Who should use this checklist

This list is for two kinds of readers. First: an OEM or EMS operations lead who already has a signed order and a fixed date, and whose SMT production line is not yet ready. The date will not move. Second: someone building a new SMT production line where the equipment lead time is already longer than the whole setup window you were planning around.

I have been on both sides of that list. I work in production scheduling. For the last five years, most of my time has been on projects where the equipment was already in motion and a client's purchase order had already fixed the date. That does not make me an expert in every piece of SMT equipment. But it does give me a strong sense of which steps in an SMT production line sourcing process actually protect the schedule, and which are decoration.

There are seven steps below. Work them in order — the sequence matters, because several of the answers change depending on what you find in an earlier step.

Step 1: Lock the product mix before you touch an equipment list

Most sourcing failures start here. A team gets excited about placement machines and requests quotes before anyone has written down what the SMT production line will actually run.

Before the first supplier call, write down three things:

  • The board size range you will build, including the largest panel you ever expect to run
  • The highest component count on any single product, and the smallest component size (01005, 0201, 0402?)
  • The high-mix versus high-volume ratio, defined in real numbers — because a line that switches from 500 to 50 units every week is a fundamentally different machine than one that runs 20,000 units continuously

The thing most people miss is the mix ratio. That number is what decides whether you are buying for flexibility or for pure throughput. Get it wrong, and every supplier downstream will happily sell you a high-throughput machine you can't change over fast enough.

Checkpoint: Have you written down a number that actually rules out some supplier options? If every supplier can meet it, the number is not specific enough.

Step 2: Do the layout walk before you quote equipment

Before anyone sends you a model-name quote, get the building dimensions, column spacing, ceiling height, floor loading requirements, and the door and dock path your equipment will travel.

A lesson I still feel. I sent a supplier a layout file and got back "no problem, it fits." It fit — until the equipment arrived and the feeder aisle was 40 cm wide. Everything was inside the building. Nobody could walk down the feeder aisle to replenish parts.

Walk it on paper first. Mark the steel and stencil storage area, the feeder carts, the rework bench, and the inspection station. Then — and this matters more than it sounds — put your actual line operators into the layout and walk their route from stencil load to AOI. If you can, walk it physically, column to column, before the SMT production line layout is approved.

Here's something vendors won't tell you: their "standard clearance" is defined around the machine, not around the human walking behind it.

Checkpoint: Can you describe the physical path from material receiving through print, reflow, and inspection without hand-waving? If it only exists in a slide, it doesn't exist yet.

Step 3: Sort the equipment list into must-have, nice-to-have, and never

Take your equipment candidate list and sort every item into three buckets:

  • Must-have: the SMT production line cannot produce your product on time without it
  • Nice-to-have: it improves yield or changeover time, but day-one output does not depend on it
  • Never: it only serves products you will not run in the next three years

The "never" bucket is where the honest savings live. It's also where honest answers get uncomfortable: if you build standard-pitch consumer boards, and someone is pitching in-line X-ray capability for ultra-fine-pitch work, ask when that capability actually gets used. If the answer is "eventually," it is probably a depreciation item, not a capability.

This is also where knowing your own boundaries matters. If a specific process is outside your team's area of experience — say, a complex mixed-technology assembly or a finish you have not run before — bring in someone who does that process rather than letting a general equipment dealer talk you into a machine you cannot evaluate.

Step 4: Validate PCB assembly integration against a real board

Before you sign anything, take one real production board through the proposed SMT production line setup. Not a demo board. The actual board with the awkward connector, the shield frame, and whatever hard-won process you have been outsourcing.

So glad I insisted on this with one supplier. Their demo video was shot on a board where every joint was visible and easy to reach. The real board exposed a connector that needed a custom nozzle — and that nozzle had a two-week lead time nobody had counted on.

What to check:

  • Is there an existing validated placement program for your actual bill of materials on the exact machine models being quoted?
  • What is the real lead time for critical component feeders, nozzles, and long-tail hardware — not the machine, but the parts around it?
  • Who is the named field process engineer for first-article bring-up — the supplier's, or yours?

If a supplier cannot put a name and a handshake on the person who will be on-site for first power-up, treat that as a scheduling risk, not a detail.

Step 5: Put utilities on the critical path, not on the punch list

Power, compressed air, nitrogen, exhaust, temperature and humidity — this is where most of the disasters live. Utilities get treated as a punch-list item behind the "real" purchase of equipment, but the utility installations are frequently the thing that actually pushes the schedule.

Lay every utility on a single sheet:

  • Power demand per machine at steady state and at startup, against your available supply capacity
  • Compressed air consumption, dew point, and pipe sizing — the compressor is the most commonly missed item on the SMT production line equipment list
  • Nitrogen purity if you run lead-free or fine-pitch reflow
  • Exhaust routing for the reflow oven, and it must not share air with other production areas

Checkpoint: Do your utility delivery dates line up with each machine's installation date? If not, every later step in the sequence gets pushed.

Step 6: Negotiate the setup sequence, not just the price

A supplier who quotes you a machine price is not taking schedule risk. A supplier who understands the SMT production line setup sequence can organize delivery, power-up, and first-article validation into a real sequence with dates.

I said "turnkey." They heard "equipment delivered and powered on." We were using the same word but meaning different things. Discovered this when the compressor and exhaust were not in the contract scope, and we spent two extra weeks finding a local contractor.

Before signing, put the setup sequence into the contract as an attachment: who moves each machine in, what needs to be on-site the day each machine arrives, who handles commissioning per machine, and what acceptance check marks completion at every hand-off point.

Checkpoint: Can you describe what happens in the first 72 hours after the first machine arrives? If not, it isn't planned — the supplier's schedule is carrying the risk, and the supplier may not know it.

Step 7: Write acceptance tests into the contract

This step gets skipped most often, and it is the one people regret most. Define in writing what counts as "done." Not equipment checked, power on. Done.

Anchor the standard to something public. For solder joint acceptance, IPC-A-610 is the widely accepted reference — say which class you're accepting against, typically Class 2 for general product and Class 3 where higher reliability is required. For reflow profiles, let your process engineer define the profile window rather than letting equipment defaults pass unchallenged.

Add acceptance gates for progress, not just final sign-off: first-article pass, daily throughput test, changeover test on a mixed line, and three consecutive days of stable production. Each gate gets a signature. That way, when the day comes, there is no argument about whether it counts.

What tends to go wrong

A few things that fall through even after the seven steps — often because everyone assumes they were handled.

  • The "standard lead time" a supplier advertises often includes their own production buffer. Their buffer is not your setup window.
  • The first-power-up field engineer is often quietly agreed as "remote support." If the supplier cannot commit to on-site attendance, put it in a contract schedule.
  • Changeover time is rarely measured with a real product mix. Time it with your actual mix, three times in a day, not with a demo.
  • Spares — especially custom nozzles and feeder tooling that get purchased once — should be scoped together with the line ramp, not discovered after a failure.

If you have to shortcut somewhere under deadline pressure, don't shortcut Step 4 or Step 7. You can compromise elsewhere. Those two are the only leverage you will have later.