Double-Sided Rigid PCB: Structure, Design and Price

What a Double-Sided Rigid PCB Is

A double-sided rigid board carries copper on both faces of a stiff insulating core and connects the two faces with plated through holes. The laminate is usually FR-4, occasionally a high glass transition grade or a composite where cost dominates, and the board holds its shape without any support, which is what makes it rigid.

It sits between two extremes. A single-sided board carries one copper layer and no plated holes, so the routing has to be solved on one plane and components are mounted on one side. A multilayer board stacks four or more copper layers with several lamination cycles. The double-sided board uses two of everything and one lamination, which is why it delivers a large step up in capability for a modest step up in cost.

The Construction

• Rigid core. A cured laminate sheet, most often FR-4, that provides the mechanical support and the electrical insulation between the two copper layers.
• Copper on both faces. Usually one ounce, or two ounces where the board carries power, patterned separately on each side.
• Plated through holes. Drilled and copper plated barrels that connect a pad on the top face to a pad on the bottom face, so a signal can change sides or reach a component mounted on the other face.
• Solder mask. Applied to both faces, protecting the copper and defining the pad openings.
• Legend. Reference designators and orientation marks, usually on one or both faces.
• Surface finish. Hot air levelling, an organic solderability preservative, immersion tin or electroless nickel immersion gold, chosen from the pad pitch and the assembly process.

Why It Is the Default for So Many Products

Routing density. Two layers with vias between them roughly square the number of crossings that can be solved, because a trace that would have to cross another can change side and pass underneath. That single capability turns an impossible single-sided layout into a comfortable one.

Two-sided assembly. Components can be placed on both faces, which halves the board area for a given circuit or allows the heavier through hole parts to be kept on the underside, away from the surface mount parts.

Ground and power distribution. One side can carry a large ground pour or a power plane, which lowers the impedance of the supply and gives the signals a defined return path. On a single-sided board this is impossible, and the noise performance suffers accordingly.

Mechanical strength. The plated holes and the copper on both sides stiffen the board, and the through hole joints are strong enough for connectors and switches that are pulled and pressed by the user.

Cost. The lamination is a single cycle and the drilling is a single pass, so a two layer board costs far less than a four layer board while covering a large part of the same circuit complexity.

Design Rules That Matter

Trace width and spacing. Standard signal traces run from about 0.15 to 0.25 millimetres, and the practical minimum spacing depends on the copper weight and the fabricator. Power traces are widened in proportion to the current, and the copper thickness is increased where the current is high.

Via design. A via carries the signal from one side to the other. Its diameter, its pad and its annular ring are chosen from the drill capability of the plant and the current it must carry, and every via changes the impedance of the trace passing through it, which matters on a fast signal.

Component placement. Keep the heavy through hole parts on the side that will be wave soldered, and remember that a component on the bottom face has to survive the reflow of the top face and, sometimes, a second reflow of its own.

Ground pour. Fill the unused copper on at least one side with a ground pour, connected to the ground net, which provides a return path and improves the thermal spreading. Keep the pour away from the edges by the fabricator minimum.

Test access. Bring the nets you will want to probe to accessible pads or vias, because a two layer board with no test points is very difficult to debug. Our notes on PCB design and layout cover these rules in more detail.

How the Board Is Made

The laminate is cut to panel size and the holes are drilled. Copper is deposited in the holes, either by an electroless process that seeds the barrel followed by electrolytic plating that builds the thickness, or by a direct metallisation process that replaces the electroless step. The two faces are then patterned, by imaging a dry film resist, exposing and developing it, and etching the exposed copper away. Copper is plated up on the traces in a pattern plating line where thicker copper is needed. The mask is printed and cured, the legend is printed over it, and the pads receive their surface finish. Electrical test checks every net for an open or a short, and the boards are inspected, packed and shipped.

The plated through hole is the step that distinguishes this board from a single-sided one, and it is also the step that fails if the process is not controlled. An incomplete plating leaves a thin barrel that cracks when the board is heated during assembly; our notes on PCB manufacturing describe the plating controls and the coupon tests that detect it.

Where It Is Used

• Consumer electronics. Televisions, set top boxes, routers, toys and appliances, where the circuit is moderate and the price is decisive.
• Automotive accessories. Lighting, interior control and low power modules outside the engine bay.
• Industrial control. Relay boards, sensor interfaces, power distribution and simple drive circuits.
• Medical devices. Monitoring instruments and diagnostic equipment, where a two layer board is sufficient and the mechanical strength is useful.
• Communication equipment. Network switches, power over Ethernet injectors and simple radio boards.

Single Sided, Double Sided or Multilayer

A single-sided board is the cheapest and is limited to simple circuits with no need for a ground plane. A double-sided rigid board adds the second layer, the vias and the possibility of a plane, which covers most medium complexity designs at a moderate cost. A multilayer board adds dedicated planes and more routing layers and is needed where there are fast interfaces, several supply rails or a dense component count.

The decision is usually made on the number of crossings that cannot be resolved, the need for a continuous reference plane, and the assembly and test requirements. Two layers and a good ground pour will carry a surprising amount of design before a fourth layer is needed, and the saving is significant.

What It Costs

Indicative pricing for a standard 1.6 millimetre FR-4 board with one ounce copper and a hot air levelled finish runs around 0.90 to 1.50 US dollars per board at one hundred pieces, and about 0.60 to 1.10 dollars at five hundred pieces for the same outline. The price moves with board area, the number and size of the holes, the copper weight, the surface finish and the delivery schedule.

Electroless nickel immersion gold adds cost over hot air levelling but is required for fine pitch surface mount assembly. An expedited build adds a substantial premium, and the premium is larger when the finish and the test are non-standard, because the plant has to insert the job into an existing sequence. Our notes on prototype PCB assembly describe how to use a first article before committing to a production volume.

Choosing a Manufacturer

Ask for the achievable minimum trace width and spacing, the smallest drill and the smallest annular ring, the finished board thickness tolerance, the surface finishes offered and the test method. Ask also how the plating thickness in the holes is verified and whether a cross section coupon accompanies the order. Those four answers, plus the certifications, separate a plant that can make a two layer board cheaply from one that can make it reliably; our notes on quality management describe the process controls behind them.

FAQ

Is a double-sided PCB the same as a two layer PCB? Yes. Both describe a board with copper on two faces of one rigid core, connected by plated through holes.

Can a double-sided board have controlled impedance? Yes, if the geometry is fixed against a reference plane. It is easier when one side carries a continuous ground plane, and it is verified with a coupon.

How much does a double-sided board cost? Around 0.90 to 1.50 US dollars per board for a standard small outline at one hundred pieces, falling with quantity.

What is the usual lead time? Five to seven working days in production, with expedited options measured in a day or two at a premium.

Conclusion

A double-sided rigid board is the workhorse of medium complexity electronics. Two copper layers and plated through holes turn a difficult layout into a simple one, allow components on both faces, and provide a ground reference at a fraction of the cost of a multilayer board. Specify the finish from the assembly process, control the plating, test every net, and use it for everything it can carry before moving to four layers. Our notes on PCB assembly describe what happens to the board next.