Double-Sided Assembly in SMT: Order, Support and Reflow Control

A battery monitoring board for an energy storage system carries several sampling harnesses, a long outline, and components on both faces. Building it is not a matter of running the same process twice. The first side has already been through reflow before the second side is printed, placed and heated again, and the second pass is influenced by what is already on the board: the mass of the parts, the direction of the pads and the way the panel is supported.

Split the Data by Side Before Anything Else

The import step for a double-sided assembly begins with a partition. The bill of materials and the placement coordinates are separated by face, and the designators, packages, polarity and rotation are checked within each group.

The most common source of confusion is a naming difference. One document may call the faces top and bottom while another calls them primary and secondary, or the coordinates may use a mirrored convention for the underside. When the naming conventions do not agree, parts are placed on the wrong face, and the error is not visible until the board is assembled.

The assembly drawing should exist per side, with orientation and polarity marked, and a customer who supplies component alternates should confirm the parameters, the package and the applicable designators. On a monitoring board the parts that require particular care are the isolated communication devices, the sampling chip and the precision resistors that set the measurement accuracy. A substitute chosen by package size alone will change the calibration, so the selection is executed against the customer confirmation rather than against a visual match.

Deciding the Order of the Two Passes

The usual sequence is to build the lighter or less populated side first, then the denser or heavier side, but the final decision depends on the board rather than on the rule. If the heavier devices are on the first side, the second pass re-melts the solder that holds them, and a component whose mass is large relative to its joint area can shift or fall off during the second reflow.

Three arrangements are used when that risk exists. The heavy parts can be moved to the second side, if the layout allows it. A small amount of adhesive can be dispensed under the heavy parts before the first pass. Or the second pass can be run with a profile that keeps the underside cooler, which is discussed below.

The order also affects the connectors. A connector that protrudes from the board makes the panel unstable on the conveyor and can interfere with the printing of the other side, so connectors are usually placed on the side that is built last where the design permits.

Two Stencils, Two Decisions

The apertures for the two sides are evaluated separately, because the pads are different. On the component side of the small precision parts, the objectives are paste release and a balanced deposit at both terminations, since an imbalance is what creates a tombstone. On a thermal pad, the objective is a controlled volume, and the aperture is often divided into windows so that the paste can release and so that the part does not float on a single large deposit.

After printing, the deposit is checked for coverage, offset, insufficient paste and bridging, and an anomalous board is corrected before placement rather than after reflow. This is the same sequence that applies to a single-sided board, and it matters more here because a defect found after the second pass involves reworking a board that has already been through two reflow cycles.

Supporting a Long Board

A monitoring board is often long and narrow, which makes it prone to sagging during printing and placement. A board that sinks under the squeegee produces a paste deposit of varying thickness along its length, and a board that moves during placement produces a placement offset that varies with position.

Carrier support is the answer, and it is arranged by the shape of the board rather than by habit. Support pins are placed under the areas where components will be placed, avoiding the parts already fitted on the underside, and a carrier is used where the outline or the connector height makes pin support impractical. The support must not press on a pad, a fitted component or a surface that will be inspected later, and it must allow the board to expand during reflow rather than holding it rigidly.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/electronic-assumbler.webp" alt="measured reflow profile for a second assembly pass” />

The Second Reflow Profile

The reflow profile for the second pass has a different problem from the first. The lower face carries components held only by solder, and their joints are re-melted while the upper face is being processed. The profile therefore has to reach a peak that forms the new joints without leaving the old ones molten for longer than necessary, and it has to do so without a thermal gradient across the board that warps it.

The practical measures are a measured profile rather than an inherited one, a check that the underside temperature stays below the melting point for as long as possible before the peak, and a board support arrangement that keeps the assembly flat. Where a design has a large thermal mass on one side and small parts on the other, the profile is a compromise that is verified with a profiler on the actual assembly.

Some designs cannot be made to work this way at all, and the alternative is a selective or a reflow-with-adhesive process. That decision is made at the design stage, which is why the assembly partner is worth consulting before the layout is released rather than after the first build.

Inspection Across Two Passes

Inspection follows the same order. Each side is inspected after its own reflow, and the second inspection is planned so that it does not disturb the parts fitted on the first side. Optical inspection covers the visible joints and the placement of every component on both faces, and any hidden joint is covered according to the customer drawing and the risk assessment for the design.

Boards held for review, boards reworked and boards that passed are kept separate through both passes, because a defect detected on the second side may have originated on the first. The electrical test at the end is the first point at which the two sides are considered together, and for a sampling board it is also the point at which the measurement accuracy of the assembled unit can be confirmed.

What the Customer Should Supply

The data package for a double-sided build is the same as for a single-sided one, with two additions: a per-side assembly drawing, and a clear statement of which side is which. Where programming or a sampling test is required, the program file, the connector definitions, the supply conditions, the sampling point description and the pass criteria belong in the package as well.

Our SMT assembly team reviews the two-sided arrangement before the first build, because the decision about which side goes first is worth making on paper rather than on the line. Where the board also has to be fabricated with a particular construction, the stack-up and the assembly arrangement are reviewed together, as described on our PCB capabilities page, and the results of both passes are recorded in our quality management system so that a defect can be traced to the pass that produced it.

FAQ

Which side should be assembled first? Usually the side with the lighter and fewer components, unless a heavy part forces a different order or an adhesive process is used.

Can the same stencil be used for both sides? Only if the pad geometry is identical. The apertures are normally evaluated separately, because the components and the paste volumes differ.

Is a carrier always needed? Not always, but a long board with components on both faces usually needs support that pin tooling alone cannot provide.