forgiving.
Not every powder pair will alloy. Copper and tungsten held at 1090 °C will not alloy appreciably in a day; you get a composite. Alloy or composite is a design input to settle up front, not something to discover on delivery.
Build envelopes are modest, approximately in the range of a 280 × 270 × 110 mm class up to a 610 × 610 × 310 mm class.
In short: right for larger, solid, low-detail metal parts where cost matters and fine features do not, and for material combinations the laser processes cannot do at all. Wrong for fine detail or tight tolerance — those go to L-PBF, or get machined after.
If the material you need is not listed, ask — the list below is what comes up most, not the limit. If your part has a service temperature, a chemical exposure or a load case, tell us that instead of naming a material and we will propose one.
| Material | Process | Use it for |
|---|---|---|
| PA12 nylon | SLS · MJF | The default functional plastic. Tough, dimensionally stable, takes heat-set inserts and snap fits. Most jigs and enclosures start here. |
| PA12 glass-filled | SLS · MJF | Stiffer and more heat resistant than unfilled PA12, at the cost of ductility. Fixture bodies and brackets that must not deflect. |
| PA11 nylon | SLS | More ductile and impact tolerant than PA12. Clips, hinges, snap features and anything that has to survive being dropped. |
| TPU | SLS · MJF · FDM | Flexible elastomer. Gaskets, bumpers, vibration isolators, soft jaws and grippers that must not mark the workpiece. |
| ABS and ASA | FDM | General purpose housings. ASA holds up outdoors where ABS chalks and embrittles under UV. |
| Polycarbonate | FDM | High stiffness and impact strength with a useful heat deflection temperature. Guards, shrouds and load-bearing covers. |
| PEI (ULTEM-type) | FDM | High temperature and flame-retardant grades. Ducting, under-hood brackets and parts that see sustained heat. |
| PPS and carbon-filled grades | FDM | Chemical resistance and stiffness. Solvent-exposed fixtures and parts replacing machined composite. |
| Rigid, tough and high-temp resins | SLA · DLP | Fine detail, smooth surfaces, crisp lettering. Fit checks, master patterns and short-service parts. |
| Clear and castable resins | SLA · DLP | Light pipes, fluidics visualisation, lenses for form review, and burnout patterns for investment casting. |
| 316L stainless | L-PBF | Corrosion-resistant general metal. Manifolds, brackets, wash-down and marine-adjacent hardware. |
| 17-4 PH stainless | L-PBF | Precipitation hardening to high strength after heat treatment. Loaded fittings, shafts, tooling inserts. |
| AlSi10Mg aluminium | L-PBF | Light, thermally conductive, machines cleanly. Heat sinks, housings, structural brackets where mass matters. |
| Ti-6Al-4V titanium | L-PBF | Strength-to-weight and corrosion resistance. Lightweight structure and chemically aggressive service. |
| Inconel 718 | L-PBF | Nickel superalloy for high temperature and hot corrosion. Combustion-side hardware and hot-gas path fixtures. |
| Build powder plus a lower-melting infill | SPD | The infill route. Any pair where the infill melts below the build powder: the kiln is held between the two melting points and the infill wicks through, giving a fully solid part with no layer-wise thermal stress. |
| Copper and tin → bronze | SPD | The alloying route. Held around 790 °C the two diffuse into a uniform bronze in roughly two hours. Cool quickly instead and they stay distinct. |
| Non-alloying pairs → composite | SPD | Copper and tungsten at 1090 °C will not alloy appreciably in a day, so the part is a metal–metal composite. Often that is the point — but decide it before the build, not after. |
| Glass, ceramics and filled composites | SPD | Anything that flows through the pourer aperture can be placed, which puts non-metals within reach of a printed part at all — laser powder bed fusion cannot process them. |
| Sintered metal, no infill | SPD | Baked hot enough to sinter the build powder on its own. Faster and simpler, but porous, and it shrinks and warps as it densifies. |
The figures below are typical general tolerances for well-conditioned equipment on parts of moderate size, and they are geometry dependent — a long thin feature and a short thick one in the same material do not behave the same. Treat them as a planning guide. Call out your critical dimensions on the drawing and we will confirm, per part, what is achievable as-built and what needs a secondary operation.
| Process | Typical general tolerance | Smallest sensible feature | As-built surface |
|---|---|---|---|
| SLS | ± 0.3 % · min ± 0.30 mm | 0.7 mm wall | Uniform matte grain |
| MJF | ± 0.3 % · min ± 0.30 mm | 0.5 mm wall | Fine matte, sharper edges |
| SLA / DLP | ± 0.2 % · min ± 0.15 mm | 0.4 mm wall | Smooth; near-injection look after finishing |
| FDM / FFF | ± 0.5 % · min ± 0.40 mm | 1.0 mm wall | Visible layer lines along the build axis |
| L-PBF metal | ± 0.2 % · min ± 0.20 mm | 0.5 mm wall | Coarse; machined on any functional face |
| SPD | Set by pourer and shrinkage — machine any close dimension | 0.9 mm wall (one pourer) | Very coarse; wire-brushed or blasted out of the kiln |
SPD sits at the bottom of that table deliberately. Its minimum feature width is one pourer diameter, and pourers run roughly 0.9 mm to 9.8 mm — one to two orders of magnitude coarser in plan than laser powder bed fusion. Minimum layer height is around 0.1 mm, so the vertical axis resolves far better than the horizontal one. If a feature needs fine detail or a close tolerance, it should be built by L-PBF or machined into an SPD part afterwards; specifying SPD and hoping will not work.
Every printed part gets some post-processing; the question is how much you are paying for. Say what the part has to look like and what it has to survive, and the quote will name the steps rather than leaving "finished" to interpretation.
Additive is not cheaper than machining or moulding for everything. It wins where the quantity is low, the geometry is awkward, the tooling does not exist, or the part is needed before a tool could be cut. These are the cases that come up most.
Assembly nests, weld and bond fixtures, drill and router guides, CMM holding fixtures, soft jaws that will not mark a finished surface, pick-and-place trays, and robot end-of-arm tooling. Conformal pockets that would take a five-axis setup print as a single body, and when the product revision changes you edit the model instead of scrapping an aluminium plate.
Enclosures, ducts and manifolds, cable management, sensor and camera mounts, shrouds and covers. Parts that go into the machine and stay there, specified in a material with a datasheet you can defend, and repeatable enough that the replacement in two years matches the original.
The bridge between a working prototype and a tool that has not been paid for. Runs in the tens to low thousands with no tooling amortisation, so unit cost is flat with quantity and a design change between batches costs drawing time rather than a tool modification. For solid metal parts at this quantity where the detail is coarse, SPD usually lands well under the laser processes on unit cost.
A part on a machine still in service that the original supplier stopped making and nobody will tool for again. We work from the surviving sample, the original drawing if it exists, or a 3D scan, rebuild it as a solid model, and print it in a material matched to the duty. You get the model back as well as the part, so the next one is a purchase order rather than a project.
Form, fit and function parts through the iterations where the design is still moving. Resin where the surface is being judged, nylon where the part is being loaded or assembled, printed metal where you need to test the thermal or structural behaviour rather than look at it.
Solid geometry quotes faster and more accurately than a mesh, because we can measure features, check wall thickness and lightweight the part without approximating anything. Send a mesh if that is what you have — it is workable — but send a solid if you can.
STL files carry no units. A model exported in inches and read as millimetres arrives 25.4 times too small, and it will still print perfectly — just useless. It is the single most common quoting error in this industry. One line in your email saying "units are millimetres" removes it entirely.
Also tell us the bounding box if the part is large. Parts beyond a single build envelope can be split and bonded or keyed together, and if that is what we intend to do it will be stated in the quote rather than discovered on delivery.
Five steps, in order. The point of the sequence is that every judgement we make about your part — orientation, process, where a feature gets machined instead of printed — is written down before anything is built, so there are no surprises in the box.
A STEP file or 3MF, a drawing if you have one, the quantity, the units, and what the part has to do in service. Ask for an NDA first if you would rather have one in place before the files move.
We check wall thickness, unsupported spans, warp and shrink risk, trapped powder or resin volumes, and features that would be cheaper or more accurate machined. Anything we would change comes back to you marked up.
Process, material, finish, build orientation, post-processing steps, unit price at the quantities you asked for, and a lead time. Assumptions are listed, so you can see what the number depends on.
Parts are built, de-powdered or de-supported, heat treated and machined where the quote said so, finished as specified, and dimensionally checked against the critical dimensions you called out.
Parts ship with the inspection notes and the build record — process, material lot, orientation and post-processing — so a repeat order in a year reproduces the part rather than reinventing it.
Yes, and plenty of jobs start that way. The limits are worth knowing: a mesh is a faceted approximation, so we cannot reliably measure a nominal diameter, move a feature, thicken a wall or change a fillet without rebuilding the model first. Tolerance expectations are correspondingly looser. If a solid file exists anywhere in your organisation, that file gets you a better part.
Tell us the units. STL does not record them.
Not automatically, and the honest comparison depends on the process. Powder bed nylon parts are reasonably consistent in all directions and behave close to published material data. Extruded parts are weaker across the layers than along them, sometimes substantially — which is why build orientation is a design decision and appears on your quote. Photopolymer resin parts should not be treated as long-term structural material at all: they creep under sustained load and degrade under UV.
Metal powder bed parts, once stress-relieved and heat treated, can be comparable to wrought material in the properties that usually matter, but they are not identical to it. If a load case is critical, say so and we will discuss the material data and whether a test coupon is worth building alongside.
When the part is large-ish, solid, metal, and does not need fine detail — and when cost matters. Selective powder deposition bakes a whole crucible in a kiln rather than melting one track at a time, so the part heats and cools as a single body. There is no steep thermal gradient, which means none of the locked-in residual stress and warping that layer-wise melting has to be designed around. With an infill metal the result is fully solid. It is also materially agnostic in a way the laser processes are not: anything that pours through the aperture can be placed, so glass, ceramics and composites are available, and metal pairs can be made to come out as a uniform alloy or as a deliberate composite depending on how fast the kiln is cooled. Cost per part is substantially below L-PBF.
The trade is resolution, and it is a big one. Minimum feature width is one pourer diameter — pourers run about 0.9 mm to 9.8 mm — against roughly 0.5 mm for L-PBF, so SPD is one to two orders of magnitude coarser in plan. Layer height gets down to about 0.1 mm, so height resolves better than width. The sintering route, with no infill, shrinks and warps as it densifies, and reaching full density with pure metals is genuinely difficult, though glass and some alloys behave much better. Not every powder pair will alloy: copper and tungsten held at 1090 °C will still not alloy appreciably after a day, so you get a composite whether or not you wanted one. Build envelopes are modest, roughly a 280 × 270 × 110 mm class up to a 610 × 610 × 310 mm class.
So: a heavy solid bracket, a counterweight, a heat-spreading block, a glass or ceramic-bearing part, or a metal combination no laser can process — SPD, and it will be cheaper. A part with 1 mm ribs, a fine thread or a tight bore — L-PBF, or SPD with those features machined in afterwards. We will say which one your part is before you order rather than after.
Usually. Send the surviving part, or the broken halves, plus any photographs or fragments of documentation. We measure or scan it, rebuild it as a solid model, and propose a material matched to what the part actually does rather than guessing at the original specification.
One caveat stated plainly: we cannot recover the original material specification from a sample by inspection. If the alloy or polymer grade matters to the duty, that needs laboratory analysis, or an engineering decision to specify a material we can justify. We will say which one applies to your part.
No — and we would rather say so here than quote around it. We do not hold industry qualification approvals, so we do not supply parts requiring certified flight status, implantable or patient-contact medical status, or food-contact compliance. Ground support equipment, laboratory and production fixtures, non-contact housings and R&D hardware in those industries are ordinary work for us; the certified article is not.
One part. There is no tooling to amortise, so a single piece is a normal order rather than an exception. The economics run the other way from moulding: quantity one is proportionally more expensive per part than quantity one thousand mainly because of setup and handling, not because of tooling.
Files are treated as confidential, used only to quote and build your parts, and are not shared outside the people working on the job. We will sign your NDA before you send anything — attach it to the quote request below, or ask for it in your first email, and we will return it executed before we need any geometry. We do not ask you to accept a document of ours. If you would rather send a simplified or feature-suppressed model for the initial quote, that works too.
The form takes attachments up to 10 MB in total. Real STEP files pass that routinely and fine STL passes it by a wide margin, so if your files are bigger — or you would simply rather not put them through a web form — tick the box on the form and send nothing. We will arrange the exchange directly once we are in contact.
Yes. Design for additive is where most of the cost and most of the failures live — wall thickness, orientation, where to put a machined datum, when to consolidate a weldment into one body and when consolidation makes the part impossible to inspect. That review is part of quoting at no charge for a normal part. Larger redesign or scan-to-CAD work is quoted separately and stated up front.
Three fields are required — your name, your email and what the part is. Everything below that is optional and only sharpens the quote. Answer what you know and leave the rest to us.