Who this checklist is for

This is for the person who's been handed a capital budget and a deadline to stand up an SMT line — not someone who's already built ten of them. My experience is based on roughly a dozen line buildouts over the past seven years, mostly 100–250-person EMS operations in North America. If you're a large OEM running a multi-site rollout, your situation is different, and you probably already have a facilities team that owns half of what's below.

The list has six steps. Do them in order. Step 2 is the one most sourcing teams treat as a formality, and it's the one that costs the most to fix later.

Step 1: Define throughput and product mix before you talk to anyone

The first vendor email I ever sent asked for "a line that can do about 20,000 CPH." That number was made up. It came from a competitor's spec sheet, not from our actual product mix. We got three quotes back that were all over the map, and I couldn't tell which one was right because I hadn't defined what "right" meant.

Before you send anything, put four numbers on paper:

  1. Mix-weighted CPH. Not peak CPH on the fastest product — the weighted average across everything you actually plan to run.
  2. Board dimensions and panelization. The longest board you'll run determines conveyor length, not the shortest.
  3. Placement density and component types. A board with 0402s and fine-pitch QFNs needs different feeder and vision specs than one with 0805s and SOICs.
  4. Volume forecast for years 1–3. This is what justifies the difference between a flexible line and a dedicated one.

Checkpoint: Can you explain why you picked each number, and what happens if you're 30% off? If not, you're not ready to request quotes.

Step 2: Draft the line layout as a constraint, not an afterthought

This is the step people skip. I used to think layout was something the integrator handled — just hand over the room dimensions and they'll figure it out. That's wrong, and it's expensive to fix.

Layout is where you discover that the floor load rating won't support a 2.5-ton reflow oven, that your compressed air drop is on the wrong wall, that ceiling height rules out a taller reflow, or that your only power panel doesn't have enough spare breakers. None of those are equipment problems. They are building problems, and they cost real money.

If you have any flexibility at all, reference IPC-2221 and IPC-7351 for the design-side constraints, then work backward to the physical line. The layout has to account for:

  • Service clearance around every machine (front and rear, and it's more than you think)
  • Conveyor direction and the flow of panels through print → place → reflow → AOI
  • ESD flooring and grounding per ANSI/ESD S20.20 if you're running sensitive components
  • Power, air, exhaust, and network drops — all of them, at the right voltage and pressure

Checkpoint: Can a rigging crew move the largest piece of equipment into position without disassembling a wall? If you're not sure, measure again.

Step 3: Convert the layout into an equipment spec sheet

Once the layout is fixed, the equipment list becomes almost mechanical. You know how many feeders you need, what vision system the mix requires, and how much conveyor the room supports. Now you can write a spec sheet that vendors can quote against.

For placement equipment, IPC-9850 defines the measurement method for CPH and placement accuracy. Ask vendors to report against IPC-9850, not against a marketing number. Most will — the ones who won't are telling you something.

The spec sheet should cover, at minimum:

  • Placement machines: gantry vs. turret, feeder count, vision capability, IPC-9850 test data
  • Screen printer: stencil size, squeegee type, aperture capability
  • Reflow oven: zones, profile capability, belt width, temperature uniformity spec
  • AOI/SPI: resolution, false-call rate, integration with your MES
  • Feeder inventory, spare parts list, and consumables (nozzles, filters, paste)

Step 4: Collect quotes, then rebuild them as TCO

This is where I've seen the most money lost. Vendors quote different scopes. One includes installation; another includes rigging but not calibration; a third "throws in" training that's really a two-day sales visit. If you compare line items, you'll pick the wrong one.

My rule: never compare quotes until you've rewritten every one of them into a single TCO template. Ours has eleven lines, and it's ugly, but it works:

  1. Equipment purchase price
  2. Rigging and moving
  3. Installation and calibration
  4. Initial feeder and nozzle inventory
  5. First-year spare parts package
  6. Training (hours, on-site vs. remote, per shift)
  7. Programming software and licenses
  8. MES / network integration
  9. Warranty coverage and what voids it
  10. Estimated energy and consumables, year 1
  11. Payment terms and currency exposure

Take this with a grain of salt — our line items won't match yours exactly. But the principle holds: if a quote doesn't have a number for a line, ask why. "Included" is not a number.

Checkpoint: Can you sort the quotes by total three-year cost and explain the gap between the cheapest and the most expensive? If the gap is bigger than 15%, something in the scope is missing.

Step 5: Lock down setup, facilities, and acceptance criteria

Setup isn't a phase that happens after purchase. It's a contract term. Put the acceptance criteria in writing before you sign, and tie final payment to passing them.

The acceptance test should follow IPC-A-610 for the assemblies you're actually producing. Don't accept a demo board. Run your board, at your volume, for a defined window — I'd suggest at least one full shift — before you release the last payment.

Facilities and setup items that always seem to slip:

  • Floor prep and anchor bolts
  • Electrical: voltage, phase, breaker size, and who pulls the wire
  • Compressed air: pressure, dew point, and who runs the drop
  • Exhaust ducting for the reflow oven
  • ESD flooring and wrist-strap stations per ANSI/ESD S20.20

Step 6: Write the after-sales terms into the contract

The last step is the one that determines whether year two looks anything like the quote. Response time, spare parts availability, remote diagnostics, software updates — all of it belongs in the contract, not in a nice email.

Ask for these in writing:

  • Response-time SLA for on-site support (4 hours, next business day, whatever — just write it down)
  • Guaranteed spare parts availability for a defined number of years
  • Remote diagnostics access and what it covers
  • Software / firmware update policy, and whether it's free during warranty
  • Escalation path when the first-line tech can't fix it

Common mistakes I've watched people make

Negotiating on day one. The first quote is a scope document, not a price. Negotiate after the scope gaps are closed, or you're just trading unknowns.

Treating layout as the vendor's problem. It's your building. It's your problem.

Comparing unit prices across different scopes. It's tempting to think you can just line up the machine prices. But identical specs from different integrators can produce wildly different total outcomes once rigging, calibration, and year-one spares are counted.

Believing "no downtime" claims. The "guaranteed 100% uptime" pitch comes from an era when vendors controlled the maintenance contract and the reporting. Today, your own OEE data will tell you what's actually happening. Trust your numbers, not theirs.

My experience is based on North American EMS operations in the 100–250 person range. If you're sourcing a line for a captive operation in a different region, your vendor landscape and infrastructure constraints will look different — some of this will translate, and some of it won't.