Matching a part to a master you already have
The characteristic problem in prototype work is that the acrylic part has to fit something that already exists — a 3D printed housing, a machined chassis, a hand-made model. The drawing for the acrylic part may be perfect and the assembly still will not close, because the master itself is not to drawing.
Printed parts shrink, warp and vary between printers. A housing that measures 100.0 mm on the drawing may arrive at 99.6 mm, and a cover machined to 100.0 mm will not fit. The tolerance that matters is not the one on your drawing — it is the gap between your part and ours.
So for prototype covers we ask for the measured dimension of the master, not the nominal one. Where that is not available, we machine the cover slightly oversize on the mating faces and let you trim, or we build in a designed clearance of 0.3–0.5 mm and tell you where it is.
Speed versus finish — you can have either first
A prototype part usually has one job: to be in someone's hands quickly so a decision can be made. That argues for as-machined edges and no polishing, which is fast and cheap and perfectly adequate for a fit check.
A presentation part has the opposite requirement. It will be photographed and handled by people judging the design, so the edges have to be polished and the faces clean. That adds a day and real cost.
We quote both versions on prototype jobs as standard, labelled "fit-check" and "presentation", so the choice is explicit rather than a surprise on the invoice. For most projects the answer is: one polished presentation sample, and the rest as-machined for fit checks.
Bridge production before you commit to a mould
If the end game is injection moulding, the prototype stage has an extra purpose: proving the design before the tool is cut. Acrylic or polycarbonate parts machined from sheet can serve as bridge production for the first hundred or thousand units, while the mould is being made.
This is not the same as moulding. Machined parts have different draft, different radii and no gate marks; the surface finish is polished rather than moulded. But for a market test or an early customer shipment, it is often good enough and it removes the schedule risk of waiting for a tool.
Where the design is heading for moulding, it is worth telling us at the prototype stage. We will flag features that cannot be machined but can be moulded, and features that are easy to machine but would need a slider in a tool.