Technical guidePDF2023

ENERGY STAR Program Requirements for Computer Servers: Final Draft Test Method (Rev. Jan-2023), Version 4.0

U.S. Environmental Protection Agency (ENERGY STAR) · U.S. Environmental Protection Agency, U.S. Department of Energy

The test method used with the ENERGY STAR Computer Servers Version 4.0 specification to obtain idle and active state power. It sets input power, ambient conditions, measurement instrument requirements and the test arrangement for blade and multi-node systems.

U.S. public domain — U.S. Government work; free to reuse

On this page

Overview

This seven-page document is the test method for determining compliance with the ENERGY STAR Computer Servers specification and for obtaining idle and active state power data for certification. It does not apply to the power supply efficiency and power factor requirements in Section 3.2 of the specification; DOE revised the overview to make clear that the voltage and related power conditions do not apply to power supply testing.

Test setup begins with three input power tables. For products with nameplate power of 1,500 W or less, voltage tolerance is 1.0% and maximum total harmonic distortion is 2.0%; above 1,500 W these become 4.0% and 5.0%. Supply voltages are 230 V or 115 V ac (100 V optional for the Japanese market), and three-phase servers use 208 V (North America) or 400 V (Europe). Frequency tolerance is 1.0%. Ambient temperature must be 25 ± 5 °C and relative humidity 15% to 80%.

The power analyzer must report true RMS power, with a current crest factor of at least 3, a minimum frequency response of 3.0 kHz, 1% or better accuracy, at least one reading per second and calibration within the past year. The temperature sensor sits no more than 50 mm upwind of the main air inlet, with ± 0.5 °C accuracy and at least 4 samples per minute. Active state testing uses the most current SPEC SERT.

Units are tested as shipped, with hardware, BIOS and power management at defaults. Power is measured between the ac source and the unit, with no UPS in between. Blade systems are tested at chassis input in a half-populated chassis (full population optional); multi-node systems are tested in a fully populated chassis, and idle power is divided by the node count. Result files such as Results.xml and Results.txt accompany the report.

Key points

  1. Ambient temperature during testing is 25 ± 5 °C and relative humidity 15% to 80%.
  2. Voltage tolerance is 1.0% and maximum THD 2.0% at 1,500 W or less; 4.0% and 5.0% above 1,500 W.
  3. The analyzer reports true RMS power with 1% or better accuracy, 3.0 kHz minimum frequency response, crest factor of at least 3 and calibration within a year.
  4. The temperature sensor is within 50 mm of the air inlet, accurate to ± 0.5 °C, logging at least 4 samples per minute.
  5. Products are tested as shipped; no UPS may sit between meter and unit, and all PSUs must be connected and operating.
  6. Blade power is measured at the chassis input; half-populated chassis is required and full population is optional.
  7. Multi-node systems are tested in a fully populated chassis, with idle power divided by the number of nodes.
  8. The Ethernet link must be live, and 802.3az capable products connect only to compliant equipment.

Using it in design

  • When importing server power data, keep the 25 ± 5 °C test ambient as metadata and warn that fan power rises at higher inlet set points.
  • Model blade and multi-node catalogue items at chassis level, matching the measurement boundary (chassis input for blades).
  • State the test configuration (half-populated blade chassis, per-node division) in the record before scaling to a fully populated chassis.
  • Note in the measurement-boundary field that PDU overhead is included in the reading and UPS losses are not.

Monthly newsletter · free

Keep your reading list current

Once a month: new library sources, changes to Turkish regulations and upcoming data centre events.

SoSo CAD

Compare the cost of design decisions

Evaluate equipment costs, energy expenses and total cost of ownership (TCO) using your inputs and assumptions in SoSo CAD.

Request a demo
CAPEX and TCO estimates in the SoSo CAD cost workspace

Who should read it

Design engineers who use server power and efficiency data in catalogues or load models, and ENERGY STAR test laboratories.

Document contents

  1. 1 Overview
  2. 2 Applicability
  3. 3 Definitions
  4. 4 Test Setup (input power, ambient temperature, relative humidity, power analyzer, temperature sensor, active state test tool, controller system)
  5. 5.1 Test Configuration
  6. 5.2 UUT Preparation
  7. Blade chassis population rules (half and full)
  8. 6 Test Procedures for All Products
  9. 6.1 Idle and Active State Testing Using SERT

Related standards

  • SPEC SERT 2.x
  • ENERGY STAR Computer Servers V4.0

Read the document

energystar-servers-test-method-v4.pdf

PDF · 0.1 MB · English

U.S. public domain — U.S. Government work; free to reuse

The document is provided unaltered because its licence permits redistribution.

Publication details

Authors
U.S. Environmental Protection Agency; U.S. Department of Energy
Publisher
U.S. Environmental Protection Agency (ENERGY STAR)
Year
2023
Document number
Rev. Jan-2023
Document language
English
Source type
Government / public
Access
Open licence — downloadable
Licence
U.S. public domain — U.S. Government work; free to reuse
Licence evidence
www.copyright.gov
Original licence note

Work of the U.S. Environmental Protection Agency ENERGY STAR program (federal agency publication on energystar.gov); not subject to copyright under 17 U.S.C. 105.

Citation

U.S. Environmental Protection Agency; U.S. Department of Energy (2023). ENERGY STAR Program Requirements for Computer Servers: Final Draft Test Method (Rev. Jan-2023), Version 4.0. U.S. Environmental Protection Agency (ENERGY STAR). Rev. Jan-2023. https://www.energystar.gov/sites/default/files/asset/document/Computer%20Servers%20Version%204.0%20Final%20Draft%20Test%20Method_0.pdf

All content and rights belong to the respective authors and publishers. SoSo CAD provides only the citation, its own summary and a link to the source, and is not responsible for the sources' content.