What Makes Lab Printing Different from Production Printing
Production printing optimises for cost per part and throughput. A research lab optimises for something the production floor rarely cares about: whether the part a student prints in March matches the part a postdoc printed in October for the same experiment. That single difference invalidates most of the buying advice written for industrial buyers.
Four characteristics separate a lab workload from a production one. Geometry is one-off and revision-heavy, so a machine's value lies in its ability to handle arbitrary geometry rather than in repeat production speed. Volume is low and unpredictable, which means the machine will sit idle for weeks and then be needed continuously before a conference deadline. Users are many and transient, with a rotating population of students and visiting researchers who each bring their own file conventions. And output is sometimes published, which is the characteristic that changes the requirements most, because a printed apparatus described in a methods section becomes something a reader may attempt to reproduce.
The consequence is that a lab should weight reliability, calibration stability and record-keeping above speed and build volume. A machine that produces a documented, repeatable part matters more than one that produces it quickly, because a fast part with no parameter record is not usable in a methods section.
Reproducibility — The Requirement Most Labs Under-Specify
Ask a lab what a printer needs to do and the answer is usually about resolution or materials. Ask what happens when the printed part is central to a published result and the requirement changes entirely, because reproducibility is not a property of the machine alone. It is a property of the machine, the material lot and the recorded parameters together.
Print-to-print variation is real and larger than most researchers expect. On a well-calibrated FDM machine running the same file, dimensional variation across repeated builds is typically in the region of ±0.1 to ±0.3 mm depending on geometry, and mechanical properties vary more than that. Variation grows when the machine drifts, which happens gradually and invisibly: belts relax, nozzles wear, and a machine that printed perfect parts in January may be producing slightly different parts by June with no single obvious failure.
Three practices close most of that gap, and none of them is expensive. First, save the profile, not just the model. A sliced file is a record of the exact layer height, temperature, flow and cooling used, whereas a raw model file is not. Second, verify calibration on a schedule rather than on demand: a periodic artefact print that is measured and logged catches drift before it contaminates a result, and the measurement itself takes minutes. Third, record the material lot, because filament and resin both vary between production batches and a change of lot mid-experiment is a variable nobody intended to introduce.
Materials for Lab Hardware
Lab components fail because of what they are exposed to, so the material should be selected from the exposure list rather than from what is already on the shelf. The common exposures in a research setting are solvents, acids and bases, autoclave or heat cycles, UV, and biological contact.
- Solvent and chemical exposure. PLA is a poor choice here: it softens and swells with many common solvents, including acetone and chlorinated solvents. PETG is more resistant but not universal. Where chemical resistance is the governing requirement, chemically resistant materials should be the starting point rather than an upgrade, and a compatibility check against the specific solvent should precede any design commitment.
- Temperature exposure. Anything that will be autoclaved, oven-dried or used near a heat source needs a material whose glass transition sits well above the service temperature, with margin. PLA is unusable above roughly 60 °C, and a part that survives one cycle may creep over many. High-performance polymers extend the ceiling substantially, and the trade-offs are set out in our guide to high-temperature materials.
- Biological and clinical contact. Where a printed part touches cells, tissue or a patient pathway, the question is not strength but leachables, sterilisability and in some cases a regulatory submission. That is a different design exercise, covered for device work in our medical regulatory guide, and it should not be improvised on a general-purpose machine.
- Optical and fluidic parts. Transparent resin is the correct route for microfluidic geometries and imaging windows, but the post-cure schedule determines the final optical clarity and must be recorded, because an under-cured transparent part degrades in clarity over weeks.
One practical warning applies across all four: do not mix materials within a single functional assembly. Different polymers shrink by different amounts during cooling, so an assembly with a PLA frame and a PETG insert will fit when printed separately and bind when assembled. Choose one material per assembly, and where a material change is unavoidable, allow for the shrinkage difference explicitly rather than at the drill press.
Multi-User Access and Print Queue Discipline
A shared lab printer fails socially before it fails mechanically. The common failure is not a broken machine but a fourteen-hour job queued by one user at 4 pm that blocks four people with a deadline, producing a policy argument that eventually damages the lab's willingness to share equipment.
Four rules resolve almost all of it. Jobs have named owners, so that a failed print can be reported to a person rather than discovered silently. Long jobs are scheduled, not queued: anything expected to run beyond a few hours books a slot, which also encourages users to section large parts into printable pieces. The queue is visible, whether on a whiteboard or a shared sheet, so nobody discovers the blockage after submitting. And failed prints are logged, not quietly discarded, because the failure rate is the only honest measure of whether the machine is healthy.
A second discipline concerns access. In a lab where the printer is available to anyone with the door code, the machine accumulates undocumented changes: a nozzle swapped for a different diameter, a bed re-levelled, a profile overwritten. Every such change invalidates the record for anything printed before it. Gating the machine behind a short induction and keeping a change log next to it costs an hour of administrative effort and protects the reproducibility of everything the lab prints.
Traceability and Documentation for Publication
A printed part that appears in a methods section is, in effect, a claim that a reader could build the same apparatus. Very few labs can currently support that claim, and the ones that can do it with a short, low-effort record created at print time.
The record should capture, per part: the machine identifier and nozzle diameter; the material brand and, where relevant, the lot number; the slicing profile name and its key parameters, at minimum layer height, extrusion temperature and infill; the date of printing; and for any part whose dimensions matter to the result, the measured dimension of a calibration artefact printed on the same machine in the same period.
Two further items are worth the small effort they cost. Post-processing must be recorded, because annealing, solvent smoothing and UV post-cure all change dimensions and mechanical properties, and a part that was annealed is not the same part as one that was not. And the model file should be archived in a versioned location rather than living on one student's laptop, because a replication request typically arrives a year or two later, long after that laptop has been retired. Where a lab maintains an electronic lab notebook, a link to the archived model and the profile name alongside the relevant experiment entry is sufficient; the goal is not a formal quality system but a record that survives staff turnover. Those working to a formal standard will find the inspection and reporting discipline in our part metrology and dimensional inspection guide a useful reference point.
Machine Classes by Lab Type
The correct machine follows from what the lab actually makes, and the four common lab profiles have genuinely different requirements. The table below sets out the governing constraint for each rather than a general recommendation.
Two of those constraints deserve expansion because they are routinely underestimated. Enclosed, actively heated machines are not a convenience in a materials lab; they are what makes engineering polymers printable at all, and a machine without a heated chamber will fail on exactly the materials the lab bought it for. And resin-based processes in a biology or chemistry lab bring handling requirements that are easier to satisfy in a dedicated room with proper ventilation and waste handling than on a shared bench, which is a facilities decision rather than a printer decision.
For labs that also run test rigs and experiment-specific fixtures, the supporting tooling question is often as significant as the printer itself, and our guide to printed tooling for forming processes covers the case where a printed mould or form sits upstream of the experiment. Where a lab needs several machines and wants them serviceable over a long equipment life, the durability and spares questions are covered in our critical spares kit design guide.
Grant-Funded Procurement — What to Put in the Justification
Equipment purchased on a grant is justified in the language of the funding body, and printer quotations that list only the machine tend to come back with questions. Three additions make a quotation defensible without inflating it.
First, state the capability the lab did not previously have and tie it to specific work, because “faster prototyping” is not a justification and “enables in-house fabrication of flow-cell geometries that currently require an external machining contract at 4–6 weeks lead time” is. Second, include the consumable and spares budget as a line item: nozzles, build surfaces, and at least one spare hotend are recurring costs, and a printer funded without them is a printer that stops working in month eight. Third, name the service expectation, so that a warranty term and a response time are visible in the award rather than discovered afterwards.
Where the machine will be shared, say so, because shared equipment is usually more defensible than single-user equipment and the funding body will want to know which groups benefit. A short list of the research groups that will use the machine is more persuasive than any technical specification.
Specifying the Machine
Four requirements follow from everything above and should be stated on any lab specification before price is considered.
- Documented parameter capture. The machine or its software must make it easy to export the full slice profile alongside the print, because a machine whose settings live only in an operator's memory cannot support a published method.
- Calibration stability and a serviceable hotend. A nozzle is a consumable and a hotend is a field-replaceable part. A lab that cannot change its own nozzle is dependent on a service call for a routine maintenance item.
- Enclosure where the workload demands it. For engineering polymers this is mandatory; for teaching labs it prevents both draught-induced failures and accidental contact with hot or moving parts.
- Multiple identical machines where throughput matters. Two mid-format machines are almost always a better lab investment than one large machine, because two machines give redundancy, parallel workloads and a smaller failure blast radius.
If your lab works with printed hardware that has to hold to a specification across users and months, that is the requirement we design our engineering-grade machines around, and we are happy to work from your actual list of printed parts, materials and required envelope rather than a general recommendation. Send the parts list and we will return a configuration with the consumables and spares budget included.
