Marshe
In January 2026, a street lighting contractor in Qatar replaced about 500 modules, each rated 120W. The lens design was sound — but the lenses yellowed and the modules burned out.
The cause was thermal, not electrical. The heat sink was 200x60x37mm, while the standard 50-60W module in the same family already uses 300x70x40mm. In Gulf summer ambient, 120W on that heat sink has no margin.
What the file did not contain: a rated ambient temperature, a measured Tc, a calculated Tj. Three numbers that would have stopped the order before it shipped. That is the point of this guide. A specification is only worth what the test conditions behind it are worth.
We build custom LED optics and PCB modules in Shenzhen and ship to lighting contractors in more than 60 countries. Below are the seven specifications we would want a buyer to challenge us on.
"Efficiency" is not one number. It is three, and a serious datasheet keeps them apart.
A realistic chain looks like this: about 160 lm/W at package level, about 140 at module level, about 125 at luminaire level. Each step is lower than the one before. If a sheet shows a module number above its package number, the two numbers were not measured on the same product, or one of them is not what the label says.
That is also how to read a 200 lm/W claim. It is not a lie by default. It is a package-level number at a defined current and temperature, usually a top bin. Ask which level, which current, which temperature, and ask for the LM-80 report behind it. Then compare only numbers from the same level.
On our own sheets the numbers come from different products and different LED generations, so they must not be chained: the 30W SKD kit uses R3 LEDs rated 122 lm/W at package level, while the 50W module is rated 130-140 lm/W at module level with PH 3030 chips. Different SKU, different era, different boundary. Ask any supplier for all three levels on the same SKU, and the comparison becomes simple.
Beam angle tells you how wide the light spreads. It does not tell you whether the road passes the lighting class.
What actually gets checked in a project:
Two modules with the same wattage and the same lumens can produce very different Uo and glare. If a supplier cannot send the IES file for the exact model, you are buying a picture, not a photometric product.
Both materials work. The mistake is treating one as premium and the other as cheap.
| Property | Optical PMMA | Optical PC |
|---|---|---|
| Transmittance | About 92% (our 12W array lens is measured at 93%) | Typically 88-90%; the best UV-stabilized optical grades reach about 90-92% |
| UV stability | Excellent; low yellowing over time | Good only when UV-stabilized; unstabilized grades yellow under strong sun |
| Heat resistance | Lower; usable range roughly 80-100°C, so keep it away from the hottest zone | Higher; roughly 120-130°C, better for high-power and high-ambient designs |
| Mechanical strength | Brittle; can crack under impact or uneven screw torque | Tough; resists impact and vibration |
| Chemical contact | Stress cracks in contact with plasticizer-containing PVC gaskets or wires | More tolerant, but still check compatibility |
Practical rules we give customers:
One number here deserves a second look. Transmittance moves with the measurement: a 2 mm molded sample reads differently from the finished lens geometry, and it shifts again with thickness and surface finish. PMMA at 92–93 % is normal. Optical PC at 93 % is not — so when a PC lens is quoted at 93 % or more, ask for the measured report on that exact grade and thickness. This is the first number an experienced buyer re-tests.
Two statements you will hear are too simple to be useful: "always use thermal pads, never grease" and "1.6mm aluminium is the minimum".
What actually matters is the thermal path and the measured result:
Then ask for the documents that prove it. ISTMT gives you Tc measured at a declared test ambient, usually 25°C, in a stated mounting condition. It is not a site measurement. If the site runs at 45°C or 50°C, the Tc you see in the field is higher, so ask for the derating calculation from the ISTMT point to your ambient, and the resulting Tj.
LM-80 describes the LED package at a fixed case temperature over the tested hours. It says nothing about the module. TM-21 turns LM-80 data into an L70 or L90 projection. LM-82 describes how electrical and photometric characteristics change with temperature. LM-84 covers lumen and colour maintenance at module and luminaire level, and TM-28 projects it.
IP66 per IEC 60529 means dust tight and protected against strong water jets. It is a result of a test on a specific design, not a property of a material.
Ask about the parts that decide it:
The driver is usually the first part to fail, and it is the part buyers negotiate hardest on price. That is a bad combination.
Engineers evaluating a bid look at color and lifetime before they look at price.
| Market | Documents to request |
|---|---|
| EU | CE declaration (EN 60598, EN 62031, EN 62471), RoHS, IP test report, IK report |
| North America | UL or ETL listing for the luminaire, DLC qualification if the project needs rebates, LM-79 and LM-80 files |
| Middle East / coastal | Salt spray report (ISO 9227), higher IP verification, UV-stabilized optics |
| All markets | IES file for the exact model, ISTMT, TM-21 projection, material declarations |
Copy this into your RFQ. It removes most of the room for different interpretations.
A 30W SKD kit with lens, aluminium PCB, gasket and driver sits in a different price band from a bare lens, and a finished luminaire sits in a third band. Quoting them against each other is where buyers lose money.
Compare per watt and per lumen at a defined boundary, and add the cost of what is missing: heat sink, driver, housing, assembly, testing, freight. A quote that looks 20% cheaper often has 20% of the content removed.
Is a 200 lm/W claim a red flag? Not by itself. Ask at what current, at what temperature, and at which boundary. Then ask for the LM-79 or LM-80 report behind it.
PMMA or PC for street light lenses? PMMA for maximum transmittance and UV stability in a cooler, sealed design with a silicone gasket. UV-stabilized PC for higher heat, impact or coastal conditions. Both are legitimate; the failure comes from mismatching material and environment.
Is 1.6mm aluminium PCB enough? It depends on copper thickness, dielectric conductivity and the wattage per area. Ask for the calculated Tj and the measured Tc, not a board thickness rule.
Thermal pad or grease? Judge by interface thermal resistance, bond line thickness and long-term stability. Pads are more consistent in production; grease performs well when the process is controlled.
Is a three-year warranty enough? For a component, it is common. For a road tender or a DLC-listed project, expect five years and confirm that labor and freight are covered.
Seven specifications, each with a document behind it. That is the difference between a datasheet and a product you can defend at handover.
Sunshineopto has been designing and manufacturing custom LED optics and PCB modules in Shenzhen since 2011, with free LED PCB Booard design on every enquiry.
Please tell me the application environment and relevant requirements for your lighting fixtures.
We will send back a specification proposal with the reports attached, usually within one working day.
Marshe
In January 2026, a street lighting contractor in Qatar replaced about 500 modules, each rated 120W. The lens design was sound — but the lenses yellowed and the modules burned out.
The cause was thermal, not electrical. The heat sink was 200x60x37mm, while the standard 50-60W module in the same family already uses 300x70x40mm. In Gulf summer ambient, 120W on that heat sink has no margin.
What the file did not contain: a rated ambient temperature, a measured Tc, a calculated Tj. Three numbers that would have stopped the order before it shipped. That is the point of this guide. A specification is only worth what the test conditions behind it are worth.
We build custom LED optics and PCB modules in Shenzhen and ship to lighting contractors in more than 60 countries. Below are the seven specifications we would want a buyer to challenge us on.
"Efficiency" is not one number. It is three, and a serious datasheet keeps them apart.
A realistic chain looks like this: about 160 lm/W at package level, about 140 at module level, about 125 at luminaire level. Each step is lower than the one before. If a sheet shows a module number above its package number, the two numbers were not measured on the same product, or one of them is not what the label says.
That is also how to read a 200 lm/W claim. It is not a lie by default. It is a package-level number at a defined current and temperature, usually a top bin. Ask which level, which current, which temperature, and ask for the LM-80 report behind it. Then compare only numbers from the same level.
On our own sheets the numbers come from different products and different LED generations, so they must not be chained: the 30W SKD kit uses R3 LEDs rated 122 lm/W at package level, while the 50W module is rated 130-140 lm/W at module level with PH 3030 chips. Different SKU, different era, different boundary. Ask any supplier for all three levels on the same SKU, and the comparison becomes simple.
Beam angle tells you how wide the light spreads. It does not tell you whether the road passes the lighting class.
What actually gets checked in a project:
Two modules with the same wattage and the same lumens can produce very different Uo and glare. If a supplier cannot send the IES file for the exact model, you are buying a picture, not a photometric product.
Both materials work. The mistake is treating one as premium and the other as cheap.
| Property | Optical PMMA | Optical PC |
|---|---|---|
| Transmittance | About 92% (our 12W array lens is measured at 93%) | Typically 88-90%; the best UV-stabilized optical grades reach about 90-92% |
| UV stability | Excellent; low yellowing over time | Good only when UV-stabilized; unstabilized grades yellow under strong sun |
| Heat resistance | Lower; usable range roughly 80-100°C, so keep it away from the hottest zone | Higher; roughly 120-130°C, better for high-power and high-ambient designs |
| Mechanical strength | Brittle; can crack under impact or uneven screw torque | Tough; resists impact and vibration |
| Chemical contact | Stress cracks in contact with plasticizer-containing PVC gaskets or wires | More tolerant, but still check compatibility |
Practical rules we give customers:
One number here deserves a second look. Transmittance moves with the measurement: a 2 mm molded sample reads differently from the finished lens geometry, and it shifts again with thickness and surface finish. PMMA at 92–93 % is normal. Optical PC at 93 % is not — so when a PC lens is quoted at 93 % or more, ask for the measured report on that exact grade and thickness. This is the first number an experienced buyer re-tests.
Two statements you will hear are too simple to be useful: "always use thermal pads, never grease" and "1.6mm aluminium is the minimum".
What actually matters is the thermal path and the measured result:
Then ask for the documents that prove it. ISTMT gives you Tc measured at a declared test ambient, usually 25°C, in a stated mounting condition. It is not a site measurement. If the site runs at 45°C or 50°C, the Tc you see in the field is higher, so ask for the derating calculation from the ISTMT point to your ambient, and the resulting Tj.
LM-80 describes the LED package at a fixed case temperature over the tested hours. It says nothing about the module. TM-21 turns LM-80 data into an L70 or L90 projection. LM-82 describes how electrical and photometric characteristics change with temperature. LM-84 covers lumen and colour maintenance at module and luminaire level, and TM-28 projects it.
IP66 per IEC 60529 means dust tight and protected against strong water jets. It is a result of a test on a specific design, not a property of a material.
Ask about the parts that decide it:
The driver is usually the first part to fail, and it is the part buyers negotiate hardest on price. That is a bad combination.
Engineers evaluating a bid look at color and lifetime before they look at price.
| Market | Documents to request |
|---|---|
| EU | CE declaration (EN 60598, EN 62031, EN 62471), RoHS, IP test report, IK report |
| North America | UL or ETL listing for the luminaire, DLC qualification if the project needs rebates, LM-79 and LM-80 files |
| Middle East / coastal | Salt spray report (ISO 9227), higher IP verification, UV-stabilized optics |
| All markets | IES file for the exact model, ISTMT, TM-21 projection, material declarations |
Copy this into your RFQ. It removes most of the room for different interpretations.
A 30W SKD kit with lens, aluminium PCB, gasket and driver sits in a different price band from a bare lens, and a finished luminaire sits in a third band. Quoting them against each other is where buyers lose money.
Compare per watt and per lumen at a defined boundary, and add the cost of what is missing: heat sink, driver, housing, assembly, testing, freight. A quote that looks 20% cheaper often has 20% of the content removed.
Is a 200 lm/W claim a red flag? Not by itself. Ask at what current, at what temperature, and at which boundary. Then ask for the LM-79 or LM-80 report behind it.
PMMA or PC for street light lenses? PMMA for maximum transmittance and UV stability in a cooler, sealed design with a silicone gasket. UV-stabilized PC for higher heat, impact or coastal conditions. Both are legitimate; the failure comes from mismatching material and environment.
Is 1.6mm aluminium PCB enough? It depends on copper thickness, dielectric conductivity and the wattage per area. Ask for the calculated Tj and the measured Tc, not a board thickness rule.
Thermal pad or grease? Judge by interface thermal resistance, bond line thickness and long-term stability. Pads are more consistent in production; grease performs well when the process is controlled.
Is a three-year warranty enough? For a component, it is common. For a road tender or a DLC-listed project, expect five years and confirm that labor and freight are covered.
Seven specifications, each with a document behind it. That is the difference between a datasheet and a product you can defend at handover.
Sunshineopto has been designing and manufacturing custom LED optics and PCB modules in Shenzhen since 2011, with free LED PCB Booard design on every enquiry.
Please tell me the application environment and relevant requirements for your lighting fixtures.
We will send back a specification proposal with the reports attached, usually within one working day.
