USA Energy Efficiency Compliance Testing Services

Check the applicable test route before booking

Start with the product’s technical characteristics, intended use, marketed configuration, model family and date of manufacture. Determine whether it is a covered consumer product under 10 CFR part 430 or covered commercial/industrial equipment under part 431, whether an applicable conservation standard and test procedure govern, and which version and compliance date control. A label such as “energy efficient” does not answer those questions.

DOE compliance, the voluntary EPA ENERGY STAR program, state programs such as California’s, and FTC labeling can involve different evidence and acceptance rules. Do not assume that passing one program means acceptance by another; request separate scopes where relevant.

Testing route matrix: what to commission

This decision matrix is a planning aid, not a product-by-product legal determination. Confirm the exact section and applicable dates before authorizing a quotation.

Sampling and represented ratings: preserve the chain of evidence

DOE generally requires randomly selected production units, or units representative of production, of each basic model; § 429.11 generally sets at least two test units, subject to product-specific and one-unit-production exceptions. The required sample size and statistical calculation are not simply a laboratory’s suggested package. Complete these checks before shipping:

Under 10 CFR § 429.11(b)(2), testing just one unit is an exception when only one unit of that basic model is manufactured. If further units are later manufactured, the default sampling and representation requirements apply; this is not a standing one-sample shortcut.

For imported covered products and equipment, DOE treats the importer as a manufacturer for these requirements. Obtain the underlying test data from the original equipment manufacturer (OEM); an OEM summary or laboratory invoice does not transfer the importer’s compliance responsibility.

• Define the basic model and every variant it represents; document energy-relevant differences in controls, components, capacity and configuration.
• Record the selection method, production status, serial numbers, firmware, accessories and test-unit custody. Prototypes are not automatically interchangeable with production-representative units.
• Locate the exact product-specific part 429 sampling plan and determine sample size, any exceptions and the calculation of represented values; a single passing unit does not establish a compliant rating.
• Identify the test-procedure version and any effective or compliance date relevant to testing, energy representations and sale. Avoid mixing metrics across old and new procedures.
• Preserve measured data and calculations, not just a pass/fail cover page; § 429.71 imposes manufacturer recordkeeping obligations.

Choosing a laboratory: six checks before purchase

DOE states that for most covered product categories it does not prescribe special test-facility qualifications, but some categories do have specific laboratory or accreditation rules. “DOE-approved laboratory” should not be treated as a universal credential. Ask for documentary answers:

An important exception to the general DOE laboratory position is the separate, voluntary ENERGY STAR product-label program: EPA requires tests from an EPA-recognized laboratory and review by an EPA-recognized certification body before a product may carry that label. A witnessed or supervised manufacturer testing laboratory may qualify only through the applicable EPA-recognized program; generic ISO/IEC 17025 accreditation alone is not proof of EPA recognition for this product and method.

• Has the site performed this exact DOE test appendix and model class—not merely generic efficiency testing?
• Can it show the current accreditation scope or program recognition where the chosen product rule or additional scheme requires one?
• Are the equipment, calibration records, environmental controls, power supply and uncertainty practices appropriate for the method?
• Who signs the report; which subcontracted tests and locations will be used; and can all raw data be supplied?
• Will the quotation include set-up, preconditioning, repeat runs, extra samples, failure analysis and shipping?
• Can the laboratory document departures and escalate an untestable design before improvising a method?

Existing reports: an evidence acceptance gate

An older report is useful only if its scope and underlying proof still support the current basic model and regulatory method. Reuse is a documented decision, never a blanket entitlement.

Failures and retesting: a change-control decision path

A failing unit, unsupported rating or design change should stop reliance on the affected evidence until investigated. Follow the sequence:

1. Quarantine the affected result and record the failed metric, operating mode, unit identity and test conditions.
2. Reproduce and investigate laboratory set-up, measurement error, production variability and actual product performance; do not erase unfavorable data.
3. Compare proposed fixes—firmware, compressor/motor, control logic, insulation, accessories or factory process—with the tested configuration.
4. Decide what new units, repeat tests, sample statistics and report revision are needed under the specific rule; check waiver or AEDM conditions if used.
5. Approve a revised represented value only after evidence supports it; assess whether DOE certification records or other program claims need updating.

Send the laboratory this seven-question brief

Before accepting the quotation, obtain a written test plan answering:

• Which DOE product/equipment definition and basic-model boundary are assumed?
• Which exact regulation, appendix, method edition and compliance date will be applied?
• What sample count, random-selection method and statistical rating calculation are required?
• Is any accredited/recognized facility or third-party test specifically required for this category or additional scheme?
• What configurations, controls, loads, ambient conditions and preconditioning will be tested?
• What raw data, calculations, deviations, traceability and reusable deliverables will be supplied?
• What are the stop-work rules and cost/time drivers for failed tests, changes and retesting?

Where the testing decision stops

This page helps specify tests and judge evidence. For classification strategy, complex scope disputes and regulatory gap analysis, see

Official sources to verify before testing

Use the current product-specific rules as controlling text; general DOE guidance is explanatory. Key starting points:

• DOE implementation, certification and enforcement
• 10 CFR § 429.11: sample selection
• 10 CFR § 429.70: authorized alternative methods
• 10 CFR § 429.71: testing records
• DOE test procedure waivers
• EPA: ENERGY STAR recognition and third-party certification
• DOE: importers’ compliance and test-data duties

Frequently asked testing questions

Does the DOE require an independent, accredited laboratory for every product?

No. DOE explains that for most products it does not impose special test-facility requirements; exceptions exist. Check the precise product rule and separately verify ENERGY STAR or other scheme recognition.

Can we submit one successful prototype test?

Not automatically. A basic model generally needs production or production-representative random sampling and the applicable part 429 statistical approach; § 429.11 usually starts at two units, with exceptions.

Will a European or ENERGY STAR report satisfy DOE?

Only if the underlying test method, unit configuration, sample plan, calculations and required evidence actually match the applicable DOE obligations. A different logo or standard is not proof of equivalence.

Can software modeling replace physical tests?

Only where the product-specific rule authorizes an alternative efficiency determination method and its validation and record conditions are fulfilled under § 429.70.

Does a redesign or failed test always force full retesting?

Not necessarily, but reuse must be justified through the affected measured parameters, basic-model grouping, applicable procedure, sampling rule and any program-specific change policy. Do not keep an unsupported rating.