The short version

If you only read one paragraph: validate the machine before you negotiate the price. Buying the wrong SMT pick and place machine costs far more than buying an overpriced one. With an overpriced machine you know exactly how much you overpaid. With a wrong machine, you find out 6–12 months later, when it's already sitting on your line, producing defects and eating your team's patience.

My own version of this lesson shipped in September 2022. A used automatic pick and place machine, purchased for roughly $11,400—or rather, closer to $12,600 once you count crating and inbound freight. The feeder interface didn't match our existing reel inventory. We found out on install day. That single omission cost us a week of production delay, about $890 in expedited feeder adapters, and a level of trust with our production manager that took a full quarter to rebuild.

Everything below is the 4-point pre-purchase checklist I now enforce on every SMT pick and place machine sourcing order. It is not elegant. It is not exhaustive. But since Q1 2024 it has caught 47 potential errors before signature—some small, a few that I genuinely believe would have ended contracts.

Why you should trust this list (or, where it came from)

I've been handling SMT and PCB equipment sourcing orders for 9 years. In my first year (2017) I made the classic error: I chose a machine because the price looked right and the vendor promised same-week delivery. Support contract? Never asked. Spare-parts lead time? Never checked. When the first vacuum nozzle failed 4 months in, the replacement took 6 weeks and we were running the line on borrowed nozzles from a neighboring shop (which, honestly, is not a network you want to depend on).

The 2022 feeder mistake was the one that finally made me build a formal checklist. The 2024 revisions came after our third rejection of a machine that looked great on paper but failed the placement-accuracy check against IPC-9850 in our own incoming inspection.

What I'm sharing is not industry consensus. It's my team's working list, maintained since 2017, refined roughly twice a year. I don't have hard data on how many sourcing errors the broader industry makes—I doubt anyone does. What I can say anecdotally is that roughly 8 out of 10 of the "bad purchases" I've seen discussed among peers trace back to one of these 4 checks being skipped.

Check #1 — Feeder compatibility, not just "works with standard feeders"

This is the one that cost me $11,400, so we start here. "Standard feeders" is the marketing phrase you will hear from nearly every sourcing agent. The operational phrase is different: does this machine physically accept the reel widths and feeder types your production line already uses?

Ask three specific questions before you commit:

  • What feeder types does the machine accept natively—tape, stick, tray, matrix tray, bulk?
  • Does the machine's feeder bank require a proprietary feeder that only the OEM sells, or can it accept generic feeders of the same width?
  • Which reel widths are supported—8mm, 12mm, 16mm, 24mm, 32mm? Every width your line runs, or just the popular ones?

Get photos of the actual feeder bank on the actual unit being quoted. Not the catalog. Not the previous model. This unit. I cannot overstate how many quotes show stock photography of a different configuration. Verify on the specific machine.

Check #2 — Placement accuracy specs against IPC-9850

Every SMD pick and place machine vendor will give you a placement accuracy number. Most of them are accurate for the conditions the vendor chose to measure under, which may not resemble your board. The industry has a standard for characterizing this—IPC-9850, "Characterization of Surface Mount Placement Equipment," maintained by IPC (ipc.org).

You don't need to run the whole IPC-9850 protocol before buying. But you do need to know what spec the vendor's number was measured under:

If the advertised accuracy is "±50µm at 3 sigma" but that was measured on a small, flat, well-anchored board at moderately slow speed, your actual placement on a dense, large, thin-core PCB could be 2–3× that error. The number isn't wrong. It's contextual.

Ask the vendor to share the test conditions alongside the accuracy claim. If they can't, or won't, treat the spec as a marketing figure and plan accordingly—because production engineers on your side will find out the real number within the first month.

Check #3 — Support and spare-parts geography

This is the check that most sourcing conversations skip entirely, and it's the one that will define whether the machine is an asset or a liability in year two.

Three questions:

  1. Where is the nearest authorized service engineer? If it's outside your country and requires a visa and flights every time a calibration drifts, that changes the total cost of ownership substantially.
  2. What is the typical lead time on the 5 most frequently replaced parts—nozzles, belts, sensors, vacuum pumps, feeder motors? A machine is not usable if one nozzle is unavailable for 6 weeks.
  3. Who supports it after the vendor's warranty period? Is it the manufacturer, a third-party service network, or "call us"?

I once ordered 40 feeder units—checked it myself, approved it, processed it—and only caught the support gap when a calibration error appeared in the third month. The vendor's response was pleasant and honest: "We can remote in but a technician visit takes about 5 weeks." $3,000 in rework, one delayed client delivery. That's when I added "support geography" to the checklist as a mandatory field, not a nice-to-have.

Check #4 — Total cost of ownership vs. sticker price

On pick and place machine price, there's plenty of range to work with. Based on quotes we collected across Q4 2024 and Q1 2025, used automatic pick and place machines from the major Japanese, Korean, and European manufacturers typically land somewhere in the $15,000–$65,000 bracket depending on year, model, and feeder count; new entry-level automatic units tend to start in the $30,000–$55,000 range. Treat those as bracket estimates for sourcing conversations, not a price list—the used market moves month to month, and feeder configuration can swing a quote by 30%. Verify current pricing directly with suppliers before quoting internally.

The point of Check #4 is this: sticker price is roughly 40–60% of the 3-year cost of a used machine. The rest sits in feeders, spare parts, calibration, consumables, training, and downtime risk. Ask for the feeder count separately from the machine price. Ask what the typical annual maintenance cost has been on the fleet for this model. Ask other buyers, if you can reach them through the vendor's references—not the references the vendor chose for you, but genuine peer shops.

The upside of a lower sticker price is real cash. The risk is that a $3,000 saving on price turns into $9,000 in catch-up costs. I have stopped pretending I know which way the tradeoff goes without this math. Some deals are worth it. Most of the "too good to be true" ones aren't.

What this checklist does not cover

Three honest boundaries, because a checklist that pretends to be universal is worse than no checklist at all.

First: prototype and low-volume work changes the math. If you're placing 200–500 components a day on prototypes, none of the feeder economics above matters the way it does on a production line. A manual or semi-automatic stage may honestly serve you better at that volume. I don't say that as an attack on pick and place machines—I say it because I once pushed an automatic machine into a shop that was still doing prototype work, and it sat idle 70% of the month.

Second: fully authorized, brand-new purchases with a full OEM warranty are a different risk profile. If you're buying new from an authorized dealer with an on-site service contract, Checks #2 and #3 are largely covered by the OEM's own warranty terms. You don't need to redo the vendor's job. But you still want to confirm them in writing—the warranty isn't the same as a service agreement, and it's worth reading the fine print.

Third: my list is one team's experience. It's tuned for shops of roughly 10–40 production staff doing mid-volume SMT work. If you're a larger shop with in-house engineering that can characterize a machine on arrival, you probably have a better version of this list than mine.

Where to actually start

If I could go back to 2017 and only fix one thing, it wouldn't be the support question. It would be this: 5 minutes of verification beats 5 days of correction. Every time. Every single time I've skipped a check "because this vendor seems solid," I've paid for it later in some form—money, time, or the particularly annoying form of both at once.

Print the 4 checks. Tape them to your sourcing desk. Fill them out before signing. Done.