Thermal measurements of LED strips. Conclusions
We thank you for your attention!
I have measured the LED strip temperatures in different strip placements.
1. LED strip suspended in the air.
Air without motion (limited volume) has low thermal conductivity. But when air moves, the situation changes significantly. Air can pick up and take away a lot of heat, and this effect is used everywhere. For example, in computers, a fan is constantly blowing cold air through the processor's heat sink. This cold air takes excess heat from the processor and gets hot. And then the hot air is blown outside.
The air moves near a hot object even without a fan. It is convection - warm air rises, and cold air falls. The LED chips are cooled by air convection in this placement. The air passing near the chips heats up and rises, taking heat from the chips.
For this LED strip cooling variant:
(pros) easiest. We don't need any additional materials.
(cons) Significant heating. LED strip bends from its gravity and does not protect against mechanical influences (easy to damage), dust, or water droplets.
2. Strip in the groove in the chipboard.
The LED chips are cooled by air convection and heat transfer to the chipboard.
For this LED strip cooling variant:
(pros) simple - the tape is glued to the furniture element (for example, under the cabinet).
(cons) Significant heating. LED chips are not protected from mechanical influences, dust, and water droplets.
Learn more about this experiment on a special page
3. Strip in the groove in the chipboard and covered with a diffuser.
The LED chips are cooled primarily by heat transfer from the chips to the chipboard. The air convection is minimal because no cold air enters the LED chips.
For this LED strip cooling variant:
(pros) simple - the strip is glued to the furniture element (for example, under the cabinet), and the diffuser is inserted into a slot in the chipboard. LED chips are protected from mechanical influences, dust, and water drops.
(cons) Extremely high heat.
Learn more about this experiment on a special page
4. LED tape on an aluminum profile in the air.
Aluminum has a high thermal conductivity. This material quickly picks up heat from the LED chips (through the metal base of the tape with high thermal conductivity) and distributes this heat evenly over its volume. This design greatly increases the air's contact area and convection (see point 1). More convection means more cooling.
For this LED strip cooling variant:
(pros) Efficient - LED chips are very well-cooled. The aluminum profile is an excellent housing, protecting the LED chips from mechanical damage. Also, the profile can be any shape and beautiful color.
(cons) LED chips are not protected against mechanical damage, dust, and water drops.
Learn more about this experiment on a special page
5. The strip on the aluminum profile in a groove in the chipboard without a diffuser.
I simulate a light fixture under a chipboard cabinet (but without a diffuser) to understand the effect of air convection.
(pros) Efficient - LED chips are well-cooled.
(cons) LED chips are not protected against mechanical damage, dust, and water drops.
Learn more about this experiment on a special page
6. The strip on the aluminum profile in a groove in the chipboard with a diffuser.
I simulate the light fixture under the chipboard cabinet entirely. The air convection is minimal because the diffuser blocks the cold air to the hot LED chips.
(pros) LED chips are well protected against mechanical damage, dust, and water droplets.
(cons) None
Learn more about this experiment on a special page
7. Strip on extruded polystyrene (XPS).
The task of this test model is to estimate the strip heating without heat dissipation in the chipboard and with minimal air convection. This data will be helpful for comparison with other data.
Learn more about this experiment on a special page
How does the power consumption affect the LED chip heating?
Modern LED chips have an efficiency of 35-45% for the top manufacturers (Philips, Cree) and an efficiency of 10-45% for the available range on the market. This parameter means "how much energy is converted into useful light". The rest of the electricity goes to useless and harmful heating.
Electricity consumption is measured in Watts or Volts*Amps.
In a previous experiment I checked the real power consumption of different LED strips. (learn more)
I calculated how much heat (in degrees) per watt the samples produced. My sample strips have different designs, characteristics, and LED chip efficiencies. So I calculated the average values and deviations from them for each sample to find commonalities for all samples.
The values for "Sample 3 (COB 5/32)" are strange and way out of line with the average. I am still trying to figure out the reason. I measured all values again for this sample and got the same result.
Therefore, I removed this sample from the list and got a deviation of less than 22%. This deviation value allows me to speak about the objectivity of my experiments and each heat transfer method. My conclusions can be applied to many other LED strips of similar power (based on the repeatability and similarity of my several results for different strip types).
LED strip cooling efficiency in different ways (within the limits of my experiments)
I've sorted my experiments and cooling methods by effectiveness.
Why is cooling important? Because at temperatures above 180-185 °F the LED chips and phosphors work under unfavorable conditions and degrade quickly. The lifetime of such a device decreases. For example, at less than 180 °F the average lifetime will be 20000-50000 hours, but at 200 °F the average lifetime will be 5000-10000 hours. The lifetime may be reduced to 1000 hours because factory errors in manufacturing cannot be excluded.
1st place (the most efficient way / experiment with the lowest chip temperature) = LED chips on Al profile in the air.
2nd place = chips on Al profile in the groove in chipboard w/o diffuser
3rd place = chips on Al profile in the groove in chipboard w diffuser
4th place = chips in the AIR
5th place = chips in the groove in chipboard w/o diffuser (4th and 5th places are almost the same)
6th place = chips in the groove in chipboard w diffuser
7th place = chips on XPS (the most inappropriate way - don't do that!!!!)
Aluminum-cased pendant luminaires have the most efficient cooling. Also, traditional aluminum-cased luminaires mounted to wooden cabinets work well.
As soon as the heat dissipation is reduced, such a luminaire is doomed to die quickly.
Can I mount the LED strip on the chipboard directly without the aluminum profile?
Yes, you can if the strip does not have a high power consumption. Experiment 3 showed sufficient cooling for LED strips (with the diffuser) of less than 6 Watts per foot. To be more sure - reduce this to 5 Watts per foot.
How much heat does the chipboard on the other side (inside the cabinet) get? Not much - no more than 100 °F.
Why is aluminum housing often used for luminaires and rarely plastic housing?
The aluminum profile has three significant main advantages and another slight main advantage.
1. Soft and rigid. Linear (in-line) housings of almost any cross-section are easy to make from aluminum.
2. Aluminum is very thermally conductive. The material conducts heat energy quickly. It easily picks up heat from warm/hot objects and gives that heat to cold objects. Aluminum is very useful as a heat sink.
3. Cheap and common. Aluminum is very common. Aluminum is cheap. The only plastic is more affordable.
4. Beautiful. Aluminum is easy to paint, and the anodized outer layer protects the aluminum from damage.
Aluminum is also easy to saw/cut, making it easier for installers and designers. Aluminum is also metal and conducts electricity, which can also be used to create fixtures. And also, and also ...
Why use plastic housing anyway? Because plastic housing is cheaper than aluminum housing.
My experiments 5, 6, 7 showed the high usefulness of aluminum in cooling LED strips and chips. The temperature of the LED chips is reduced by 30 degrees or more.
Here is what it looks like in the temperature comparison:
Is COB technology really more efficient at dissipating heat from LED chips than traditional SMD technology?
Yes! The direct contact of the chip with the strip base ensures direct heat transfer from the chip to the copper base of the strip with a continued transfer of this heat to the external environment. This design perfectly reduces the overall temperature of the chip and strip.
This effect is poorly visible in the tables but well visible in the graphs.
Experiments 1, 2, and 3.
Please note. The electrical power consumption of sample 1 (COB) is greater than that of samples 5 and 7 (SMD). More power means more electricity is used for heating. But the temperature of the COB sample is less than the SMD sample.
Additionally, I did some calculations.
I measured the strip base temperature in air (experiment 1). This experiment's primary heat sink for the LED strips is air (convection).
And in experiment 2, I placed the strips in a groove in the chipboard and measured the temperature of the back side of the chipboard in this area. The primary heat sink for the strips is the chipboard in this experiment.
I calculated the temperature difference at the back of the chipboard and the strip base in the air.
The temperature delta nature of COB strips is quite different from SMD strips. The heat is transferred mainly through the base of the COB strips.
Why? The LED chips are soldered directly onto the base metal with high thermal conductivity. The heat from the chips goes immediately into the material with lower thermal resistance. On the other chip side is silicon with a phosphor with a much higher thermal resistance and thermal conductivity - 2800 times more (Silicone = 0.081 and Copper = 228 in the table (link, new tab).
Cooling through the strip base (not through convection) for COB strips is much more effective in prefabricated/assembled luminaires with housing and a diffuser.
