The quote looked fine. The line didn't fit.
I'm the quality manager at a contract electronics manufacturer. Every purchase order crosses my desk before a supplier sees money — roughly 180 to 200 of them a year. Last year, on line-related projects, I kicked back about 40% of first-pass proposals. Not because of price. Because the proposal didn't hold up.
Most buyers approach SMT production line sourcing as an equipment purchase. List the pick-and-place, the printer, the reflow oven, the AOI. Get quotes from three sources. Compare unit prices per machine. Place the order. Wait for the container.
That's not sourcing. That's shopping.
And the gap between those two things is worth a quarter of your capacity.
What you're actually buying is a system
An SMT production line isn't a set of machines standing near each other. It's a system with a bottleneck, a material flow, and a set of physical constraints that don't show up on any spec sheet.
The bottleneck sets your output. Full stop. A 60,000 CPH placement machine sitting behind a printer with a 25-second cycle is a 25-second line. The buyer sees 60,000 on the datasheet. They get 25 seconds on the floor.
Most buyers focus on the placement machine's rated speed and completely miss the conveyor heights, the buffer conveyors, and the loaders — which together can run 15% to 20% of total line cost and 100% of your line's ability to actually flow.
Here's the question everyone asks: "How much per machine?" The question they should ask: "What's the cycle time of the slowest station, and who owns that number?"
The deeper problem: layout gets treated as an afterthought
This is where it usually goes sideways. The equipment list gets built first. Then somebody draws a floor plan around whatever was ordered. Then the electrical and air drops get routed.
That sequence is backwards. And it's expensive.
Layout isn't about fitting machines into a rectangle. The rectangle is the easy part — it's the least interesting constraint in the whole project. What actually decides whether your line works:
- Floor loading and vibration isolation — some placement machines care, some don't, and the ones that do will tell you after installation
- Compressed air and vacuum routing, sized for the whole line rather than per machine
- Reflow exhaust and fume extraction, which usually needs more ceiling space than anyone planned for
- Power distribution at line level, which is a genuinely different problem from power at machine level
- Solder paste and stencil storage within a realistic walking distance of the printer
None of that is exotic. All of it gets skipped, routinely. (This was back in 2024, and I don't think the pattern has changed.)
Here's something vendors won't tell you: on a "turnkey" line quote, the peripheral equipment is often sourced after the contract is signed. The core machines get locked first because that's where the margin lives. The loaders, conveyors, and transfer units get bought last. By then you've lost your leverage, and the timeline has quietly absorbed an extra three to five weeks that nobody wrote down.
What it costs when you get this wrong
Not the equipment price. That part is usually fine — it's negotiated, it's documented, it's a number both sides agreed to.
The cost is time. And time compounds.
We had a project in 2023 where the equipment got specified first and the layout drawing came second. When the machines landed, the infeed conveyor sat 12 mm off the printer's load height. Sounds trivial. A shim job, right?
It wasn't. The conveyor height fed back into the electrical routing, which fed back into the exhaust ducting. Re-routing the ducting meant taking an adjacent line down for four days. Four days of an adjacent line down is not a line item on any quote.
Realistic damage from a layout redo at that stage: four to eight weeks of schedule. Power, air, and exhaust all re-run. Re-permitting, if you're somewhere that cares. And you're burning your launch window the entire time.
I went back and forth for about three weeks on our own line — cheaper imported equipment with our team handling integration, versus paying a premium to a provider that owned the whole scope. On paper, self-integrating saved 22%. My gut said we'd be the ones holding the bag when the conveyor heights didn't line up and nobody's name was on that part of the drawing.
We paid the premium. Honestly? Best money we spent that year.
Here's the piece that never gets priced in: consistency. A line that runs is not the same as a line that runs the same way on Tuesday as it did on Monday. That's the thing your customer actually notices. Your customer doesn't see your placement rate. They see whether the third shipment looks like the first one.
On acceptance, we hold a new line to a process capability index of 1.33 on critical characteristics — the conventional threshold where a process is generally considered capable. Note that it's a line-level number, not a machine-level number. You cannot buy it by upgrading one component. Published standards exist for the components — IPC-A-610 and J-STD-001 cover acceptability and soldering requirements, IPC-9850 addresses placement equipment performance parameters — but no standard tells you what your line does. Only your line tells you that.
Once the line is on your floor, every layout decision pays rent. In downtime. In rework. In overtime you pull out of a budget you'd rather not touch. Before that, they're just lines on a drawing.
So what actually works
Short version, because the problem is the point here, not the fix:
- Lock the layout before the RFQ goes out. Not a sketch. A real drawing with utilities, clearances, and traffic paths. If a supplier can't work from it, that tells you something.
- Specify the peripherals in the same document as the core machines. Conveyors, loaders, buffers, transfer units. Same scope, same quote, same timeline. No separate purchase order later.
- Ask for the engineering package, not just the equipment list. Who signs off on the utility interfaces? If the answer is "we'll figure it out on site," you've found your deal-breaker.
- Write acceptance criteria at line level. Throughput, first-pass yield, changeover time. Machine specs don't tell you whether the line works.
None of this is complicated. It's just usually skipped, because it all happens before there's a purchase order to argue about.
Equipment pricing is negotiable. Your launch window is not.