Comparing bare-board and assembled-system cost

Compare a rigid-flex architecture at the level of the deliverable you need. A bare-board quotation covers a different scope from an assembled and tested system. If an integrated board removes interconnect parts or changes assembly operations, account for those changes explicitly before deciding which architecture costs less.

Define an equal comparison boundary

Choose a boundary such as “assembled interconnect subsystem ready for product integration.” Then include every item required to reach that boundary for both options. If one offer stops at bare boards while the other includes assembly and testing, the printed totals are not directly comparable.

These invented values show why removing parts does not automatically guarantee a lower subtotal. They are not market prices or our quotation. The example also excludes one-time tooling, freight, taxes, inventory effects, and service costs; a real comparison should identify and evaluate its own exclusions.

Separate recurring and one-time costs

If an architecture requires a fixture or other one-time charge, keep it in a separate row and evaluate it at the intended order quantity. Do not mix a prototype setup charge into one offer’s unit price while excluding it from the other. When quantities change, recalculate the comparison using valid offers for that quantity.

For a simple arithmetic extension, a 500-unit purchase at the illustrative subtotals gives 22,500 and 23,000 units of currency before excluded charges. Adding a hypothetical 1,000 one-time charge only to the first concept changes its total to 23,500. The ranking changes because the comparison now includes another stated input.

Record operational differences without assigning imaginary savings

Integration may change assembly steps, test access, rework options, and the unit that must be replaced after a defect. List those differences, then request measured or quoted evidence where they affect the decision. Do not assume a fixed yield improvement or labor saving simply because fewer connectors appear in the drawing.

A practical worksheet has three columns for each claimed saving: the changed operation, its evidence source, and its cost treatment. If the evidence is missing, label the saving as unconfirmed rather than subtracting it from the total.

Keep the engineering equivalence visible

Würth Elektronik’s flex-solutions design guide distinguishes PCB cost from system cost and includes assembly, installation, testing, and quality costs in the broader view. That distinction supports a comparison framework; it does not establish a universal saving for a particular rigid-flex design. The following examples are original arithmetic exercises using invented inputs.

Write the deliverable in one sentence

Before entering prices, describe exactly what the comparison buys. For example, the fictional boundary could be two populated functional islands connected, programmed, checked, and ready to install in an enclosure. This statement makes it easier to notice when a quoted item is missing. A connector purchased separately still belongs in the comparison if it is required to reach that boundary.

List exclusions immediately below the boundary. If enclosure installation is excluded, exclude it from both alternatives and state that the result is a subsystem comparison. If firmware programming is included in one assembly quote, do not add it again as an independent operation. Equal boundaries matter more than a visually tidy table containing two totals that describe different work.

Make assembly-time assumptions reproducible

Suppose a proposed architecture change is expected to remove 90 seconds of manual work per unit. At an invented burdened labor rate of 30 currency units per hour, the arithmetic saving would be 90 / 3,600 × 30 = 0.75 currency units per unit. Across 500 units, that is 375 currency units. The calculation is straightforward; the uncertain part is whether 90 seconds is a valid time reduction for the actual production process.

Identify which operations disappear and which new operations appear. Removing connector mating might add a controlled folding or retention step. A timing study should use a defined work sequence and representative conditions, rather than comparing an experienced operator on one design with an early prototype build of the other. Keep the observation record and its limitations beside the calculated saving.

Keep tooling allocation tied to a quantity

A one-time fixture charge of 1,000 currency units contributes 10 per unit when allocated across 100 units, but only 2 per unit when allocated across 500 units. Neither allocation changes the amount paid for the fixture. It changes how the comparison presents that amount against a chosen production quantity. State the quantity and whether the fixture can actually be reused for later orders.

Do not spread a charge across an optimistic lifetime volume while comparing the other architecture at the committed first order. Use the same decision horizon. If future demand is uncertain, show multiple quantity scenarios and keep the original charge visible. A reviewer can then see whether the apparent advantage depends on purchasing a volume that has not yet been authorized.

Calculate a break-even point without hiding its assumptions

Use the earlier example with recurring costs of 45 for the separate-board concept and 46 for rigid-flex. Add a one-time cost of 1,000 to the separate-board concept and zero to the integrated concept solely for this arithmetic exercise. The totals become 1,000 + 45Q and 46Q, where Q is the number of subsystems. They are equal at Q = 1,000.

Below that quantity, the integrated concept has the lower total under these specific inputs; above it, the separate-board concept has the lower total. At 500 units the totals are 23,500 and 23,000. At 2,000 units they are 91,000 and 92,000. This reversal illustrates why a preferred architecture can change with quantity. Real quotations may have price breaks, minimum orders, or additional charges that make the simple linear model inappropriate.

Handle yield and rework as evidence-based inputs

A quoted cost per started assembly differs from cost per accepted assembly when some units do not meet the defined acceptance boundary. For an invented batch costing 4,600 currency units with 95 accepted units, dividing the batch cost by accepted quantity gives approximately 48.42 per accepted unit. That result depends on what costs were included and whether rejected units have recoverable value or require further work.

Do not use that example as an assumed yield for rigid-flex or a separate-board design. Obtain project-specific evidence or label the input as a scenario. If rework costs are already included in the batch total, adding another blanket rework percentage would double count them. Keep the numerator, accepted quantity, and treatment of remaining units explicit so another reviewer can reproduce the calculation.

Review the replacement unit and service strategy

An integrated assembly may change what must be replaced when one portion is damaged. A separate-board concept might permit replacing one island, while an integrated concept may require a larger replacement unit. Whether that difference matters depends on the product’s service strategy, repair access, diagnostic process, and actual failure experience. Record the architectural difference before attempting to assign a cost.

If evidence is insufficient, present service cost as an unresolved comparison item or a bounded scenario. Avoid manufacturing a failure rate merely to complete the spreadsheet. A transparent open question is more useful than a precise lifetime total built from unsupported assumptions. The design decision can still proceed if the responsible team understands and accepts the remaining uncertainty.

Test which assumptions can change the decision

Vary one uncertain input at a time within a justified range. For example, compare the result with and without the proposed 0.75 labor saving while keeping all other inputs fixed. If the preferred option changes, labor evidence deserves attention before the decision is closed. If the ranking remains unchanged across plausible values, that particular uncertainty may be less important than another unresolved item.

Keep technical acceptance outside this arithmetic shortcut. A cheaper architecture that fails the required geometry, electrical function, or validated motion is not an equivalent alternative. The final record should identify both the engineering basis for eligibility and the cost assumptions used to choose among eligible options. That keeps a purchasing comparison connected to the product that must actually be built.

Check that both architectures meet the same functional and mechanical requirements. If one accepts a different service strategy or validation scope, record the difference alongside cost. Purchasing should not accidentally approve an engineering tradeoff hidden in a lower number.

Date the comparison and identify the validity of each quotation. A later material change, revised volume, or expired offer can make a previously correct total unsuitable for the next order. Update the affected inputs and preserve the earlier version so the decision history remains clear. The worksheet should show which values are current quotes, which are measured observations, and which remain planning assumptions.

Use our engineering and cost-review tools and manufacturing overview. Submit both architecture options for quotation review with the same quantities, release status, and required deliverables.