Heavy Copper PCB Mechanical Strength: What to Put in the RFQ | PCB168

Nobody arrives at heavy copper for structural reasons. You get there because a rail has to carry tens of amps, or because a power device needs somewhere to dump its heat, and copper weight gets pushed from the usual 1-2oz up past 3oz. The decision is electrical and thermal.

What shows up immediately afterwards is structural:

• assembly says the boards sit slightly bowed and the placement machine struggles to hold them flat;
• mechanical engineering asks whether this board can be bolted straight onto a heatsink, and whether the mounting points need washers;
• production asks whether the array is V-scored or routed, and whether the press-fit connector can survive depaneling;
• reliability asks whether those big holes will crack first after a few hundred thermal cycles.

By the time these reach the supplier they have usually been compressed into one sentence: "is this heavy copper board strong?" That sentence is the single most misleading thing in the whole conversation.

1. Copper weight is not stiffness: heavy copper changes consequences, not strength

A PCB's bending stiffness comes overwhelmingly from the laminate - the resin and glass cloth - not from the copper. Copper has a higher modulus than FR-4, but it occupies a small fraction of the board thickness and it is discontinuous, etched into traces and planes. Going from 1oz to 6oz does not hand you a stiffer board.

What heavy copper genuinely does is three things: it increases conductor cross-section (the thing you wanted), it changes how material is distributed inside the board, and it changes the stress state during manufacturing. The last two are where every structural surprise comes from.

This distinction is not wordplay, because it decides how the RFQ gets written. Believe that heavy copper equals strength and you will write "requires heavy copper PCB with high mechanical strength" - a line no quoting engineer can act on, because it maps to no process decision. Accept that heavy copper changes consequences and the same requirement becomes five sets of specific conditions, each of which a quoting engineer can read, each of which lands on a drawing, and each of which can be checked at acceptance.

How copper weight itself gets chosen - current, temperature rise, duty cycle - is a separate question this article does not repeat. That one is covered in current and thermal budget to copper weight. Here, copper weight is already frozen. This is the structural bill that arrives afterwards.

2. First item: copper weight eats your thickness budget

PCB168 publishes heavy copper capability as 3-20oz copper, 1-16 layers, and 0.8-6.0mm board thickness. Buyers routinely treat those three lines as three independent dropdowns. They are coupled.

The arithmetic is unavoidable: finished thickness equals dielectric layers plus copper layers plus finish and mask. Moving from 1oz to 6oz takes a single copper layer from roughly 35µm to roughly 210µm, and on a multilayer board that increment is multiplied by however many copper layers you have. The copper has to come from somewhere - either the board gets thicker, or the dielectric gets squeezed. And dielectric thickness is simultaneously carrying voltage withstand, impedance, and layer-to-layer reliability.

The practical consequence for a buyer is this: board thickness is not a number you casually type in. It is an item where you either give a range or give the constraint that produced it.

• If thickness is set by mechanical fit - a card slot, a clip, a shield can, a press-fit depth - say so, and give workable limits.
• If thickness is just "same as last revision", say that it can float. That gives engineering room to trade between copper and dielectric.
• If you simultaneously lock high copper weight, high layer count, and thin finished thickness, those three may not be physically compatible, and your quote turns into a feasibility review.

The mutual ceiling between layer count and thickness is its own topic, covered in what limits layer count on thick copper multilayers. The point to carry out of this section is narrower: thickness tolerance is a structural line item, not a typo. PCB168 does not publish a thickness tolerance figure, so tolerance has to be confirmed against drawing and order; the general industry basis is IPC-6012 (industry specification, not a PCB168 internal control value, and the referenced revision governs).

3. Second item: warp and twist come from copper imbalance

When a buyer says heavy copper boards warp, the mechanism is usually not "too much copper". It is copper distributed unevenly inside the board.

Lamination heats, presses, and then cools. On cooling, copper and resin do not contract by the same amount. If a layer's copper is concentrated on one side of the panel, or if vertically opposed layers have very different copper coverage, that imbalance freezes into the finished board as bow or twist. Thicker copper amplifies it. So heavy copper boards really are more prone to warp - but the cause sits in the distribution, not the thickness.

Which matters because it means the fix is mostly on the design side, and supplier-side compensation is limited. So the valuable RFQ input here is design information, not an adjective:

• Is copper distribution symmetric? Do full planes appear on only some layers? Is coverage on vertically paired layers comparable?
• Is added copper permitted? Copper thieving or balance patterns - electrically dead copper placed in empty areas - is the standard remedy. State explicitly whether it is allowed, and name the exclusion zones (antennas, shielded regions, high-voltage isolation, test point areas).
• Is an asymmetric stackup intentional? If the stackup is deliberately asymmetric for impedance or cost reasons, say so, otherwise engineering may flag it as a suspected error and come back to you.
• Where does the flatness requirement come from? Is it an incoming inspection criterion, or a real limit on your assembly equipment? If it is the equipment, telling the supplier what the machine tolerates is far more useful than an abstract figure.
• What assembly does it see? Single-sided versus double-sided reflow load the board differently. Put the assembly route in the RFQ.

Flatness criteria should reference a standard (IPC-6012 / IPC-A-600 - industry specifications, not PCB168 internal control values) and must state the measurement condition: at room temperature or after reflow, whole board or local area, how supported. A flatness requirement without a measurement condition reliably turns into a dispute at acceptance.

4. Third item: copper weight changes depaneling, but do not specify the process yourself

"V-score or rout?" is the most common structural question on heavy copper, and the most common buyer mistake is specifying the process in the RFQ.

V-scoring cuts a groove into both faces, leaves a web, and the board is broken apart. It is cheap, fast, and suits straight-edged arrays. The catch is that breaking applies bending stress near the separation line, and as the board and the copper get thicker the force needed to break it rises, along with the stress transmitted into the board and into nearby parts. Routing mills the outline directly and imposes much less stress, but it is slower, costs more, and needs retaining tabs.

So heavy copper does tilt this trade toward routing. But whether V-scoring is viable, and what web thickness is appropriate, is a manufacturing feasibility judgement based on thickness, copper weight, material, and equipment. Specifying "V-cut" in your RFQ gets you one of two outcomes: the supplier complies and the risk stays with you, or it comes back for discussion anyway.

Give inputs instead, and leave the process judgement with the party who can carry it:

1. Outline and array intent - outline file, your preferred array and up count, or an explicit "array optimisation up to supplier".
2. Stress-sensitive locations - which parts sit near the board edge? Any ceramic capacitors, BGAs, connectors, or crystals near separation lines? This is the most frequently omitted and most valuable line in the whole section.
3. Board edge requirements - plated edge? Sliding or insertion contact? Burr limits, since deburring requirements themselves influence process choice.
4. Who depanels - do you receive arrays or singles? If you separate them, what equipment do you use?
5. Where outline tolerance comes from - driven by a mating part, or inherited?

The acceptance criteria for edges and depaneling themselves - burrs, delamination, chipping, dimensions - are covered separately in board edge and depaneling acceptance criteria. This section answers only how heavy copper shifts the choice and what you owe the supplier.

5. Fourth item: fastening and mating interfaces, the real question behind "strength"

Back to that opening sentence. When someone asks whether the board is strong, they are almost never thinking about an unsupported board snapping. They are thinking about three interfaces.

Mounting screws, board to chassis or heatsink. Tightening concentrates load on a small area around the pad and hole, and on a heavy, stiff board overtightening is the common failure, not loosening. The RFQ should carry: mounting hole locations and diameters, plated or not, whether they tie to copper or ground, washer and hardware types, the torque range you plan to use, and whether there are parts or copper on the opposite face to avoid. Note the framing: torque is your assembly condition, not a PCB parameter. You state it so engineering can judge whether the pad, annular ring, and surrounding copper are adequate - not so the supplier can guarantee a torque figure.

Press-fit connectors and pressed-in hardware. Press-fit forms a mechanical and electrical joint by pressing a terminal into a hole, and it depends simultaneously on hole diameter tolerance, hole wall copper, and board thickness. Spell out: connector part number and the manufacturer's recommended hole tolerance window, who performs the press operation, and how many insertion cycles are allowed including any rework. Do not omit this one - if hole diameter drifts outside the connector window the result is not degraded performance, it is a terminal that will not go in or a cracked barrel.

Busbars and large-area joints. High-current designs frequently bolt or solder copper bar onto a large plane. Provide joint location and area, finish requirements (large solderable or pressable areas drive different finish choices), mating part material and thickness, and assembly sequence.

Vibration and shock magnify all three, but translating environmental figures into quotable inputs is a different article: industrial heavy copper operating environment in the RFQ. The only note needed here: if vibration is in play, fastening interface detail moves from nice-to-have to mandatory.

6. Fifth item: hole life does not come from copper weight, it comes from stating what the hole must survive

Heavy copper invites the assumption that thick copper means strong holes. Plated through-hole reliability actually rests on the integrity and uniformity of the barrel copper, and on the strain the hole sees during thermal cycling - which is set by board thickness, hole diameter, and material expansion behaviour. Heavy copper boards tend to be thicker, so aspect ratios run higher, which puts them on the more demanding side of that equation rather than the safer side.

You do not need to specify plating parameters, and you should not. What you can give, and what quoting actually needs, are service conditions:

• Thermal cycling expectation - how often does the equipment power cycle, what is the operating temperature span, and is there a defined cycling test requirement? If there is, give the test conditions directly.
• Assembly passes - how many reflow cycles, any wave solder, hand soldering, or rework, and what is the rework limit?
• What each hole is for - which holes carry high current, which are mechanical, which are signal only? The three have different concerns.
• Material already fixed - PCB168 publishes FR-4 TG170/TG180 for heavy copper, and high-Tg laminate is inherently kinder to thermal stress. If your application needs something else, say so rather than assuming.

Hole acceptance again references IPC-6012 (industry specification, not a PCB168 internal control value; referenced revision governs), and the class - Class 2 or Class 3 - must be declared by you. That is a buyer decision, not a supplier default.

7. Making it acceptable: what belongs in first article, what belongs in routine lots

Structural requirements usually fail not because nobody wrote them, but because nobody said how they would be verified. Splitting evidence into two tiers removes most of the later argument.

First article - proves the structure was built as designed:

• dimensional report on outline and mounting hole positions, against your outline file;
• measured finished thickness, against the range or constraint you supplied;
• confirmation of how stackup and copper distribution were realised, including any balance copper;
• flatness measured under the condition you declared;
• cross-section analysis to confirm barrel copper and layer structure - PCB168 lists Cross-Section Analysis among its published quality operations, so it can be requested at first article;
• for press-fit or screw interfaces, measured hole diameters against your mating window.

Routine lots - prove the result stays consistent:

• AOI on every layer and 100% electrical test (both published as standard PCB168 operations);
• sampled thickness and key dimensions;
• appearance and board edge - burrs, chipping, delamination;
• sampled flatness, same condition as first article;
• sampled diameters on large and press-fit holes.

One working principle: first article proves the design was realised correctly; routine lots prove realisation stays stable. Migrating every first-article item into routine inspection raises cost and slows delivery while rarely reducing real risk.

Tying that evidence back to physical goods at receipt is covered in incoming inspection and acceptance planning, and checking these requirements against drawing and specification for contradictions is covered in the fabrication notes conflict checklist. Treat this whole article as the expansion of the mechanical and structural section inside the broader heavy copper purchase specification and acceptance framework.

8. Five RFQ lines that turn "strong" into something quotable

The seven sections above, compressed into five lines you can paste into an RFQ. Each maps to something a supplier can act on.

Line 1 - Structural conditions (why there is a structural requirement at all)

> This is a [power / drive / ...] board. Copper weight is fixed at [N]oz on [inner / outer / mixed] layers, [N] layers, finished thickness [target or range], thickness constraint driven by [mechanical fit / inherited / may float]. Mounting method is [bolted to heatsink / clipped into chassis / ...]. Assembly route is [single-sided reflow / double-sided reflow / includes wave solder].

Line 2 - Outline and depaneling constraints

> Outline per [file]. Array [supplier to optimise / per attached]. Stress-sensitive parts near separation lines: [types and locations, or "none"]. Board edge [requires / does not require] plating, [has / has no] insertion contact. Depaneling performed by [us / supplier]. Please select the depaneling process on feasibility grounds and state it in the quotation.

Line 3 - Fastening and mating interfaces

> Mounting holes: [locations / diameters / plated / tied to copper]. Hardware [screw and washer spec]. Planned torque range [value]. Press-fit: [connector part number] requires hole tolerance [window]; press operation by [us / supplier]; rework re-insertion [permitted / not permitted]. Busbar joints: [location / area / mating part material and thickness / finish requirement].

Line 4 - Acceptance basis and measurement conditions

> Acceptance class [IPC-6012 Class 2 / Class 3] (state the revision relied upon). Flatness requirement [value and criterion], measured [room temperature / after reflow], [whole board / local], support method [describe]. First article to include: outline and mounting hole dimensional report, measured finished thickness, stackup and copper realisation confirmation, cross-section analysis, [measured press-fit hole diameters]. Routine lots: AOI every layer, 100% electrical test, sampled thickness and key dimensions, appearance and board edge, sampled flatness.

Line 5 - Change triggers

> Written notice and re-confirmation required for any change to: stackup or copper distribution, extent of balance copper, material grade or Tg, depaneling process, machining method for mounting or press-fit holes, surface finish. After a change, first article [is / is not] to be resubmitted.

How to negotiate the change clause itself is covered in repeat order and change control terms. PCB168's published heavy copper range sits on the heavy copper PCB product page and the manufacturing capabilities page; once you have the five lines above, submit them for quotation.

9. One sentence

Heavy copper makes the conductor stronger. The structural outcome is decided by copper distribution, thickness budget, depaneling method, fastening interfaces, and what the holes have to survive - so writing those five as conditions with acceptance criteria beats writing "high mechanical strength required", because the first can be quoted and the second can only be guessed at.

FAQ

Q1: Does thicker copper mean the board bends less?

No. Bending stiffness comes mainly from the laminate; copper is a small, discontinuous fraction of board thickness. Heavy copper changes conductor cross-section and internal material distribution, and the second of those actually raises warp risk, because copper imbalance gets amplified. Control deformation by keeping copper distribution symmetric and permitting balance copper, not by adding copper weight.

Q2: Can heavy copper boards still be V-scored?

That is a manufacturing feasibility judgement, dependent on thickness, copper weight, material, and equipment, and it belongs with the supplier at quotation rather than with the buyer in the RFQ. Heavy copper and thick boards do push the trade toward routing, because the break force is higher and more stress reaches the board and nearby parts. What you should supply is outline, array intent, stress-sensitive part locations, and edge requirements.

Q3: Should I put a warp figure in the RFQ?

You can, but only together with the measurement condition - room temperature or after reflow, whole board or local, how supported - and a referenced criterion such as IPC-6012 or IPC-A-600 (industry specifications, not PCB168 internal control values). A figure without a measurement condition cannot be reproduced at acceptance and will become a dispute. Often more useful: state what your assembly equipment actually tolerates.

Q4: Who defines the mounting torque requirement?

Torque is your assembly condition, not a PCB delivery parameter. Put the planned range, hardware spec, and opposite-face obstructions in the RFQ so engineering can judge whether pad, annular ring, and surrounding copper are adequate. Do not ask a supplier to guarantee that the board withstands a given torque; that is not something a PCB specification can promise.

Q5: Class 2 or Class 3 - who decides?

The buyer declares it. It is an acceptance class reflecting your application, and a supplier should not default it on your behalf. Declare the standard revision alongside it, since criteria differ between revisions.