The Prototype Works. The 100-Unit Batch Doesn't. Here Is Why.

Author: Srihari Maddula  •  Founder & Technical Lead, Eurth Techtronics Pvt Ltd

Category: Manufacturing Realities  •  Estimated Reading Time: 18–20 minutes

Published: July 2026

The Prototype Is a Sample of One

A prototype that works is a necessary condition for product launch. It is not a sufficient one. A prototype is assembled by hand, with careful component selection, on a PCB that was reviewed multiple times, by an engineer who understands the circuit and can compensate for marginal conditions through judgement and adjustment. A production batch of 100 units is assembled by a process, with components from a BOM that was sourced at the time of procurement, on PCBs manufactured to a tolerance specification, by operators following a work instruction.

The prototype and the batch are not the same thing. The gap between them — the transition from one working unit to a hundred consistent units — is where embedded product development either matures or fails. This is the stage that first-time hardware product teams are least prepared for, because engineering education and early-stage development culture both optimise for prototype success, not manufacturing consistency.

Failure Mode 1: Component Tolerance Stack-Up

Every passive component — resistor, capacitor, inductor — has a tolerance on its value. A 10-kilohm resistor with 5% tolerance could be anywhere from 9.5 to 10.5 kilohms. In a circuit where the resistor value is not critical, this is irrelevant. In a circuit where the resistor sets a threshold, a reference, or a time constant, 5% variation can push some units into incorrect operation.

The problem compounds in circuits where multiple tolerances stack. A voltage divider using two resistors each with 5% tolerance can produce a divider ratio anywhere from 90.25% to 110.25% of the nominal. If this divider sets a battery low-voltage detection threshold, some units will detect low battery too early and some too late. In the prototype, the specific components used happened to be close to nominal. In the batch, the distribution of values across units produces a distribution of behaviours — some correct, some marginal, some wrong.

The engineering discipline required is worst-case tolerance analysis on every critical circuit before finalising the design. Use 1% resistors where the value is circuit-critical, not 5%. For capacitors, account for voltage coefficient — ceramic capacitors lose significant capacitance under DC bias, sometimes 50 to 80% at rated voltage. The 10 μF capacitor in your circuit may actually be providing 3 μF under operating conditions.

Failure Mode 2: Solder Joint Variability

In prototype assembly, solder joints are inspected individually by the person who made them. In production assembly, solder joints are inspected by automated optical inspection (AOI) with criteria that catch gross defects but miss subtle issues — insufficient solder, marginal wetting, cold joints that look acceptable at room temperature but develop intermittent contact under thermal cycling.

Fine-pitch QFN packages — the flat no-lead packages common on modern MCUs and RF chips — are particularly vulnerable. The thermal pad underneath the package is the ground and thermal connection, and it requires precise stencil design and reflow profile control to solder correctly. An improperly soldered thermal pad produces a unit that works initially but fails under thermal load as the marginal joint develops cracks. This is a classic root cause for field failures that appear 2 to 6 months after deployment.

The solution is not inspecting your way to quality — it is designing for manufacturability and then qualifying the process. This means working with your PCBA vendor to design an appropriate solder stencil (aperture size and shape for the thermal pad), agreeing on a reflow profile that has been validated for your specific board stackup, and establishing a first-article inspection process that includes cross-section analysis of critical joints on the first batch.

Failure Mode 3: BOM Drift and EOL Substitutions

The component in your prototype BOM may not be available when you go to production. Component availability in the global electronics supply chain is volatile — a part that was readily available six months ago may now have a 52-week lead time or may have been discontinued by the manufacturer. Your PCBA vendor will substitute an 'equivalent' component without necessarily consulting you.

The substituted component may have a different package variant with slightly different land pattern requirements. It may have a different electrical characteristic that matters in your specific circuit. It may have a different temperature coefficient that shifts your circuit's behaviour across the operating range. The vendor's substitution is in good faith, but it is based on electrical specifications, not on how that component interacts with your specific circuit.

The discipline here is an approved vendor list (AVL) for every critical component, with a defined substitution approval process. Any deviation from the AVL must go through engineering review before production. This is standard practice in the medical device and automotive industries, where it is enforced by regulation. In commercial IoT products, it requires deliberate enforcement, but the cost of not doing it — in field failures and recall investigations — is significantly higher than the administrative overhead of doing it.

Failure Mode 4: Test Coverage — The Functional Test Gap

The prototype was tested extensively by engineers who understood the system. Every function was exercised. Every edge case was explored. The production batch is tested by a functional test fixture that tests what was specified in the test plan.

The gap between 'what engineers tested' and 'what the test plan specifies' is where production escapes happen. A production escape is a unit that passes the production test but has a defect that manifests in the field. Common escapes include: marginal RF performance that passes at room temperature but fails at temperature extremes; intermittent firmware bugs triggered by specific input sequences not covered in the test script; sensor calibration drift that is within tolerance at assembly but outside tolerance after thermal cycling.

Test jig design is a first-class engineering activity. The test jig should exercise every electrical interface of the device, apply stimulus at its operating extremes, and check responses against calibrated limits derived from characterisation data — not from datasheet nominal values. The test engineer should be involved in product design review, not handed a finished product and asked to write a test plan for it.

Failure Mode 5: Firmware Version Mismatch in Production

Production firmware is different from development firmware. Development firmware has debug logging enabled, may have test modes active, and may have assertions and error checking that are removed in production builds to save code space. The transition from development to production firmware introduces a risk: the production firmware has different behaviour in edge cases, because the edge-case handling paths were tested in the development firmware but not in the production build.

The production firmware build must be treated as a distinct engineering deliverable, with its own validation test pass before it enters production. The programming fixture must be configured to flash a specific, hash-verified firmware binary — not 'the latest build from the development machine.' Firmware traceability — knowing which specific firmware binary is running on which serial-numbered unit — is required for any product where field failures need to be investigated.

What Design for Manufacturing Actually Means

Design for manufacturing is not a checklist of layout rules. It is a mindset that considers the production process as part of the design specification. It means designing circuits with tolerances that accommodate component variation. It means selecting package types that your PCBA vendor can solder consistently. It means writing an approved vendor list before finalising the BOM. It means designing the test fixture as part of the product design, not after it.

At EurthTech, with 26+ delivered products, the manufacturing lessons are ones we learned on real production runs, with real defects, and real field returns. The engineering instinct to optimise the design and hand it to manufacturing is natural. The discipline to design the manufacturing process as carefully as the electronics is what separates products that scale from prototypes that worked.

THE RULE: Your prototype proves the concept.

Your manufacturing process proves the product. Invest in both with equal rigour.

© 2026 Eurth Techtronics Pvt Ltd  |  eurthtech.com  |  All rights reserved.