Your system came with that additional copper sheet and it sits between the CPU and the copper heat plate?
That's neat. I'm glad they're working on improving the thermals, even if it is in a weird way.
Topton Jasper Lake Quad i225V Mini PC Report
- Thread starter ReturnedSword
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
That's neat. I'm glad they're working on improving the thermals, even if it is in a weird way.
Yes and Yes ;-) factory mounted, as posted earlier the thermals are really great 31 degrees in a closed cabinet.
If you do some quick maths, the N100 still outperforms the N5105 in "calculation power per watt".. Use the cinebench R20 with N5105 at 233 and the N100 at 352 and add the power consumption with N5105:26W and N100 = 34W for this:
(352/233) * (26/34) = 1.15 or around 15% more performance per watt under full load. This is nice.
Now I am just unsure if I want to become an "early adopter" and wait for a new n100 4xi226 nic mini router with potentially the same issues the N5105 had in V1, V2 or if I can wait for the N100 in V3, V4...
They should make the copper the correct thickness, adding a plate is trying to correct the mistake and it's much better than leaving a gap.
Last edited:
There shouldn't be any block, copper or otherwise, it should be part of the lid and just CNC milled into the correct level like this Raspberry Pi 4 case.
Even that specific example is not perfect, as those PI blocks are usually hollow
But what you describe, would be the correct solution.
Are they hollow? they wouldn't be, that would make the casting nearly impossible.Even that specific example is not perfect, as those blocks are hollow
But what you describe, would be the correct solution.
... but let's go with the idea that they are hollow, they would in fact be better at transferring heat because air isn't a good conductor of heat and the heat would walk up the thin sides to the next closest thing of a lower heat value, AKA the top. This would occur until the top thick section was saturated with heat at which point 'cooling' (transference of energy) would slow down.
No that's wrong.
The thicker the material, the more time it takes to transfer the same amount of heat.
We are not talking about thicker material, we are comparing same thickness material, both have the same distance from A to B, but one is hollow, the other solid.
What you are saying is that a hollow cube of copper is a better thermal conductor than a solid cube of copper.
Same size, same thickness, same A to B distance.
Let's say you want to insulate a house (contrary of conducting heat), would you use hollow metal walls or solid metal walls?
The wall thickness is the same, but metal is a better thermal conductor than air.
Regarding the hollow box I saw, probably wasn't the one in your photo, I can't see the other side.
Last edited:
No, they have different thicknesses. Your terminology is wrong, you are saying they have the same O.D. and that is correct, but the thickness or internal depth of the alloy (which we are talking about) is greater in a solid piece.
Take our example of a hollow block, in calculating the mass of the alloy we would need to know both the external measurements and the wall thickness of the material (for a hollow one).
In this use case, we don't want it to heat up something with a large mass, we want to create the largest surface area to dissipate (transfer) the heat (energy) and we want to do it as fast as possible. We definitely don't want something with a large mass storing energy/heat so close to the area that we are trying to get it away from.
A hollow pipe/block/whatever will move more heat through the item and it will do it faster than a solid one and thus it is "better".
Yes, it is a better pathway to the larger heatsink area, in this case, the top of the case. It would be better because it would be a faster pathway.
This is all basic thermodynamics.
Here you go, you can have a look here;
Geekworm Raspberry Pi 4 Heavy-duty Aluminum Passive Cooling Metal Case (P173)
This Raspberry Pi 4 Aluminum Case Pi 4B Heat Dissipation passve cooling metal case housing is specially designed for the latest Raspberry Pi 4 Model B computer, it is compatible with Raspbery Pi 4B 1GB 2GB 4GB 8GB version. It is heavy dute raspberry pi 4 case with case weight 184g which can...
geekworm.com
That's why a SOLID piece of copper/aluminium would dissipate heat faster.
I have the Icybox, way better than the Geekworm's one, and the alumium pipes are solid. If they were hollow, they would heat up quickly and unable to dissipate the heat in due time, they would saturate in a heartbeat.Here you go, you can have a look here;
Geekworm Raspberry Pi 4 Heavy-duty Aluminum Passive Cooling Metal Case (P173)
This Raspberry Pi 4 Aluminum Case Pi 4B Heat Dissipation passve cooling metal case housing is specially designed for the latest Raspberry Pi 4 Model B computer, it is compatible with Raspbery Pi 4B 1GB 2GB 4GB 8GB version. It is heavy dute raspberry pi 4 case with case weight 184g which can...
Copper is expensive, and at least those factories in China are including extra layers of it with their shipments due to their not-so-refined manufacturing process, so kudos to them.
No, it has more mass and therefore would take more energy to heat up, correspondingly holding more heat and taking longer to dissipate the heat.
I don't have any intention to teach you physics here, but again: an hollow solid would saturate with heat too quickly than it can dissipate. Solids are used so they can take a lot of heat and dissipate it quickly enough, being actively or passively cooled.
You don't have to trust me, you just have to look around you: thermaltake products would be a good starting point.
@dazagrt
I'm sure we all know that heat transfer is directly proportional to the surface area size in contact to heat.
Let's say your hollow cube is soldered to the die on bottom side and to the case on top side.
So bottom side is closer to cpu temp. and top side is close to box temp.
How heat goes from bottom side to top side of the cube, it climbs the thin walls and the thin walls have a very small total area of contact with heat in bottom side and the same small area of contact to top cold side.
Now increase the thickness of the walls, the total area of contact increases, therefore heat transfer increases.
Increase it the max you can and you have the best case scenario a solid cube.
I'm sure we all know that heat transfer is directly proportional to the surface area size in contact to heat.
Let's say your hollow cube is soldered to the die on bottom side and to the case on top side.
So bottom side is closer to cpu temp. and top side is close to box temp.
How heat goes from bottom side to top side of the cube, it climbs the thin walls and the thin walls have a very small total area of contact with heat in bottom side and the same small area of contact to top cold side.
Now increase the thickness of the walls, the total area of contact increases, therefore heat transfer increases.
Increase it the max you can and you have the best case scenario a solid cube.
Last edited:
Not relevant to the topic or discussion, but this was one of the hollow block cases for PI I saw:
FLIRC Raspberry Pi 4 Case
FLIRC Raspberry Pi 4 Case
It wouldn't be realistic to expect full CNC milled cases at current price point. So we get much cheaper extrusion molding (with CNCed hole for copper block) and it makes impossible to have extra bumps
Also I'm pretty sure RasPI case is just mold casting instead of CNC.
So my personal preference is ~200$ mini PC with "good enough" extrusion molded case vs 300$+ mini PC that runs ~5 to 10 C cooler.
I think the simpler way would be to add 4 screw holes (or even just two, on each long end) around the CPU like a mounting plate would be, then attach the standoffs on the heatsink/top of the case. Let the pressure on the CPU be put on by those screws.
BTW with just an off the cuff janky solution like some object that goes between the bottom plate and the empty space behind the CPU, which has some give (like rubber), so that pressure is applied by screwing the bottom plate, I was able to reduce temps so that the CPU does not throttle at all, even running Prime95 for an hour and a half. Where the original test throttled at 90°C after about 30-35 minutes, this one did not go above 65°C, meaning that this smallest heatsink can dissipate about 18W (CPU package power reading from HWInfo, monitor at the outlet was reading 20.7W or so) easily, only if it is done efficiently.
BTW with just an off the cuff janky solution like some object that goes between the bottom plate and the empty space behind the CPU, which has some give (like rubber), so that pressure is applied by screwing the bottom plate, I was able to reduce temps so that the CPU does not throttle at all, even running Prime95 for an hour and a half. Where the original test throttled at 90°C after about 30-35 minutes, this one did not go above 65°C, meaning that this smallest heatsink can dissipate about 18W (CPU package power reading from HWInfo, monitor at the outlet was reading 20.7W or so) easily, only if it is done efficiently.
Last edited: