0201 Placement at 80k CPH — What Changes on the Line and in the Design

An 0201 is 0.6mm × 0.3mm. On a placement machine running at 80,000 components per hour, the nozzle visits 22 components every second. The acceleration profile of the head means the part experiences forces above 8G between pick and place. At that speed, every margin in your design, your stencil, and your line setup either holds or shows up as a yield number that ruins the cost model.

What looks fine on a prototype line at 15,000 CPH starts breaking at production speed in these ways:

• Tombstoning climbs from background to 0.05–0.15% on poorly designed pads. At 80k CPH and 10,000 0201s per board, that's 5–15 defects per board.
• Skewed placement increases because the nozzle is moving faster, the vision system has less time to centre, and any tape pocket variability matters more.
• Solder bridging between adjacent 0201s gets worse because the stencil release becomes less consistent at the smaller aperture sizes the line is now running.
• Component-to-board offset shows up in AOI as marginal pass/fail decisions — you're inside spec, but only just.

"At 80k CPH, the line is unforgiving. Every design margin you took at prototype becomes a yield number at volume." — Pioneer Horizon SMT line lead

This article is the conversation we have with customers before they commit to high-density consumer or medical boards that drop into 0201 territory. The trade-offs are real and the design decisions cascade — get them wrong on the schematic and the line can't bail you out.

The IPC-7351B nominal land pattern for an 0201 is 0.30mm × 0.30mm pad with 0.15mm spacing. That's the textbook starting point and almost nobody runs exactly that in production. Here's what we run on the Madurai line and why.

Pad dimensions

• Pad length: 0.30mm — IPC nominal. We sometimes trim to 0.28mm on density-constrained boards but the wetting margin drops.
• Pad width: 0.30mm — matches the 0201 termination width. Going wider than 0.32mm increases bridging risk; going narrower than 0.27mm risks insufficient wetting.
• Pad spacing: 0.15mm — IPC nominal. We push to 0.20mm for any board we expect to run at 80k+ CPH because the wider gap forgives placement skew.

Paste aperture sizing

For 0201s on a 100µm stencil, we run apertures at 0.27mm × 0.30mm — slightly narrower than the pad, full-length. This is roughly 90% pad coverage and it consistently delivers wetting without bridging at production speed. Going to 100% coverage at 80k CPH produces bridging at a rate of 0.02–0.04% — small absolute number but it's the dominant rework driver on 0201-heavy boards.

Stencil considerations

• Thickness: 100µm for any 0201-bearing area. A step stencil down to 100µm is mandatory on boards that mix 0201s with thermal-pad BGAs.
• Aperture wall finish: electroformed or laser-cut with electropolish. Standard laser-cut walls have enough surface texture to hang on to paste in the smallest apertures.
• Aperture corner radius: 0.05mm minimum. Sharp corners lock paste in the stencil and release inconsistently.

Solder mask

Non-solder-mask-defined (NSMD) pads for 0201s. The mask opening should be 0.05mm larger than the copper on each side. SMD pads concentrate stress at the mask edge and shorten thermal-cycle life by an order of magnitude on these tiny terminations.

The aperture and mask numbers above are not theoretical — they are what we run today on a consumer audio programme delivering 12,000 0201s per board at 1.8% reject rate. Earlier rev of the same board ran 5.4% with textbook nominal apertures.

An 0201 tombstones when one termination wets and the other hasn't yet. The wetted end lifts off the pad before the second end can catch hold. At production speed, three forces decide whether tombstoning happens:

Asymmetric paste deposit

If the stencil deposits 5% more paste on one pad than the other, the heavier pad heats slightly faster and wets first. Stencil aperture symmetry matters here — we measure paste-deposit volume by 3D SPI on every panel and reject any panel with >15% asymmetry between the two pads of a 0201 pair.

Asymmetric thermal mass

If one pad is connected to a copper pour and the other isn't, the unconnected pad heats faster and wets first. The fix is thermal symmetry in the layout — either both pads connect to copper or neither does. For 0201s sitting on a power plane, this is the single biggest tombstone driver.

Asymmetric routing entry

A 0201 with a thick trace entering one pad and a thin trace entering the other has built-in asymmetry. Use thermal reliefs or match trace widths within 2:1.

Profile contribution

Faster ramp rates through the soak amplify any of the above. Reducing ramp rate from 2.5°C/s to 1.5°C/s through soak typically cuts tombstoning by 40–60% on a marginal board. It won't fix a board that's structurally asymmetric — but it'll buy enough margin to ship.

What good looks like

On our well-engineered 0201 programmes we hold tombstoning to 0.01% or below — one part per 10,000. On a board with 10,000 0201s that's one rework event per board, which is acceptable. On marginal designs we've measured 0.2%, which is 20 events per board — a programme killer at any non-trivial volume.

The fix order is: layout symmetry first (free), then aperture symmetry (stencil retool — cheap), then profile (validation cost — moderate), then component substitution as a last resort. Most boards we audit have all three of the cheap fixes available and unused.

Density is what drives the decision to go 0201 in the first place — a 0201 footprint is roughly 35% of a 0402 footprint. But the routing and placement reality is that you don't get all of that density back unless your courtyard discipline is right.

The IPC courtyard

IPC-7351B's "least" courtyard excess for 0201s is 0.1mm. The "nominal" is 0.15mm. The "most" is 0.20mm. We default to nominal — 0.15mm — for any board running at 80k CPH or above. The 0.1mm courtyard is feasible on a slower line but adds rework risk at production speed because the placement-skew tolerance no longer fits inside the courtyard.

What courtyard discipline buys you

• AOI clearance. AOI algorithms need adjacent-component clearance to image clearly. Components inside each other's courtyards trigger false fails.
• Rework access. A bridging defect between two 0201s with 0.1mm of clearance can be reworked. A defect between two 0201s with 0.05mm clearance often takes both parts off — and now you're rebuilding rather than reworking.
• Repeatable solder reflow. Adjacent components disturb the thermal field around each other. Tight courtyards produce more tombstoning, especially when one neighbour has high thermal mass.

Trace routing inside the courtyard

Don't route signal traces between two 0201s at minimum courtyard. The trace can pull solder during reflow (solder thieving), creating opens. If you must route through, do it on an inner layer with a via on each side. Surface traces between 0201s at 80k CPH are a yield problem looking for a place to land.

Polarity marks

Diodes and tantalum caps in 0201 size need polarity markings on copper or in silk. Silk on 0201 territory is generally illegible — go to white-on-copper or use a polarity dot in the silkscreen at least 0.5mm away from the part. Vision systems can be set to look for polarity by die orientation, but only if the part vendor's die orientation is consistent across reels.

For a deeper walk-through of how courtyards interact with the rest of the panel constraints, see our stencil aperture and density guide.

The case for 0201 is density. The case against 0201 is yield, cost-per-component, and design-engineering time. Here's how we frame the trade-off when customers ask whether they should commit.

Component cost

An 0201 R/C is roughly 1.4–1.8× the cost of a 0402 equivalent at any meaningful volume. On a board with 5,000 R/Cs, that's a real BOM line item — typically $0.30–0.80 per board, depending on stock and grade.

Placement cost

Time per placement is essentially the same at 0201 as at 0402 on a modern machine — both run at the head's max speed. So placement-cost-per-part is flat. The placement-cost-per-board, on the other hand, goes up if 0201 forces a slower head profile to maintain accuracy.

Yield cost

• Well-engineered 0201 board: 1.5–2.0% first-pass defect rate on the 0201 population, all causes. Rework time per board: 3–6 minutes.
• Well-engineered 0402 board: 0.3–0.5% first-pass defect rate on the 0402 population, all causes. Rework time per board: under 1 minute.

The 0201 yield penalty is real — typically 1.5–2 absolute percentage points on the chip-component population. On a board with 5,000 chip components, that's 75–100 additional rework events per board. At ₹15–30 per rework event labour cost on our line, that's ₹1,100–3,000 per board.

When 0201 is the right call

1. Consumer wearables and handhelds where board area is the binding constraint and unit volumes justify the engineering investment in pad and stencil optimisation.
2. Medical implants and hearing aids where 0402 simply doesn't fit and density isn't optional.
3. High-end smartphones and tablets where the cost of one square millimetre of board real estate exceeds the cost of the 0201 yield penalty.

When 0201 is the wrong call

1. Industrial control boards where space isn't constrained and reliability is paramount.
2. Automotive boards where the thermal-cycling fatigue life of 0201s under hood is not yet at parity with 0402s.
3. Low-volume programmes under 5,000 units/year where the engineering investment in 0201 optimisation can't be amortised.

"0201 is not a default. It's an answer to a specific density question. If you can't articulate the density question, stay at 0402." — Pioneer Horizon DFM lead

If you're weighing the decision for your next revision, talk to us before tapeout — a thirty-minute review of your placement density and BOM will tell you whether 0201 pays back.