IPC Class 2 vs Class 3: Which PCB Quality Class Do You Need?

Somewhere in every fabrication note is a line that quietly sets your board's cost and reliability: the IPC class. Pick Class 3 when you didn't need it and you've overpaid; pick Class 2 for a board that flies or keeps someone alive and you've under-built it. Here's what the classes actually mean and how to choose.

What IPC classes are

The IPC standards — IPC-A-600 (acceptability) and IPC-6012 (performance) — sort boards into three classes by how demanding their end use is. The class isn't a grade of "better" or "worse"; it's a statement of how much margin the product needs.

Class	Intended for	Reliability demand
Class 1	General electronics — toys, throwaway devices	Function is enough; short life acceptable
Class 2	Dedicated service electronics — most industrial, consumer, commercial	Continued performance, extended life; occasional failure tolerable
Class 3	High-reliability — aerospace, medical, defense, automotive safety	Must not fail; downtime or failure is unacceptable

Class 2 covers the large majority of commercial hardware. Class 3 is the high-reliability tier.

Where Class 3 is actually stricter

The classes diverge on the details that decide whether a board survives stress, thermal cycling and time. The big ones:

Annular ring — the copper collar around a drilled hole. Class 2 allows minimum ring with some tolerance; Class 3 demands more copper and far less breakout, because a thin ring is where barrels crack. See annular ring.
Plating thickness — Class 3 requires more copper in the hole barrel for thermal-cycling survival.
Acceptance criteria — voids, nicks, and defects tolerated at Class 2 are rejected at Class 3. Inspection is tighter and more of it is mandatory.
Coupons and reporting — Class 3 typically requires microsection coupons and documented test data on every panel; Class 2 often doesn't.

Engineering note

Class 3 is not "Class 2 done more carefully" — it changes the design. Larger annular rings and pads need more room, which can push a dense board to more layers or into HDI. That's why the class has to be decided before layout is finished, not bolted on at fabrication.

What it costs

Class 3 costs more, and the cost is real: tighter tolerances lower yield, mandatory coupons and microsections add test, and the extra copper and inspection add process. Depending on complexity, expect a meaningful premium over the same board at Class 2. That premium is worth every rupee on a board that can't fail — and pure waste on one that can.

Which class do you need?

Work backwards from consequences of failure:

Class 2 — consumer products, industrial controls, IoT, most commercial electronics. The default unless something specific pushes you up.
Class 3 — aerospace and defense, medical devices, automotive safety systems, and anything where a field failure is a safety or mission event. Often contractually required.

There's also Class 3/A and space-grade addenda for the most extreme cases, but for most teams the real decision is simply 2 vs 3.

Get the class right in the front end

The class you declare drives DFM rules, stackup, and how the data package is prepared. Declaring Class 3 late — or discovering mid-build that your design can't meet it — is the expensive path. Deciding early, and having the front end verify the design against that class, is the cheap one.

Not sure which class your board needs, or whether your current design meets it? Send us your files and we'll tell you where it stands — before it reaches the fab.