This checklist is for anyone sourcing an SMT pick and place machine — new, used, or refurbished — where a wrong call costs six figures and a month of downtime. I'm the quality and brand compliance manager at a contract electronics manufacturer. My job is to review every piece of capital equipment before it's signed off for the floor: roughly 35–40 machines, fixtures, and major subassemblies a year, across three SMT lines. In 2023 I rejected 22% of first-article equipment deliveries. Most of those rejections were visible on the quote and the spec sheet weeks before anything showed up at the dock.

Eight steps, in the order I actually run them. The last two are the ones people skip, and they're the ones that decide whether the machine is still productive in year four.

The 8-step sourcing checklist

1. Write your board profile before you look at a single machine

Start from the board, not the brochure. I keep a one-page profile per product family: largest and smallest panel, smallest chip component we actually place (0402 today, 0201 on two assemblies), tallest component, heaviest board, panel dimensions after routing, annual volume, and how often we change over. Then I add a three-year outlook, because the machine you buy for today's mix is the machine you're stuck with when the mix moves.

That page does two jobs. It filters the market down to machines that can physically do the work, and it gives you something to hold a supplier to. “It'll handle your boards” isn't a spec. “Panel 460 × 460 mm, 0.8 mm thickness, 0201 minimum, 25 mm maximum component height” is.

2. Tell the difference between a placement rate and a marketing number

IPC-9850 is the standard for characterizing placement equipment performance, and the useful thing about it is that it forces the test conditions into the open: component mix, feeders mounted, vision on or off, board size, whether board transfer time is included. A chipshooter figure of 60,000 CPH with vision off and four feeders running tells you almost nothing about a real line with 90 feeders and a mixed BOM.

So ask for the number that matters. Measured throughput on a component mix close to yours, over a full shift, on a machine that's been installed at a customer for at least a year. If the supplier can't produce that, they can produce a demo. Those are different things.

3. Price the feeders and nozzles before you price the machine

This step surprises people. On the used and refurbished quotes I've reviewed, the feeder set and nozzle kit together have run anywhere from a quarter to nearly half of the quoted machine price. That's my own sample of quotes, not an industry benchmark — treat it as a reason to ask the question, not as a number to plan against.

Ask specifically: how many feeder slots does the machine have, how many feeders are included, which tape widths, electric or pneumatic, and whether the nozzles are a complete set or just the ones currently on the head. A machine with 30 slots and no feeders is a machine you can't run.

4. Build the landed-cost sheet, line by line

The quote is not the cost. I make suppliers break it out: machine, feeders, nozzle set, spare parts kit, freight, rigging and crane, installation, calibration, training, electrical and air drops on our side, software licenses, warranty, and re-validation of the line after install. Then I add our internal hours for the qualification run. If you're working through a sourcing partner, ask for the same format from them — a partner who can't separate rigging from calibration is handing you the ambiguity back.

Price reference: based on publicly listed prices on used-equipment marketplaces in Q1 2026, working used SMD pick and place machines with a basic feeder set commonly list in the US$15,000–60,000 range, while current-generation automatic pick and place machines are listed well above US$100,000 depending on configuration. These are asking prices, not transaction prices, and they exclude freight, rigging, feeders, and installation. Verify current numbers before you budget anything.

Speed, accuracy, support. Pick two — that's the honest version of the tradeoff. What I won't accept is a single-line quote that hides which two we're picking.

5. Get a demo on your board, or don't book the demo

Demo boards are designed to win demos. Bring your hardest assembly — the one with the fine-pitch connector and the part that likes to tombstone. If you can't travel, send the Gerber and the BOM and ask for a recorded run with your component mix plus the placement data log.

I have mixed feelings about demo visits. On one hand, half of them are theater. On the other, two hours standing next to a machine while an applications engineer loads feeders tells you more about how a supplier operates than a week of email. So I still go. I just don't decide anything in the room.

6. Audit the utilities and the floor, not just the footprint

Phase, voltage, kVA, compressed air pressure and flow and dew point, floor loading, leveling, ceiling height for the enclosure, vibration from neighboring equipment, network drops for the line software, ambient temperature. Send the supplier your site drawing and ask them to confirm in writing. Then have your facilities engineer confirm it back. “It'll fit” isn't a confirmation.

I once watched a machine sit on a dock for eleven days because nobody checked the door width. Not glamorous. Expensive.

7. Call three references running the same model, installed 18 months or longer

Not the references the supplier picks alone — give them the model and the year of install, and let them find three. Then ask about actual uptime, mean time to repair, lead time on the ten wear parts they've really had to replace, how many service visits came in the first year, and which component they wish the machine could place. That last question gets the most honest answer of the whole call.

The numbers said one thing on our last line upgrade: two quotes, comparable specs, one about 15% lower on landed cost. My gut said stay with the higher quote. I went with the gut, and roughly seven months later I found out the cheaper supplier's spare parts coverage in our region was effectively one person with a van. That's a story, not a data point — but it's why reference calls are step 7 and not step 10.

8. Negotiate the support envelope, not just the number

Price is the easiest thing to move and the least useful thing to win. What I want in writing: how many years spare parts are guaranteed available, what response time applies and what starts the clock, how many training seats and whether they're on-site or at the supplier's facility, exactly what's covered and excluded under warranty (nozzles? feeders? belts?), software update terms, and what happens if the machine can't be brought to spec after install.

Refurbished machines need this section doubled. “Refurbished” isn't a standard — it's a word. I've seen it mean a full teardown with new belts, bearings, and a calibration certificate, and I've seen it mean a wash and a power-on test. Ask for the work scope in writing, and ask which wear parts were replaced versus inspected.

Four mistakes I see most often

Buying the machine instead of the capability

Throughput on a datasheet is not throughput on your floor, and one very fast machine is not always the cheapest way to hit your volume. Two mid-speed machines occasionally beat one high-speed machine on total cost, because redundancy has a price and line-down time has a bigger one. Run the arithmetic on your actual mix and your actual downtime exposure.

Treating “fully compatible with your line” as a given

Nobody can guarantee that from a spec sheet, and anyone who does is telling you something about how they sell. Compatibility is a list of specific checks: conveyor height and SMEMA interface, board edge clearance, panel handling, upstream and downstream buffers, line software protocol, and how the machine behaves when the printer stops. Get those as line items, not as an adjective.

Skipping the qualification run

Install day is not acceptance day. Plan a first-article run against the IPC-A-610 class your customer requires — Class 2 for much of general product, Class 3 if that's what's in the contract — and write the acceptance criteria into the PO before anyone signs. If you plan to run high-mix production next year, qualify on a high-mix sample too, not just the easy board.

Not asking the supplier what they don't do

This is the one I'd put first if I could. The sourcing partner who told me “specialty nozzle work isn't our strength, here's who does it better” earned my trust on everything else, and I've since bought twice from them. A supplier claiming to cover the whole scope — sourcing, calibration, spares, integration, training — is either genuinely large or bluffing. Ask the question directly: what part of this do you hand off? The answer tells you where the risk sits before it becomes your problem instead of theirs.