Production Volume Calculator: Do You Need a Pick And Place Machine?

“About 10,000 boards a month.”

I’ve heard that answer more times than I can count, usually five minutes before somebody asks which pick and place machine they should buy. The problem? Board volume, standing alone, tells me almost nothing about the placement workload hiding underneath.

Not nearly enough.

Ten thousand LED boards carrying 25 components each create 250,000 placements. Ten thousand industrial controller boards carrying 700 components each create seven million. Same board count. Completely different factory.

That’s why I don’t begin with boards per month, and I frankly don’t trust any supplier who does. I begin with verified placements per productive hour—after yield loss, feeder loading, product changeovers, first-article checks, nozzle swaps, maintenance, inspection and rework have taken their bite.

Because they will.

So, do you actually need a pick and place machine?

You probably do when your required placement volume has moved beyond what manual assembly or outsourcing can sustain, and when the resulting labor savings, repeatability, lead-time control and added capacity can repay the complete installed system—not just the mounter—within a period your business can tolerate.

That last bit matters.

The Production Volume Calculator

Before browsing machine catalogues, opening quotation spreadsheets or arguing about brands, run four numbers. It isn’t glamorous, but neither is explaining six months later why a supposedly “fast” machine can’t keep up with the real BOM.

Step 1: Calculate daily placements

Required daily placements = Good boards per day × Placements per board

Say the production plan calls for 500 good boards per day, with 180 placed components on each board:

500 × 180 = 90,000 good placements per day

Simple enough.

But count every physical placement. Don’t count only unique part numbers or feeder types. A board may use ten resistor values and require ten feeder positions, yet if it carries 160 resistors, the machine still has 160 resistor placements to make.

That’s the workload.

Step 2: Correct for yield

Here’s where optimistic spreadsheets begin lying.

Production demand normally means good, shippable boards. A line operating below 100% first-pass yield must process more than the sales requirement because some assemblies will need diagnosis, touch-up or complete rework.

Use:

Gross placement requirement = Required good placements ÷ First-pass yield

At 97% first-pass yield:

90,000 ÷ 0.97 = 92,784 placements per day

Small difference?

Not really. Leave yield out of a capacity calculation and a marginal machine suddenly looks acceptable on paper, right up until the backlog starts growing and Saturday becomes a normal production day.

Step 3: Calculate productive hours

An eight-hour shift isn’t eight hours of placement. Anybody who has run an SMT floor knows this, yet plenty of capital requests still pretend otherwise because eight divided by eight looks pleasantly clean.

Real shifts contain feeder loading, line clearance, stencil issues, first-off inspection, breaks, material shortages, barcode problems, nozzle cleaning, setup verification and the occasional board jam that somehow consumes 25 minutes.

So let’s assume an eight-hour shift provides 6.5 productive hours:

92,784 ÷ 6.5 = 14,274 effective placements per hour

That figure is already more useful than monthly board volume.

Much more useful.

Step 4: Apply realistic utilization

Now comes the brochure-speed trap.

A machine’s headline CPH is usually measured with a favorable component mix, short head travel, controlled feeder positions and very little of the awkward stuff that shows up in an actual production schedule. Throw in tray-fed QFNs, tall connectors, nozzle changes and mixed lots, and the glossy number starts leaking.

Fast.

At 55% utilization:

14,274 ÷ 0.55 = 25,953 rated CPH

Then add 20% for growth:

25,953 × 1.20 = 31,144 rated CPH

So the requirement isn’t merely “a machine that can build 500 boards.” It’s a configuration capable of roughly 31,000 rated CPH, assuming the utilization and growth inputs genuinely match your product mix.

That’s a different buying conversation.

Complete Formula

“`text id=”prgr1x” Required Rated CPH = ( Good Boards Per Day × Placements Per Board ) ÷ First-Pass Yield ÷ Net Production Hours ÷ Expected Machine Utilization × Growth Factor

Spreadsheet formula:

```text id="kt3mdo"
=(Boards_Per_Day*Placements_Per_Board)
 /First_Pass_Yield
 /Net_Hours
 /Utilization
 *Growth_Factor

Enter percentages as decimals:

• 97% yield = 0.97
• 55% utilization = 0.55
• 20% growth allowance = 1.20

Don’t run one scenario and call the job finished. Run several—especially the unpleasant one.

I wouldn’t approve a machine from the optimistic case alone. Frankly, that scenario often describes the factory everyone hopes to have after commissioning, not the one they’ll actually be running on a wet Wednesday with two missing reels and an engineer rewriting the placement program.

The conservative case hurts.

Good. It should.

Why Advertised CPH Is Not Real Throughput

CPH means components per hour. It doesn’t mean finished boards per hour, good boards per hour or—despite what some rushed sales presentations imply—money per hour.

Actual throughput drops when production includes:

• BGAs, QFNs or fine-pitch ICs
• Tray-fed or tall components
• Frequent nozzle changes
• Long feeder-to-board travel
• Limited feeder capacity
• Mixed product batches
• Warped panels or unstable board support

Consider the JUKI RS-1XL. JUKI publishes 42,000 CPH under optimum conditions but 29,000 CPH under IPC-9850 conditions. Same machine, different test reality—and a gap large enough to break a badly constructed capacity plan.

Panasonic lists up to 42,000 CPH for one NPM-W2 configuration, along with single- or dual-lane operation and as many as 120 component supply positions under the stated setup.

Fuji, at the faster end, reports up to 120,000 CPH for a particular NXTR high-speed configuration using 2RV modules, RH28 heads and four-panel processing.

Those numbers are legitimate within their stated conditions. But compare them carelessly and they become useless.

Here’s the ugly truth: if three suppliers are quoting against three different assumptions, you aren’t comparing machines. You’re comparing sales stories.

Send each supplier the same BOM, centroid file, panel drawing, package list, feeder requirements and production target. Ask for a simulated cycle-time report. Then ask which components, stoppages and setup activities were excluded.

The exclusions matter.

Why Manufacturers Are Automating

Labor pressure isn’t imaginary.

In February 2024, IPC reported that 59% of surveyed electronics manufacturers were facing higher labor costs, while 45% were dealing with higher material costs.

And automation investment hasn’t exactly slowed down. The International Federation of Robotics counted 4,281,585 industrial robots operating worldwide in 2023—up 10%—with 541,302 new installations during that year.

The U.S. Census Bureau also reported that robotics expenditure in semiconductor and electronic-component manufacturing increased from $579 million in 2020 to $1 billion in 2022.

Big movement.

But don’t misread it. Industry-wide automation spending doesn’t prove that your company should buy a machine next quarter, any more than rising warehouse construction proves you personally need a forklift.

Your numbers decide.

When Manual Assembly Still Makes Sense

Manual placement gets mocked too easily.

For prototypes, engineering builds and genuinely tiny production batches, it can still be the least bad option—especially when the design is changing every week and nobody knows whether the current enclosure will survive the next customer review.

Manual or semi-automatic production may remain practical when:

• Production is limited to prototypes
• Designs change frequently
• Monthly demand is irregular
• Boards contain few components
• Capital must be preserved
• Outsourcing already provides acceptable pricing and lead times

For an early-stage operation, a Prototypowa i małoseryjna linia SMT may make far more sense than a high-speed system carrying unused heads, feeder banks and depreciation.

Idle speed earns nothing.

And in high-mix work, a supposedly slower machine with painless program changeover and sensible feeder commonality can outperform a faster platform that spends half the shift being set up.

That happens often.

When Outsourcing Is Better

Outsourcing isn’t failure. Sometimes it’s simply the cheaper way to buy capability you don’t need to own every day.

But compare honestly.

The contractor’s unit price must be weighed against your complete risk-adjusted in-house cost, which includes far more than the number printed beside “pick and place machine” on a quotation.

An honest ownership calculation includes:

• Machine purchase and freight
• Installation and utilities
• Feeders, nozzles and tooling
• Software and training
• Maintenance and spare parts
• Printer, reflow oven, SPI and AOI
• Conveyors and ESD controls
• Inżynieria procesowa
• Scrap, rework and working capital

I’ve seen buyers obsess over the mounter while barely discussing solder paste printing or reflow profiling. That’s backwards. A placement machine can position every component perfectly and still feed a disaster if paste volume is unstable or the thermal recipe is wrong.

Fast defects.

Businesses bringing production in-house should therefore examine a complete turnkey SMT line, not a mounter in isolation.

Calculate the Real ROI

The ROI model should include the value created and the costs added. Obvious, yes—but many proposals mysteriously count every labor saving and almost none of the maintenance, finance or process-engineering burden.

Use:

“`text id=”g2p6b4″ Annual Net Benefit = Outsourcing Cost Avoided

• Labor and Overtime Saved
• Rework and Scrap Reduction
• Expediting Cost Avoided
• Margin From Added Capacity
• Operating and Maintenance Costs
• Financing Costs “`

Then:

“`text id=”2rg87f” Simple Payback Period = Total Installed Investment ÷ Annual Net Benefit

And:

```text id="mqtgdr"
Cost Per Good Board =
Total Annual Ownership Cost ÷ Annual Good Boards Produced

Suppose the installed project costs $350,000 and produces $140,000 in annual net benefit:

$350,000 ÷ $140,000 = 2.5 years

Looks healthy.

Now reduce output to 70% of forecast, perhaps because the contract launch slips, product approvals drag or utilization never reaches the promised figure. Annual benefit falls to $98,000:

$350,000 ÷ $98,000 = 3.57 years

Which payback should carry more weight?

The second. Every time.

Forecast demand is cheap to type into a spreadsheet. Missing it is expensive.

Production Bands for Initial Screening

These ranges are only a first-pass filter. They don’t replace a board-level cycle-time study, but they’ll stop you from shopping in a completely unrealistic equipment class.

For sustained high-volume demand, a high-speed mass-production SMT line may be appropriate.

May be.

Before signing anything, look at feeder commonality, product families, operator headcount, floor space, line redundancy and downstream capacity. One screaming-fast mounter feeding a slow oven—or stopping the whole line whenever it faults—isn’t a clever system design.

It’s a bottleneck with good marketing.

Signs You Are Buying Too Early

A sales forecast arrives. Everyone gets excited. Suddenly a machine purchase that wasn’t discussed last month is being treated as an operational emergency.

Slow down.

You may be buying too early when:

• The BOM or board design is still changing
• Current assembly cost is unknown
• No one owns the SMT process
• The mounter is faster than printing, reflow or inspection
• The purchase depends on one unconfirmed contract

And don’t confuse a machine operator with a process owner.

The operator loads feeders, clears alarms and keeps production moving. The process owner controls stencil design, paste handling, placement programming, thermal profiles, defect analysis, traceability and continuous improvement.

Different responsibilities.

Very different.

Signs You Are Waiting Too Long

The opposite mistake is easier to spot because it usually arrives wearing an overtime sheet.

You may have waited too long when:

• Overtime is becoming routine
• Outsourcing lead times are losing orders
• Manual placement errors are increasing
• Minimum order quantities create excess stock
• Engineering changes take too long
• Demand is supported by recurring orders
• Several products share components and feeders

At that stage, compare available pick and place machine configurations against the calculated capacity—not against whatever brand name happens to be familiar.

Brand loyalty is comforting.

Cycle time pays.

Published customer production cases may help you identify similar applications, but don’t let them replace a calculation based on your BOM, your panel and your shift pattern.

What to Send a Supplier

“Please quote a machine for medium-volume PCB production” isn’t a useful request. It’s an invitation to receive a generic quotation.

Send real files.

Provide:

1. Gerber files and panel drawing
2. Centroid or XY file
3. BOM with manufacturer part numbers
4. Package sizes and component heights
5. Placements per board
6. Tape, tray and tube requirements
7. Daily and monthly output targets
8. Changeover frequency
9. Shift pattern
10. Yield and traceability targets
11. Available floor space and utilities
12. Three-year growth forecast

Then request:

• Estimated cycle time
• Efektywne CPH
• Feeder and nozzle configuration
• Czas przełączenia
• Accuracy conditions
• Training and warranty scope
• Spare-parts recommendation
• Acceptance-test criteria

No acceptance criteria, no deal.

I mean that literally. The factory acceptance test should use representative boards, real component packages, agreed quality limits and a sustained production run—not a tray of identical chips flying onto a demonstration panel for three minutes.

That proves very little.

Często zadawane pytania

How many boards justify buying a pick and place machine?

A pick and place machine is justified when required good-board output, multiplied by placements per board and adjusted for yield, exceeds the sustainable capacity of manual production or outsourcing, while savings from labor, quality and shorter lead times repay the complete installed system within the company’s acceptable investment period.

There isn’t a universal board threshold. Anybody giving you one without asking for placements per board, product mix and shift availability is guessing.

How do I calculate PCB production volume?

PCB production volume is calculated by multiplying required good boards by placements per board, correcting the result for first-pass yield, dividing by productive operating hours and then adjusting for realistic machine utilization, changeovers, feeder replenishment, maintenance, inspection and expected production growth.

After that, convert the placement result back into boards per hour using the actual component count. Not the average from three unrelated products—the real board.

What CPH pick and place machine do I need?

The required machine CPH is the placement rate needed to meet good-board demand after accounting for yield, available production hours, downtime and realistic utilization, with additional capacity included for expected growth rather than relying on the machine’s highest advertised speed.

Run expected and conservative cases. I’d rather discover a shortfall in Excel than on a production floor with a delivery truck arriving Friday morning.

Do I need a complete SMT line?

A complete SMT line is a connected production system that normally includes solder paste printing, placement, reflow, inspection, conveyors, material handling and traceability, and it becomes necessary when any upstream or downstream process would otherwise limit the speed or quality of the pick and place machine.

The exact configuration depends on product complexity, compliance requirements and defect cost. A cheap consumer board and a safety-related industrial controller don’t carry the same inspection risk.

Is a used pick and place machine worth buying?

A used pick and place machine can be worthwhile when software, feeders, nozzles, spare parts, calibration support and service remain available, and when the seller can demonstrate acceptable placement accuracy and sustained output using components and boards similar to your actual production.

A bargain machine with locked software, worn feeders and obsolete controller boards isn’t a bargain. It’s a parts hunt.

Calculate Before You Automate

Don’t begin with:

“What is the best pick and place machine?”

Begin here:

What verified production capacity do we need, what equipment must surround it, and what return survives a conservative forecast?

That question is less exciting than comparing machine videos. It’s also far more likely to keep you from buying the wrong line.

Run the calculator. Prepare the BOM, centroid file and panel data. Push suppliers for a cycle-time analysis based on your actual product—not an idealized demo board.

And when the assumptions look too neat, challenge them.

For equipment selection, feeder planning and installed-cost estimates, contact an SMT production specialist.