The "4,000lm" LED Ceiling Lights Differed in Brightness and Color From Room to Room
LED Lumen & Color Temperature Test Guide - Why LED Lights That Passed Inspection Get "Too Dim" and "Every Fixture Is a Different Color" Complaints After Installation, Watts vs Lumens, Reading CCT, CRI and Color Consistency (SDCM), Integrating Sphere Tests Compared With Power, Light Distribution, Lumen Maintenance and Flicker Tests, Approximate Criteria, Gaps in Factory Reports, Supplier Substitution Such as Cheaper Chips, Mixed Bins and Swapped Drivers, Spec Sheet Items, Stage-by-Stage Checks and a Checklist

Hello, this is StartMade.

A distributor of interior materials imported 2,000 LED ceiling lights from a factory in Guzhen, Zhongshan, China's lighting hub. The rated specs were 50W, 4,000lm, 4000K neutral white and CRI Ra80 or higher. The approval sample looked bright enough on the office ceiling, and the pre-shipment inspection found no problems with appearance, lighting, remote control operation or packaging.

About a month after installations began, the calls started. Some customers said the lights were dimmer than their old ones, but the harder complaint was "the living room light and the bedroom light are different colors." Both were 4000K units, yet one looked yellowish and the other bluish. Five units from remaining stock went into a lab's integrating sphere: luminous flux came to 2,850-3,100lm, power was 38-41W, and color temperature ranged from 3,650K to 4,450K. CRI was Ra72. A teardown showed the same number of LED chips as the approval sample, but from a different chip maker, and chips from different color bins were mixed even within a single fixture.

Brightness and color defects in LED lighting are tricky because they barely show at inspection. A light that switches on looks bright, and when units are lit one at a time, color differences are hard to notice. The problem only surfaces once a customer hangs several of them side by side in the same space. This article starts with how to read watts, lumens and color temperature, then covers test types and criteria, common supplier tricks, and how to build lumen and color testing into your spec sheet and inspection checklist.

A reply of "Yes, it's 50W 4,000lm" or a switch-on check of one unit at a time is not enough. To prevent post-installation claims, set in numbers the luminous flux (lm) of the finished fixture with its cover on, power (W), the color temperature range, color consistency (SDCM), CRI, and the LED chip maker and bin codes in your spec sheet, then run integrating sphere tests on samples drawn from production lots.

1. Why Two "50W 4,000lm" Lights Can Differ in Brightness and Color

Start with the units. Watts (W) only tell you how much electricity a light uses; how bright it actually is shows up in luminous flux (lm, lumens). Two 50W lights can deliver 3,000lm or 5,000lm depending on the chips, driver and thermal design. That makes luminous efficacy (lm/W), lumens divided by watts, the yardstick for quality. For today's indoor LED fixtures, around 80-120lm/W for the finished fixture with its cover on is typical.

Color comes down to three numbers. Correlated color temperature (CCT) expresses in kelvin (K) whether the light leans yellow or blue, and the color rendering index (CRI, Ra) shows how naturally objects' colors appear. The one most often missed in practice is color consistency (SDCM). LED chips vary slightly in color even from the same line, so chip makers sort and sell them in grades (bins) by color and brightness. Two 4000K lights using chips from different bins can look visibly different side by side. SDCM expresses that difference in steps; the smaller the number, the more uniform the color.

When the label and reality are far apart, the cause is usually one of these five.

• False lumen labeling (虚标): The factory adds up the chip-level flux from the LED datasheet and lists it as fixture flux. Losses from the cover, lens and heat are left out, so the figure runs 20-30% high.
• Chip substitution: Approval samples use chips from a known maker, while production uses cheaper, lower-output chips.
• Mixed bins (混bin): To use up leftover stock, chips from different color bins are assembled together. Color then varies within one fixture or between fixtures.
• Swapped drivers or lowered drive current: Replacing the power circuit (driver) with a cheaper one or setting a lower current drops both power and brightness, and can lower the power factor or cause flicker.
• Poor heat dissipation: A thinner heat sink or a cheaper board in place of an aluminum board raises chip temperature, so brightness drops within tens of minutes of switching on. A few months later the light is noticeably dimmer.

Brightness and color need to be checked separately from safety certification. Safety certifications such as KC look at shock and fire risk and electromagnetic emissions; they do not guarantee that the labeled lumens or color temperature are accurate. We covered lighting imports in general in Episode 78 and certification in Episode 10. This article focuses on whether the labeled brightness and color actually come out of the fixture.

2. Test Types — Integrating Sphere Tests and Supporting Tests Compared

Ask simply for "a brightness test" and the factory may take one floor reading with a lux meter or send an LED chip datasheet and call it done. Results for the same product change with the measurement method, time after switch-on, whether the cover is fitted and ambient temperature, so agree on the method and conditions first.

Get Values Measured After Stabilization

An LED is brightest right after switch-on, then dims gradually as chip temperature rises and stabilizes after 30 minutes to an hour. The worse the thermal design, the bigger the drop. If a factory report lists values measured right at switch-on, the number is higher than the brightness in real use. Write the conditions into the spec sheet, for example "finished fixture with cover, rated voltage, ambient 25±1℃, measured after light output stabilizes," and ask for a spectral report showing flux, power, CCT, CRI and SDCM on one page.

Screening on Site First

There are simple checks you can run before sending anything to a lab. Pick several units from the same lot and light them side by side against a white wall; color differences show up surprisingly well. Differences you miss when lighting units one by one become visible once there is something next to them for comparison. A handheld spectral illuminance meter gives a rough comparison of CCT and CRI, and a plug-in power meter quickly catches units whose power is far off the label. If power is more than 20% below the label, brightness is likely low by about as much.

3. Setting the Criteria

The ranges below are approximate figures commonly referenced in the industry. Requirements vary by destination country, sales channel and installation space, so use the table only as a starting point and set actual criteria based on your product and a testing lab's advice.

Comparison claims such as "equivalent to a ○○W fluorescent" or "twice as bright as before" are easy to overlook in sales copy. If you use them without fixing the flux of what you are comparing against, customer expectations can miss the mark even when measured flux matches the label. If you sell into a country with an energy efficiency labeling scheme, also confirm that efficacy and power match the registered values.

4. Why You Shouldn't Take Factory Reports at Face Value

Ask a lighting factory for brightness data and it will usually send the LED chip maker's datasheet or a certification test report. These documents have a few gaps.

5. Common Supplier Tricks

Once LED chips are under the cover they can't be seen from outside, and the maker or bin is hard to tell from a chip's appearance, which makes substitution easy. Here are the types we have actually encountered in consultations.

The most common is the one in the opening case: only the approval samples get good chips. Appearance, lighting and remote operation are identical, so an inspector switching units on one at a time can't tell. Keep a few approval samples aside and light them next to production samples, or sphere-test both under the same conditions, and most cases get caught. We explained the same principle of material substitution for plastics in Episode 125.

6. The Damage Brightness and Color Defects Leave Behind

Brightness and color defects don't show at shipment and only appear after installation, but the damage spreads well beyond the price of the product.

• Returns that include installation costs: With lighting, swapping the product isn't the end of it. Removal, reinstallation and call-out costs come with it.
• Partial replacement doesn't work: In a store or office with dozens of the same fixture, if a few are off-color you often end up re-matching the whole space with one lot.
• Labeling and advertising issues: If labeled flux or power is far from reality, it can be treated as false labeling.
• Early failure: Poor heat dissipation doesn't just reduce brightness; it shortens chip and driver life, so failure claims pile up within one or two years.

See Episode 53 for product liability and recalls and Episode 133 for electronics burn-in testing.

7. Items to Put in the Spec Sheet

A single line such as "50W 4,000lm 4000K" gets read differently by every factory. Put the items below into a table and attach it to your spec sheet. See Episode 59 for writing spec sheets and Episode 45 for color control between lots.

Asking for chip reel label photos may seem like extra hassle, but it does more of the work. A sphere test only checks a few samples, while reel labels show whether the chips used in production have the maker and bin you specified. See Episode 119 for the procedure when chips or drivers change.

8. Building Lumen and Color Tests Into Your QC Checklist

Sample Stage

When approval samples arrive, set a few aside and send one to a third-party lab for a sphere test separate from the factory's report. If possible, tear one down and photograph the chip maker markings and count, heat sink weight, board material and driver model. That record becomes the baseline for comparing production units.

Production Stage

During production, check the maker and bin code on the reel labels of incoming chips, and require that bins are not mixed within a lot. Get sphere test records for each lot, but if the results show the same number every time and there are no photos of the measurement screen, first question whether the test was actually done. Getting pre-shipment burn-in records as well reduces early failures. See Episode 102 for production tracking.

Pre-Shipment Inspection Stage

The inspector picks packed units personally, lights several side by side and checks for color differences. They measure power with a power meter and tear down one or two units to confirm the chips and heat sink match the approval sample. A handheld spectral meter is used to compare CCT and CRI, and leftover samples go to a lab for sphere testing. See Episode 100 for pre-shipment inspection.

9. Sample Request Wording for the Factory

Putting the wording below into your purchase order or spec sheet takes chip and driver choices out of the factory's discretion and makes it easier to assign responsibility if brightness or color claims arise.

"本产品须使用与确认样相同品牌、型号及色温·光通量分档（BIN）的LED芯片，禁止混用不同分档芯片。成品（含灯罩）在额定电压、环境温度25±1℃、光输出稳定后以积分球测试，光通量不得低于4000lm，功率50W±10%，色温4000K（符合ANSI C78.377范围），色容差不超过3 SDCM，显色指数Ra≥80，R9>0。未经我方书面同意不得更换LED芯片品牌、型号、分档、驱动电源、散热器、灯罩材料或生产工厂，如需更换须事先书面通知并重新测试。每批须提交注明品号、批号、测试条件、结果及照片的积分球测试报告，以及芯片出货凭证和芯片料盘标签照片。测试样品由我方验货员随机抽取并可拆机检查。如有不合格，须加抽复测；复测仍不合格的批次暂停出货。"
(This product must use LED chips of the same maker, model and CCT/flux bin (BIN) as the approval sample; mixing chips from different bins is prohibited. When the finished fixture (including the cover) is measured in an integrating sphere at rated voltage and an ambient temperature of 25±1℃ after light output stabilizes, flux must be at least 4,000lm, power 50W ±10%, CCT 4000K (within the ANSI C78.377 range), color consistency within 3 SDCM, CRI Ra80 or higher and R9 above 0. Without our written consent, the LED chip maker, model or bin, driver, heat sink, cover material or production factory may not be changed; any change requires prior written notice and retesting. Each lot must be accompanied by an integrating sphere test report stating part number, lot number, test conditions, results and photos, together with chip shipment records and photos of chip reel labels. Test samples are drawn at random by our inspector and may be torn down for inspection. On a failure, additional samples must be drawn and retested; any lot that still fails will be held from shipment.)

10. Three Real Cases

These are anonymized versions of patterns we see often in consultations. Amounts and figures are examples.

Case 1 — LED Ceiling Lights

This is the case from the opening. The distributor took the test results and teardown photos to the factory, but the spec sheet said only "50W 4,000lm 4000K" plus outer dimensions, with no chip maker, measurement conditions or color consistency criterion, so the factory held out with "it's correct according to the chip datasheet." In the end the factory admitted it had used chips different from the approval sample and refunded part of the amount, but with return visits to installation sites and replacement costs, the loss still came to about KRW 31 million. From the next order on, the distributor puts the chip maker, bin codes and flux with the cover fitted into the spec sheet, and lights samples from every lot side by side for comparison.

Case 2 — Track Spotlights for a Café Franchise

A supplier that delivered 600 track lights to a café franchise placed a reorder each time a new store opened, and at the third store it heard "the lighting color is different from our other stores." Measurements showed a color difference of around 7 SDCM between the first and third deliveries, with wide variation within the third delivery itself. The factory had changed chip suppliers, and bin control had slipped along with it. Since then, the supplier keeps the first approval sample as a reference sample, has put bin codes and a 3 SDCM limit into the spec sheet, and compares every reorder side by side with the reference sample.

Case 3 — LED High Bay Lights for a Warehouse

A supplier that delivered 300 high bay lights to a logistics warehouse heard six months after installation that the lights were "clearly dimmer than at first." On-site illuminance readings were nearly 30% below the levels right after installation. Teardown of returned units showed heat sinks 40% lighter than the approval sample, with fewer fins. With the chips running hot continuously, fast degradation was inevitable. The supplier now puts heat sink weight and thickness into the spec sheet and orders new models only after receiving 1,000-hour lumen maintenance test results.

11. Common Mistakes

• Looking only at "50W" power without setting finished-fixture flux with the cover on in numbers
• Checking only that each unit lights up, without lighting several side by side to compare color
• Accepting an LED chip datasheet or safety certification report as an optical report for the finished product
• Writing only a nominal CCT without setting a tolerance range, SDCM or bin codes
• Judging brightness by values measured right at switch-on, without looking at thermal design or lumen maintenance
• Trying to match color with the previous lot on a reorder without a reference sample

12. LED Lumen and Color Control Checklist

Specification Stage

• Set the LED chip maker, model, count and bin codes
• Set flux, power and efficacy for the finished fixture with cover
• Wrote down the CCT range, SDCM and CRI (Ra·R9) criteria
• Wrote down measurement conditions (voltage, temperature, stabilization, number of samples)
• Wrote down driver model, power factor, heat sink weight and lumen maintenance criteria

Sample and Production Stage

• Kept a few approval samples aside as reference samples and tore one down for the record
• Had an approval sample sphere-tested separately at a third-party lab
• Received sphere test records for each lot with photos
• Checked chip shipment records and chip reel labels (bin codes)

Pre-Shipment Inspection Stage

• Lit several units side by side and checked color differences
• Measured power and confirmed it was within tolerance
• Tore down random samples to confirm chips and heat sink match the approval sample
• Sent leftover samples to a lab for sphere testing

Wrapping Up - Check Lumens and Bin Codes, Not One Light at a Time

Taken out of the box and switched on one at a time, LED lights all look alike. The difference only shows once a customer hangs the same fixture in the living room and the bedroom. Writing the chip maker, bin codes, flux with the cover fitted and color consistency into the spec sheet, and checking sphere test results for every lot, doesn't cost much. Put a few hundred thousand won in testing next to tens of millions of won in site revisits and replacements, and it's clear which is cheaper.

There is also something you can do right now. Ask the factory currently producing for you to send the maker and bin codes of the LED chips in use, photos of the chip reel labels, and the integrating sphere report for the latest lot. If the answer is a single line saying "Yes, it's 4,000lm," draw up a lumen and color control table and send it before your next order.

StartMade supports you on the ground with spec sheets covering chip specs and lumen and color criteria, third-party testing of approval samples, chip checks on production lots and pre-shipment inspection. If you have already received brightness or color claims, we start with testing and teardown of remaining stock; if you are preparing a new product, we start with chip selection and setting criteria.