One board, two requirements that pull in opposite directions
Heavy copper demand starts with current. Power converters, inverters, welding equipment, UPS systems and industrial control boards have to move tens of amps or more across a limited board area, and conductor cross-section is the variable you can actually change. Sunking's heavy copper range is 3oz to over 12oz copper, 1 to 16 layers, 0.8 to 6.0 mm board thickness, 6 mil minimum trace width, on FR-4 TG170/TG180.
The complication is that the same board usually also carries control signals, sense returns and gate drive. Those nets want density and impedance consistency, not copper weight. As copper gets thicker, more material has to be etched away at a given trace width, the absolute width deviation grows, and the achievable minimum trace width backs off. Making every layer heavy forces your signal layers to absorb the penalties of heavy copper; making every layer thin means the current path does not work at all.
So in manufacturing a heavy copper multilayer board is normally not "a thick copper board" but a mixed-weight stackup: different copper weights on different layers, each layer doing one job. This is exactly what the heavy copper capability page describes — thin inner layers (1–2oz) carry signals while heavy copper layers (3oz to over 12oz) handle high current, supported by sequential lamination, resin-rich prepreg fill and heavy copper etch compensation.
How a mixed-weight stackup is arranged
Answer three questions, in this order.
First, which nets genuinely need heavy copper. Only high-current paths and areas that rely on copper for heat spreading. Removing layers you were thickening "while you were at it" is the single most effective way to control both board thickness and cost.
Second, whether the heavy layers go inside or outside. Outer-layer heavy copper gives a short thermal path and lets you solder high-current terminals directly, but outer etching happens before solder mask and finish, so width deviation is exposed on the finished part. Inner-layer heavy copper is enclosed in dielectric and dimensionally more stable, but heat has to leave through vias and copper planes, and resin fill between heavy layers is more demanding. There is no universal answer here; it depends on the current path and the mechanical design, and it belongs in engineering review.
Third, how symmetry is maintained. Mixed weights are inherently asymmetric: 3oz on one side against 1oz on the other means the copper volume difference shows up as stress difference through lamination and later thermal exposure. You can mirror the copper distribution about the board centreline as closely as possible, or use dielectric thickness to offset the copper imbalance. How much asymmetry is acceptable depends on board thickness, layer count and flatness requirements, and is again a review item rather than a number the designer can set alone.
A reasonable starting point: conventional copper on the two outer layers for routing and soldering, an adjacent pair of inner layers in heavy copper for current, separated from the signal layers by a thicker core. Treat it as a starting point, not a template — once layer count rises, heavy layer placement has to be re-planned together with impedance layers and reference planes.
Three consequences that come with the heavy layers
Mixed-weight copper is not "raise the number on a few layers". It moves three things at once.
Resin fill. After a heavy layer is etched, there is a pronounced height step between remaining copper and cleared areas, and prepreg resin has to flow in and fill it during lamination. Thicker copper and larger cleared areas mean more resin is required, which is why heavy copper builds use resin-rich prepreg and often sequential lamination. What the designer can contribute: avoid large copper-free areas on heavy layers, keep copper where copper is allowed, and bring the step height and fill demand down.
Board thickness accumulation. Every layer's copper adds to finished thickness, heavy layers noticeably so, and the extra dielectric needed for fill adds on top. Finished thickness on a mixed-weight build is often meaningfully greater than an estimate based on the signal layers. Sunking's thickness tolerance is ±10% or ±0.005" (≈0.13 mm), whichever is greater, over a heavy copper range of 0.8 to 6.0 mm. If enclosures, connectors or press-fit parts impose a hard thickness limit, state it as a stackup constraint up front rather than trying to squeeze it after the stackup is set.
Etched line width deviation. Trace width tolerance is ±20%, and copper-to-copper spacing tolerance is also ±20%. Thicker copper makes that percentage a larger absolute deviation, and at the same time the achievable minimum width backs off as copper weight rises. Fine lines on a heavy layer are the design choice most likely to come back from manufacturing. The actual etch compensation depends on copper weight, pattern density and equipment, and must be confirmed by engineering review against your data — no fixed coefficient is given here, because no single coefficient holds across board types.
These three interact. Adding copper for fill changes the impedance reference; thinning dielectric to hold thickness erodes fill margin. A stackup is solved as a whole, not filled in layer by layer.
The usable ranges for layer count and copper weight
The heavy copper page gives 1 to 16 layers and 3oz to over 12oz. Read those two ranges separately.
- Layer count: Sunking's overall capability is 40 layers standard production and up to 64 layers, but the normal range for a heavy copper structure is 1–16 layers. Beyond the range on the heavy copper page you are combining heavy copper with high layer count, which needs its own review — do not multiply two independent maxima together.
- Copper weight: the stated maximum is over 12oz, with 16oz and 18oz already produced. This is conditional, advanced-tier capability and must always be presented with "subject to engineering review".
- Parameters that get constrained alongside: mechanical minimum hole 0.2 mm standard (0.15 mm advanced), laser minimum hole 0.1 mm standard (0.075 mm advanced), aspect ratio 10:1, impedance tolerance ±10% standard (±5% advanced). Asking a heavy copper stackup for small holes, high aspect ratio and tight impedance at the same time is where conflicts surface in review.
The rule is straightforward: values inside the standard tier can be designed against directly; anything in the advanced tier, or anything that pushes two capability axes at once, must be called out in your submission and confirmed by engineering. Do not assume that because each item is individually feasible, the combination is.
What the stackup drawing has to state
Manufacturing reads a stackup drawing looking for executable inputs; every missing item costs a query cycle. A heavy copper multilayer stackup should state at minimum:
- Copper weight per layer, listed layer by layer, distinguishing base copper from plated copper. Not a single blanket "heavy copper" label.
- Dielectric designations and nominal thicknesses, cores and prepreg separately. Where you are not sure, "to be set by manufacturer per fill requirement" is better than a blank.
- Layer order and reference planes, identifying which layers are signal, which carry current, and which plane each controlled-impedance net references.
- Impedance requirements, target and tolerance (±10% standard; mark ±5% explicitly as an advanced requirement), with permission for the manufacturer to adjust trace width to hit the target.
- Symmetry and thickness requirements, including the finished thickness target and any hard mechanical limit.
- Copper-free area notes on heavy layers, stating whether large cleared regions may be filled with copper or thieving, since this directly drives the fill strategy.
How it is verified in manufacturing
Once the stackup is fixed, verification runs on the manufacturing side against IPC-6012 Class 2/3 criteria:
- Incoming copper foil inspection — confirms base copper weight and foil condition.
- Post-etch copper thickness and trace width measurement — whether heavy layer widths land inside ±20% is the critical process point on a mixed-weight build.
- Multilayer lamination registration — mixed weights bring larger layer-to-layer movement differences, and registration determines whether inner-layer hole-to-copper still holds.
- Microsection analysis — fill completeness, voiding, and layer thicknesses against the stackup drawing.
- Current-carrying and dielectric testing — verifies the current path and interlayer insulation.
What designers can do is turn those checks into measurable callouts on the drawing: which layer's copper thickness, which trace segment's width, which net's impedance is to be measured. Anything not called out gets judged against generic criteria, which may not be the answer you wanted.
When to stop and ask
Take these to engineering review rather than deciding in design: copper above 12oz; heavy copper combined with 16+ layers; trace widths approaching 6 mil on a heavy layer; a finished thickness limit that cannot move; impedance tightened to ±5%; heavy layers that also need 0.15 mm class mechanical holes or high aspect ratio. None of these are automatic refusals — feasibility depends on the specific combination and has to be judged against your actual data.
Submit data through the heavy copper capability page. Process windows are on manufacturing tolerances and the capabilities overview; inspection and release criteria are on quality assurance; base material options are in the materials library. For pricing use the quote request.
FAQ
Can a heavy copper multilayer board use 3oz+ on every layer?
It can be built, but it is usually unnecessary and expensive. All-heavy construction forces signal layers to accept the minimum-width penalty and larger width deviation of heavy copper, while board thickness and fill demand both rise. The purpose of a mixed-weight stackup is to put heavy copper only where current flows.
Are heavy layers better on the inside or the outside?
It depends on the current path and thermal strategy. Outer heavy copper gives a short thermal path and direct terminal soldering but exposes width deviation on the finished part; inner heavy copper is dimensionally more stable but must move heat through vias and planes and demands more of the resin fill. Let engineering review decide against the actual structure.
Does mixing copper weights make board thickness uncontrollable?
No, but it will come out thicker than an estimate based on signal layers. Both the copper itself and the additional dielectric needed for fill have to be counted. Finished thickness tolerance is ±10% or ±0.005" (≈0.13 mm), whichever is greater. Where a hard thickness limit exists, raise it as a stackup constraint.
What is the minimum trace width on a heavy copper board?
The heavy copper page states 6 mil minimum trace width. The achievable minimum backs off as copper weight rises, and the specific combination must be confirmed by engineering review against your data and equipment conditions. Trace width tolerance is ±20%.
Who provides the etch compensation?
The manufacturer, based on copper weight, pattern density and equipment conditions, confirmed in engineering review. You do not need to pre-compensate trace widths, but you should identify which widths are functionally critical so they can be prioritised.
What is the maximum copper weight available?
The stated maximum is over 12oz, with 16oz and 18oz already produced. This is conditional advanced-tier capability requiring engineering review of feasibility and applicable structure.