Begin With the Order Profile, Not the Equipment List
Small-batch work fails in planning long before it fails on the line. The first task is to pin down, in numbers, the shape of demand you are buying capacity for: how many distinct product families you expect to run in a year, how many units per family, how often the recipe changes, and how quickly you need to respond when a customer changes the specification. A plan that never writes those four answers down will produce an equipment list that looks modern and behaves like a prototype lab.
Write the profile as a sentence your engineers can argue with. Something close to: "we expect to run n product families, in lots of m pieces, with a recipe change on x per cent of orders, and we must ship within y weeks of an order." Everything downstream—how many machines, how much automation, how much manual fixturing—is a response to that sentence, and you can revisit it when the order book changes.
Budget Changeover Time as Production Time
In a high-volume plant, a changeover that costs twenty minutes is noise. In a small-batch plant running twelve lots a week, the same twenty minutes plus the setup and first-article check can consume a working day per change. That is why the changeover figure belongs in the capacity model next to cycle time, not in a footnote.
When you read supplier data, look for the figures that support changeover rather than peak throughput. On the semiconductor process equipment listings, some cards publish a cycle time against a stated test condition; ask the supplier for the equivalent figure for your own changeover: how long to move from one product recipe to the next, what has to be changed by hand, and what has to be re-qualified afterwards. Ask whether recipes are stored and recalled as named programs, because the ability to call a recipe by name is what keeps a changeover repeatable when the operator is not the person who wrote it. If your parts move between laminating, bonding and debubbling steps, ask the same question for every station in the route, since the slowest changeover sets the pace of the whole cell.
Decide What the Recipe Library Has to Contain
Small-batch production lives and dies on documentation. Before you sign an equipment order, define what a recipe means on your site: the process parameters, the fixture set, the material lot references, the torque or pressure settings, and the inspection result that closes the lot. Then ask whether the machine can hold that record and export it, and whether the export is a file you can keep rather than a screen someone photographs.
This is also where the engineering services attached to a machine earn their cost. If your team is small, ask about engineering services for the process qualification stage: who writes the first recipes, who trains your operators, and who stays responsible when a recipe has to be revised after a customer complaint. Get those answers in the quotation. An equipment order with no named owner for process documentation transfers that work to whoever is least busy on your floor, which is how a small-batch line accumulates undocumented settings.
Separate Hard Automation From Flexible Fixturing
Fully automated lines are built for repetition. Small batches are built for variety. The planning decision is therefore not "automated or manual" but which operations are stable enough to automate and which need tooling a technician can change in minutes.
A practical way to split the line is to automate the steps that are hazardous, temperature-critical, or alignment-critical, and to keep flexible fixturing at the handling and loading steps. Where your product family is genuinely unsettled, ask about non-standard automation rather than buying a fixed machine and adapting the product to it—non-standard automation equipment is built to the process you describe, and the design review that produces it is also a review of your assumptions. Ask the supplier to show how a fixture is changed and how long that takes, then time it yourself during acceptance.
Keep one station deliberately manual and instrumented. It becomes your reference point for first articles, for the odd one-off order, and for the recovery run when an automated station is being serviced. A line with no manual fallback loses every order the moment a single station stops.
Plan Maintenance Against Your Spare Capacity, Not the Calendar
With one route and no duplicate machine, an unplanned stop is a missed delivery. The plan therefore has to say who owns preventive maintenance on each machine, what the wear parts are, what the supplier keeps in stock, and how fast a replacement reaches you. Ask for the consumables list per machine and place the first spares order with the equipment, not after the first failure. Ask which faults the machine reports on its own and how that information reaches your maintenance team.
Then write down the minimum you will run at: which single machine, if it were down, still lets you ship. If the honest answer is "none", your plan needs either a spare machine, a second route, or an agreed external capacity—and that decision is far cheaper to make during planning than in the middle of an order.
Put the Numbers in One Table Before You Compare Suppliers
Collect the same six figures for each line you are considering: changeover time per recipe, cycle time at your lot size, first-article time, scrap allowance you are willing to carry, uptime the supplier will stand behind, and the documentation package delivered with the machine. A comparison table built from those columns exposes a proposal that looks cheap but is slow to change, and one that costs more but returns hours every week.
Keep that table as the record of the decision. It is the document your production team will use a year later to justify the next purchase, and it is the checklist you will run against the acceptance test on the day the equipment arrives.
