Msslc Dallas2012 Steinberg (1)

2012 LED Street Lighting Workshop – Dallas, TX

A Rational View of LM-79 Reports, IES Files,

and Product Variation Gary A. Steinberg Systems Mgr, Outdoor SSL Fixtures GE Lighting Solutions

• Φ (Tj) ~ 0.12 - 0.25 %/°C

• Vf (Tj) ~ -2.5 – 4.0 mV/ °C

• Tj (Ta, t)

Why LM-79 …


Temp/Orientation

• Natural convection

• 25 ± 1 °C

• Heatsinks may be orientation sensitive

• Orient per mfgr’s recommendation


Thermally Stabilize

• Because Φ (Tj) and Vf (Tj) …

• More Mass => More Time


No Initial Seasoning

• Up or down in first 1- 2K hrs

• 0-3% Effect

• Test at 0 hrs


Electrical Requirements • Steady state AC or DC • Max Power or Mfgr’s Modes • AC Voltage Uncertainty < 0.2% • AC Power Uncertainty< 0.5% • DC Voltage/Current Uncertainty < 0.1% • Low Distortion/Low Impedance AC • Excellent AC Regulation +/- 0.2%


Standardize • Conditions


• Methods

• Content

• NOT FORMAT

Lab A

Lab B

“All Significant Data” • Test Lab/Date • Product Mfgr and Product Designation • Measurement Qtys (TLF, LPW, CCT, CRI) • Electricals (Volts, Amps, Watts, PF, THD) • Hours on Product and Hours in test • Ambient Temperature • Orientation • Stabilization Time • Test Method • Calibration Method/Reference Standard

• Correction factors • Measurement Geometry (2∏ or 4∏) • Luminous Intensity Distribution

(graphical and/or candela table) • Color Quantities • SPD • Spectroradiometer Parameters • Equipment List • Statement of Uncertainties (if req’ed) • Deviations from SOP if any


Sample Report


What

How

Why Who

Accreditation

Sphere Photometry • Absolute Photometry • Total Luminous Flux • Color Characteristics • 4Π or 2Π Methods • Self absorption, spectral,

and spatial corrections • Spectroradiometer • Photometer head (not

preferred) Image used with permission courtesy Labsphere



Sphere Data

Parameter Table

Chromaticity Diagram

Spectral Plot

SPD Table

Goniophotometry • Absolute Photometry

• Intensity Distribution • Angles per typical LM’s • Spectral Corrections • Type C



Gonio Data

TM-15 LCS Graph

CU Graph

Polar Plots

Candela Table Iso-E Plot

How many LM-79 Reports?



Typical Modular Roadway Cat Logic

> 2400 Photometric/Electrical Variations

Distributions Lumen Outputs

CCT Options

PSU Options

Supply Voltages

Drive Currents

Simple LM-79 Financials


>2400 x $500 = > $1,200,000

>$1,000 => REALLY BIG

Captive Internal Cost – LM-79 Sphere + Gonio

External Lab Price – LM-79 Sphere + Gonio

Rational Approach


• Detailed System Model

• LM-79’s at Ends of Scalable Ranges

• Tested/Validated Scaling Factors

System Model


• LED Factors • Thermal Factors • Power Supply Characterization • Optical Efficiencies • 2nd Order Effects • Application Testing

Scalable Ranges


• Limited Tsp Differences

• Known Optical Efficiency

Tested/Validated Scaling


CTQ’s Lumens Power Application Performance

Parent SKU’s

…

Child SKU

…

Child SKU

… Model

Predict Measure

Correct

Pop Quiz


Q: Will a purchased unit “match” the data in the submitted LM-79 report?

A: Mostly

Q:Why?

A: Tolerances and Variation

Uncertainty … up to a few percent

Measurement Uncertainty • Calibration Standards • Electrical Sources and

Instruments • Corrections • Optical Instruments • Geometric Influences

Image used with permission courtesy Labsphere


Tolerances > zero • CCT - ~± 300K • CRI - ~ -2 to +X Pts • THD - ≤ 20% • PF - ≥ 0.9 • Watts - ± 10 % • Flux - ± 10 % • Distrib – 10% Max CP


Typical Power Variation


VF Tol = +/- 4% 60 LEDs/Class 2 PSU IF Tol = 5% Tsp +/- 3 °C of Avg

ηPSU = +/- 2% “Normal” Data

Histogram

0

200

400

600

800

1000

1200

1400

1600

1800

89 91 93 95 97 99 101

103

105

More

Bin

Freq

uenc

y

.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

FrequencyCumulative %

Final Test Power ERMC 95W (Pulsed)

Rated USL

Where is your LM-79 sample? Random? Selected?


Which sample criteria? • LM-79 … ‘consider need to average’ …

no specific requirement • Other LM’s … ‘sample should be

representative’ • Confirmance audits … e.g. Energy Star • 6σ methods … Z-Scores and confidence

intervals


LM-79 Expectations


• No performance guarantee

• Sample of ONE is in spec

• Tolerance Knowledge Req’ed

Should Published IES File = LM-79?

• Sampling • Ratings


LM-79 Report Limits • LM-79 typically has ONE IES file … from ONE sample


Pessimistic? Optimistic?

Reasonable?

Published IES files should … • Predict expected “typical” field performance.

Test 100810XX Test 10081014 Test 10081015


…

“Published” IES File Criteria


• “Rated” typical lumens

• “Rated” typical power

“Published” IES File Criteria


• Representative distribution

• Application impact

IES File Opportunities


• “Industry accepted” methods to create files … sampling guidance.

• Means to capture or represent spatial variability in the distribution … or application

• Standard methods to capture or represent flux and power variability

• Lots of opportunities to enhance/improve!

Thank You! Gary A. Steinberg

Systems Mgr, Outdoor SSL Fixtures GE Lighting Solutions

[email protected]

Msslc Dallas2012 Steinberg (1)

Documents