Does Putting a USB Charger in an Aluminum Case Actually Cool It Down?
イチケン / ICHIKEN
Viewers of the creator's USB charger teardown videos kept suggesting that chargers would run cooler in a metal enclosure. To test this, the creator moved the internals of a 65W USB charger into two custom aluminum housings and measured them against the unmodified charger. The short answer was that aluminum did lower the surface temperature, but by less than expected. The creator then used a simple thermal model to explain why, and gave three reasons metal-bodied chargers are rarely sold.
Two Aluminum Enclosures
The creator made two aluminum housings, both CNC-machined through JLCCNC, a machining service run by the video's sponsor.
The first is a simple case shaped like the original charger, with no cooling fins. It was bead-blasted and anodized. The second has as many cooling fins as could fit on the body to increase surface area, and was left as-machined with no surface treatment. The finless case has a surface area of 69.4 cm². The finned case has 341.4 cm², about five times more.
Copper would have been interesting but much more expensive, so the creator chose aluminum. The quoted prices were $33 per unit for the finless case and $100 per unit for the finned one. The creator said the finned design needs a 5-axis machine and called the price surprisingly low.
Moving the Charger Into the Metal Case
The original plastic shell was removed because it would get in the way of heat transfer. Only the internal circuit board was fitted into the aluminum. That creates a safety risk: if high-voltage parts touch the metal housing, the circuit could short or the user could get a shock. To prevent this, the creator lined the entire inside of the aluminum case with Kapton tape.
The creator pointed out that Kapton's thermal conductivity is only about 0.16 W/m·K. The tape therefore adds thermal resistance and hurts heat dissipation, but it was accepted as necessary for safety.
Some of the charger's original thermally conductive silicone was stripped off before the board was inserted. At that point the board and the aluminum were separated by air, so almost no heat could pass between them. To bridge that gap, the creator injected a thermally conductive silicone potting compound, Shin-Etsu Chemical's KE-1285P. It was chosen because it has low viscosity and flows well, which the creator judged important for filling a complex shape. Areas the potting compound should not reach, such as around the USB port, were packed with silicone grease instead.
The potting compound comes in two parts, A and B, mixed in equal amounts. The creator had no proper degassing equipment. Instead, they degassed the mixture using a vacuum-sealed bag, loaded it into a syringe, and used small tubes to steer the flow into narrow gaps. Curing at room temperature would take a very long time, so the assembly was heated on a hot plate kept below roughly 100°C for about an hour. The creator explained that higher temperatures cure the compound faster but could damage the electrolytic capacitors inside.
After curing, the compound had visibly swelled compared with before heating. Cutting into a raised area revealed air bubbles. The creator concluded that degassing had been insufficient and said it should have been done more thoroughly.
Test Setup
All three units were tested at the same time: the stock charger, the finless aluminum version, and the finned aluminum version. Each ran at the charger's maximum rated output of 65W. Surface temperature was measured with a thermocouple attached to each case, and output power was logged alongside it.
Results for Each Charger
Stock charger. Temperature rose as soon as the test began at 65W. After about 30 minutes, the surface reached 80°C and the charger's overheat protection kicked in, reducing output to 55W. The creator said this is the charger's normal behavior. In the graphs, output power jumps between 0W, 65W, and 55W. The creator explained that this happened because they manually adjusted the electronic load to follow each charger's throttling. After throttling began, the temperature slowly fell, since lower output means less power loss and less heat. It settled at about 72°C. Small dips along the way were attributed to the black-body tape on the surface.
Finless aluminum. Again, temperature rose right away, and at about 30 minutes the charger throttled to 55W. The creator noted that the aluminum housing did not prevent throttling and promised to explain why later. The maximum surface temperature was 67°C, nearly 13°C below the plastic original. The temperature barely changed afterward and held steady at 67°C.
Finned aluminum. This unit also throttled around the 30-minute mark, even though its surface was only 54°C at that point. Unlike the others, it did not stay throttled and quickly went back to full 65W output. The creator's explanation after thinking it through was that heat transfer here has a large time constant. The surface stays relatively cool while the interior likely heats up much more, creating a large temperature gap between inside and outside. At 65 minutes it throttled again and again returned immediately to 65W. After that it ran at full 65W with no further throttling. Its maximum surface temperature was 63°C, which the creator took as evidence that the fins were helping considerably.
Comparing the Three
The stock charger had the highest maximum and equilibrium temperatures, the finless aluminum case was next, and the finned case was lowest. The creator stressed an important caveat: the stock and finless units ended up throttled to 55W, while the finned unit ran steadily at 65W. The finned unit was therefore working under the hardest conditions and still ran coolest, which the creator attributed to its much larger surface area.
The creator also noted two things from the spec comparison. First, the finless aluminum case has about the same surface area as the plastic original, yet its maximum surface temperature was 13°C lower. Even so, it still throttled. Second, the finned case has roughly four to five times the surface area of the others, but its surface temperature did not fall as much as the creator had hoped. The creator's view was that with only natural convection, temperatures just don't drop very far. To get real cooling, heat has to be removed actively, and the creator said this experiment confirmed that once again.
Why the Finless Case Still Throttled: A Thermal Circuit Model
The creator went back to the finless result. Its surface was clearly cooler than the plastic original, so why did it still throttle? The creator's hypothesis was that the interior stayed hot, meaning there was a large difference between internal and surface temperatures. Reviewing the build footage, the creator noticed that they had left extra space between the board and the aluminum to make room for the potting compound. They suspected this distance was blocking heat flow.
To test the idea, the creator built a thermal circuit, a simple series of thermal resistances, and emphasized that it relies on several assumptions. On one end is room temperature, 24°C. On the other is the heat-generating circuitry, assumed to be at 100°C. The heat flow Q is 5.2W, which is the charger's own power loss. Between them are three resistances in series: the silicone potting (Rsp), the Kapton tape (Rtp), and the aluminum housing (Ral). Higher thermal resistance means less heat gets through. As an example, the creator said a plastic part would have high resistance and the same shape in aluminum would have low resistance.
From the experimental data, the creator derived an aluminum housing resistance of 8.3°C/W. For the Kapton tape, the creator reasoned that one layer could not create a large temperature drop and assumed a 5°C difference, which gives 0.96°C/W. That puts the boundary between the potting compound and the tape at 72°C. The remaining drop from the assumed 100°C source then gives the potting layer a resistance of 5.4°C/W.
The creator's interpretation was that the resistance between the aluminum housing and the air is significant, but so is the potting layer. No matter how much effort goes into the aluminum, including adding fins, heat is not moving well from the circuit to the case, and so it cannot be dissipated.
The Potting Compound Is the Bottleneck
The creator checked the potting compound's specs again. Its thermal conductivity is 0.8 W/m·K. For comparison, a typical thermal grease for electronics is about 0.84 W/m·K, so this compound transfers heat about as well. The 6061 aluminum used for the housings is about 170 W/m·K.
According to the creator, there are two ways to lower the potting layer's thermal resistance: shorten the distance heat has to travel, or use a compound with higher thermal conductivity. In practice, that means keeping the gap between the circuit's heat sources and the aluminum as small as possible and filling it with no voids. The creator acknowledged this build may not have done that.
Using the model, the creator estimated the heat source temperature under two scenarios. With forced-air cooling that halves the heat sink resistance, the source drops to about 78°C. With natural cooling but an optimized internal structure that halves the potting layer's resistance, it drops to about 86°C. From this comparison, the creator concluded that optimizing the potting-filled region looks like the more cost-effective option. The creator also mentioned that high-conductivity potting compounds exist but tend to be more viscous, which makes filling tight spaces difficult.
Why Metal-Enclosure Chargers Are Rare
The creator gave three reasons commercial chargers stay with plastic: cost, insulation, and contact temperature.
Cost. Metal parts with high thermal conductivity cost more. The CNC-machined cases here were expensive, especially the $100 finned one. The creator was unsure whether casting could produce these shapes, and final shaping would still need CNC work. Overall, a metal enclosure would make the product much more expensive than plastic, which the creator did not see as realistic.
Insulation. Plastic is an insulator, so touching it carries no shock risk. Metal conducts, so if a live part of the circuit touches the housing, a person touching the outside could be shocked. According to the creator, a metal enclosure would require a ground connection, which is impractical for a charger.
Contact temperature. During the tests, the creator touched the metal-cased charger and found it too hot to hold. The stock plastic charger had a higher surface temperature, but could be held for a long time. The creator explained this with a sauna comparison: 100°C air in a sauna is barely bearable, but 100°C water would be unbearable, because air conducts heat poorly and water conducts it well. Similarly, when a finger touches plastic, the contact spot cools and heat from the rest of the case moves in slowly. With metal, heat rushes in from the surrounding material, so it feels hotter.
The creator cited JIS C 62368-1, which sets upper temperature limits for parts touched during operation: 70°C for metal and 87°C for plastic. With natural convection and the same internal heat, the creator said it is hard to lower the temperature enough for a metal case to meet its stricter limit. They believe this is why real products use plastic, which has a more lenient contact-temperature limit.
Conclusion
The creator's verdict was that a metal enclosure does cool a USB charger, but not as much as they expected. Real cooling requires thermal design across the whole device and careful internal construction. To cool a charger properly with a metal case, the surface area has to be increased substantially, for example with fins. Without that, the design needs forced-air cooling or an optimized internal heat path. Given cost, insulation, and contact-temperature limits on top of that, the creator concluded that a metal-bodied USB charger is not very realistic.
Hello. Today, we'll modify a USB charger and verify whether it can cool down. I've been working on USB charger disassembly and verification videos for a while now. I've been getting a lot of comments from viewers saying that fitting a USB charger inside a metal enclosure might make it cooler. So let's actually try it.
We made these metal enclosures. If we put a USB charger in one and run it, how much will the temperature drop, and how much will it help with heat dissipation? We'll verify that. But first, here is today's sponsor.
If you want to make printed circuit boards at a lower cost, JLCPCB is recommended. Lately, they've been working particularly hard on manufacturing multilayer boards. They would like everyone to use six-layer boards! They are making them on a very large scale in China, and prices are quite cheap: 35 for five six-layer 100 × 100 mm boards. Via-in-pad, which is essential for high-density design, has recently become available for free! They also distribute coupons for six-layer boards. We encourage everyone to try different things!
Here are the metal enclosures we made for this project. We'll put the USB charger inside. This time we made two kinds. The first has a simple structure without cooling fins. It really looks just like the charger, doesn't it? We also treated the surface with bead blasting and anodizing.
For the second, we installed as many cooling fins as possible on the main unit to increase the surface area. This one has no special surface treatment; we're leaving it as machined. The smaller one's surface area is 69.4 cm², and the one with cooling fins is 341.4 cm². Just adding these fins gives about five times the surface area.
The material is aluminum. I think it would be interesting to make it with copper, but the price would be quite high, so this time we're using aluminum. All of these were made by CNC machining. We used a CNC cutting service called JLCCNC, run by our sponsor, JLC Group. The one without fins is $33 per piece, and the one with fins is $100 per piece. Especially for the one with fins, it seems you have to use a five-axis machine. I think it's a shockingly low price. What do you think?
From here, we'll work on placing the USB charger inside the aluminum enclosure. We won't leave the plastic enclosure intact because it affects heat transfer. We'll take out only the internal circuitry of the USB charger and put it in here.
However, if we put it in a metal enclosure as it is, the high-voltage parts of the charger circuit could contact the metal housing. The circuit could short, or the person using it could receive an electric shock. So we'll do a little work with Kapton tape.
We'll apply this Kapton tape inside the aluminum housing. The thermal conductivity of the tape is about 0.16 W/(m·K). It increases thermal resistance, which is not good for heat dissipation, but it's for safety reasons, so that can't be helped. We've now covered the entire inside surface with Kapton tape.
We'll put the contents of the charger here. I stripped off some of the thermally conductive silicone originally attached to the charger. The charger fits inside. In this state, the charger and aluminum are not in close contact. There's air in between, so there is no heat transfer at all. Therefore, we'll inject new thermally conductive silicone potting material.
This time, we have Shin-Etsu Chemical's thermally conductive potting material, KE-1285P. For filling complex shapes like this, a low-viscosity potting material seems like a good idea. This potting material has low viscosity and high fluidity, so it seems likely to fill well.
We're going to put in silicone potting material, but there are some places we don't want it to enter, such as the USB port. We'll fill those places with silicone grease.
We'll prepare the potting material. There's an A and a B, and you take equal amounts of each and mix them well. The beaker was a little too small. I don't have equipment for removing bubbles, so I'll use this simple vacuum-packing setup to degas it. I wonder if it's ready.
Then we'll fill the syringe with potting material and fill the USB charger. To get it into the gaps as well, we used small pipes to direct the flow. We've filled it with potting material like this.
Leaving it at room temperature until it hardens would take an awful lot of time, so I'll heat it on a hot plate to make it harden quickly. The hot plate temperature is generally below 100°C. We'll heat it for about an hour. The higher the temperature, the faster it hardens, but if you raise it too high, I'm afraid the electrolytic capacitors inside might be damaged. I'll leave it like this for a while at about 100°C or less.
The silicone potting material has hardened like this. Compared to before heating, the whole thing looks raised. It's possible that insufficient degassing left air bubbles inside that expanded. Looking at the raised part, there are some air bubbles. I guess the potting material wasn't sufficiently degassed. I should have degassed it thoroughly.
Now we'll actually operate it as a charger and check how high the temperature rises. With this setup, we conducted three experiments simultaneously: the stock product, the aluminum housing without fins, and the aluminum housing with fins. The load is set to draw the charger's maximum output of 65 W. Temperature was measured with a thermocouple attached to the surface. We measured the temperature and the USB charger's output power simultaneously.
Here are the measurement results. This is the temperature and output power graph for the stock charger without modifications. The horizontal axis shows the time elapsed since the start of the experiment. The left vertical axis is the charger's surface temperature, and the right vertical axis is its output power.
Immediately after the start of the experiment, the USB charger is operating at 65 W output power. At the same time, its temperature begins to rise. After about 30 minutes, the temperature has risen considerably, and the surface temperature is now 80°C. At this point, the charger's thermal throttling overheat protection activates, reducing its output power to 55 W. This is normal operation for this charger.
For all the chargers this time, we manually adjusted the electronic load to match thermal throttling. As a result, the power goes back and forth between 0 W, 65 W, and 55 W. This is normal.
Once thermal throttling activates, the temperature decreases gradually. Reducing the output power reduces the charger's power loss, meaning it generates less heat. Finally, the surface temperature settled at about 72°C. There are places where the temperature drops during the process. Those are effects of things like applying blackbody tape to the surface.
Next are the results with the charger housed in the aluminum enclosure without fins. Here too, the temperature began to rise immediately after the experiment started. After 30 minutes, the output power drops to 55 W as thermal throttling activates. Simply putting it in an aluminum housing doesn't seem to lower the temperature enough to eliminate thermal throttling. We'll discuss this later.
The maximum temperature is 67°C, so compared with the plastic enclosure, the temperature has dropped nearly 13°C. Even after operating for a long time, there is little or no temperature change. It reaches equilibrium at 67°C.
Next are the results with the charger housed in the aluminum enclosure with fins. Similarly, the temperature began to rise immediately after the experiment started. After 30 minutes, thermal throttling activated the overheat protection. The surface temperature at this time is 54°C. However, it doesn't remain in a thermally throttled state. It immediately returns to operating at the full 65 W.
I've given the reason for this behavior a lot of thought. It may be because the time constant for heat transfer is quite large. There is a large temperature difference between the inside and outside. Although the surface temperature is low, I suspect the temperature inside may have risen considerably.
Continuing to operate it as it is, thermal throttling activated again at 65 minutes. It immediately recovered its 65 W output capability again. After that, no thermal throttling or other problems occurred, and it could continue operating at the full 65 W. The maximum surface temperature is 63°C. You can see that the fins are quite effective at dissipating heat.
This graph summarizes the temperature trends of the three units tested. The maximum temperature and equilibrium temperature are highest for the stock product, followed by the aluminum housing without fins. The aluminum housing with fins has the lowest temperature.
One point to be careful about here is that the stock charger and the finless aluminum enclosure have their output power reduced to 55 W by thermal throttling, while the finned aluminum enclosure operates stably at 65 W. Therefore, the operating conditions are most demanding for the one with fins. Adding fins makes the surface area much larger than the other two, so it dissipates heat well. Looking at it this way, you can see how cooling changes with and without fins and with different enclosure materials.
A comparison in a specifications table looks like this. Looking at surface area, the aluminum enclosure without fins actually hasn't changed that much compared with the stock product. Even so, its maximum surface temperature is 13°C lower. The surface is quite cool, but thermal throttling still occurred.
The aluminum enclosure with fins has about four to five times the surface area of the others, but the surface temperature hasn't dropped as much as I thought it would. In this experiment, heat dissipation was by natural convection, which may explain why the temperature didn't drop that much. After all, you need airflow to carry heat away forcibly; otherwise, the temperature doesn't fall that much. This experiment confirmed that once again.
There were some points in the experiment that I'd like to discuss, so we'll take another look. In the temperature measurement experiment with the aluminum enclosure without fins, we were able to keep the surface temperature lower than with the stock plastic enclosure. I was wondering why thermal throttling still occurred. We believe the internal temperature actually remained high. In other words, the difference between the surface and internal temperatures had become quite large.
This is footage from the assembly. To fill it with potting material, we left a little more distance between the charger and the aluminum. After considering it, we believe this distance may be preventing heat transfer.
This is what we call a thermal circuit. We can make a number of assumptions using this diagram. We'll calculate the thermal resistance, or how difficult it is for heat to pass through the Kapton tape, silicone potting material, and aluminum enclosure used this time.
At the far left is the room temperature of 24°C. At the far right is the temperature of the charger generating heat. In this case, we assume that the charger's circuit is at 100°C. The heat emitted by that heat source is expressed as Q, which is 5.2 W here. This 5.2 W is the power loss emitted by the charger itself—in other words, heat. That heat, Q, passes through the thermal resistances to the room-temperature surroundings. That's what the diagram shows.
R_SP is the thermal resistance of the silicone potting material, R_TP is the thermal resistance of the Kapton tape, and R_Al is the thermal resistance of this aluminum housing. The greater the thermal resistance, the less heat is transferred. For example, with the same shape, plastic has high thermal resistance, while aluminum has low thermal resistance. Let's use this to think it through.
First, from the experimental results, we can determine the thermal resistance of the aluminum housing: 8.3°C/W. Next is the Kapton tape. The idea is that a single layer of tape wouldn't create that much of a temperature difference, so we assume a temperature difference of 5°C. That gives the Kapton tape a thermal resistance of 0.96°C/W.
Then we know the temperature where the silicone potting material contacts the Kapton tape. Since the temperature difference assumed earlier is 5°C, the temperature at that contact point is 72°C. We can then determine the thermal resistance of the silicone potting material, which is 5.4°C/W.
Looking at it this way, the thermal resistance between the aluminum housing and the air is a dominant factor, but the silicone potting material also has a major influence on heat dissipation. Therefore, no matter how much effort you put into the aluminum housing, adding fins and trying to dissipate heat, if the heat isn't being transferred properly inside, it cannot be dissipated.
Let's carefully check the specifications of the silicone potting material used in this project again. Its thermal conductivity is 0.8 W/(m·K). The higher this number, the better the heat transfer. A common thermal grease applied to electronic components has a thermal conductivity of 0.84 W/(m·K), so the potting material used here transfers about as much heat. For reference, the aluminum used in this project is 6061, with a thermal conductivity of about 170 W/(m·K).
To lower the thermal resistance of the potting material, you basically need to shorten the distance over which heat is transferred or use potting material with higher thermal conductivity. In this case, to transfer heat effectively to the outside, we needed to make the distance between the heat-generating part of the circuit and the aluminum enclosure as short as possible and fill it with potting material without gaps. We may not have achieved that this time.
These are estimated heat-source temperatures with forced-air cooling and with optimization of the silicone potting material section. For forced-air cooling, assuming the heat sink's thermal resistance is halved, the heat-source temperature drops to 78°C. With an optimized internal structure and natural air cooling, suppose the thermal resistance of the potting material could be halved. The heat-source temperature drops to 86°C. Looking at it this way, optimizing the structure of the area filled with potting material seems more cost-effective.
Potting materials with high thermal conductivity do exist, but they tend to be more viscous. Filling small spaces with them is quite a challenge!
Now let's consider why metal enclosures aren't used in actual products. I have about three reasons: cost, insulation, and contact temperature.
First, let's talk about cost. Using metal parts with high thermal conductivity naturally increases costs. In particular, we used machining this time, so the cost was high. The one with fins costs $100 each. I'm not sure whether the metal enclosures could be made by casting or similar methods, but final shaping would require CNC machining. The cost would still be high. Looking at it this way, metal enclosures make the cost rise dramatically compared with plastic enclosures, so they don't seem like a very realistic option.
Next, let's talk about insulation. Consider a plastic enclosure. Naturally, even if people touch it, plastic is an insulator.
There is no current flow and no electric shock. On the other hand, the metal enclosure is a conductor. When a part of the electrical circuit carrying voltage is in contact with the metal enclosure, there is a possibility of electric shock when you touch the metal part of this exterior. If a metal enclosure is used, I need to connect a ground wire for grounding. It doesn't look like a very practical design for a charger.
And finally, contact temperature. During this experiment, I touched the charger with a metal housing. In fact, it feels so hot I can't touch it. On the other hand, the unmodified plastic enclosure has a high surface temperature, but I could touch it for a long time. This is due to the difference in thermal conductivity of the materials.
For example, if you go into a sauna with a temperature of 100°C, it's just barely tolerable heat. Going into 100°C water would be outrageous. This is due to the low thermal conductivity of air: it is difficult to transfer heat. Water, on the other hand, has high thermal conductivity and conveys heat well. For chargers, too, comparing metal with plastic, plastic has a lower thermal conductivity. It doesn't feel that hot to the touch.
If you hold the two like this, the plastic one has a low thermal conductivity. The temperature drops where I touch it, and it is difficult for heat from other parts of the enclosure to be transferred. On the other hand, what about the metal? When the temperature of the touched area drops, heat can be transferred from other parts of the enclosure very quickly. That's why it feels hot.
These temperature limits for each material are determined by JIS C 62368-1. The upper temperature limit for the parts that people touch for operation is 70°C for metals and 87°C for plastic.
Even if a metal enclosure is used, it is naturally air-cooled, and the amount of heat generated inside has not changed, so it's actually quite difficult to lower the temperature. Therefore, I think actual products use plastic enclosures with looser contact temperature limits because metal cannot be kept within the contact temperature range.
So today I've been testing to see if a USB charger can be cooled down by putting it inside a metal enclosure. The conclusion is that it does get cooler, but not as cool as I thought it would. The internal structure must be well thought out. Thermal design must be done for the entire equipment.
If the metal enclosure were to be cooled properly, the surface area of the body would need to be increased with fins, etc. If the surface area were not increased, forced air cooling or optimization of internal heat transfer would be required. Then there's the cost, insulation, contact temperature, etc. to consider. I thought a USB charger with a metal housing was not very realistic.
So, if you found this video helpful, please give us a high rating and subscribe to our channel! If you have any concerns, please contact us. Please try to leave a comment as well! Thank you for watching the video until the end.
Article published · Updated