MSI Plays Show And Tell With Motherboard M.2 Cooling Solution

Status
Not open for further replies.
Others have, and have dealt with it in various ways. Gigabyte's X99 Designare has a similar system with the M.2 under an aluminum plate. Asus X99 RoG uses the extended ATX size to place it on the lead edge by the SATA ports, away from most GPU exhaust. ASRock sometimes places the M.2 above the PCIe slots. MSI just seems to want to make it somewhat standard on their boards.
 

mpdahaxing

Honorable
Nov 29, 2012
43
0
10,540
I'm glad I have a case whose side panel allows for a 3x3 array of 12cm fans. Blowing air directly onto the components and using deltas as exhaust to get the heat out.
 

Brian_R170

Honorable
Jun 24, 2014
288
2
10,785
Unfortunately, it doesn't help dual-sided boards. Luckily, there aren't many dual-sided M.2 drives available, yet.

Does anybody make a kit that includes a PCIe-to-M.2 adapter card and a stick-on heatsink that you can use with any M.2 SSD? I've seen these bundled with a specific M.2 SSD, but not sold as separate kits.
 
Don't think you'd need an additional heatsink if you put the M.2 on a PCIe card ( and yes, you can get those adapters ). This is more to shield the M.2 from the waste heat of the GPU, under which they normally sit. Put it on its own card and now the drive is inline with most front case fans so it should get a healthy dose of airflow.
 
It doesn't take much to cool most PC components ... as we saw with EVGA's VRMs on the 1070 / 1080 cards, chip temps can be drastically lowered rather simply. Adding a simple backplate to a GPU with proper TIM or thermal pad can do wonders for VRM temps. And like most other comparisons, what thermal pad and what heat sink material matters.

But I do have one issue with related to questions unanswered. When I see an article entitled, "MSI Plays Show And Tell With Motherboard M.2 Cooling Solution" I expected to see the "show and tell part". Is a followup forthcoming where the Mister Miaggi "Heatsink On / Heatsink Off" test will be performed to show / tell us the results ?
 


I don't get it.
 

bit_user

Polypheme
Ambassador
I hope you enjoy cleaning dust out of your case, because you'll be doing it a lot.
 

bit_user

Polypheme
Ambassador
If the goal is to shield it from the GPU, then it should be a thermal insulator and have some other exhaust port to remove heat from the SSD. If you make it thermally conductive, and the heat coming off the GPU is greater, then this would transfer the heat into the SSD. The only way to make it uni-directional would be to put a peltier device in between, and those are pretty inefficient (unless your goal is to generate heat).

So, given that they say it's thermally conductive, then it can't be shielding the SSD from GPU heat. It must be acting as a heat sink to remove heat from the SSD. Common sense says this is so, since SSDs will throttle around 70 or 80 degrees C, whereas the air exhausted by a GPU should be nowhere near that.

Anyway, some benchmarks should be illuminating, especially if there's a way to remove the shield and run the bare SSD installed in the same spot.
 

ah

Reputable
Oct 29, 2014
69
0
4,630
Talking about M.2 cooling, I've lost the M.2 screw, I contacted MSI, asking for help with the screw, I haven't from them since. It's a pity, the reviewers fail to ask the M.2 drives vendors like Samsung to include some screws in the package. One pay hundreds of dollars, and the box comes with only a bary drive!!!
 
You're thinking about this in the wrong way. Yes, aluminum is a conductor, however this also includes a thermal pad in touch with the SSD, meaning the aluminum plate can act as not only a shield from the direct GPU exhaust heat, but also as a heat sink of sorts.

While I'm sure it would be nice to have some active cooling dedicated to the M.2, adding a fan and ventilation channels would dramatically add to the mboard cost. Besides, this is just a plate on one side of the M.2, it's not completely enclosing it in its own sealed chamber.

Again, you're not considering all the variables here. If the air off the GPU is hotter than the flash dies themselves, and if the aluminum were in direct contact with the flash, yes, this would heat up the drive. However two of those conditions aren't true. First, the GPU waste heat isn't hotter than the flash, and the M.2 shield has a thermal pad.

Those are functionally one and the same in this instance. No, the GPU waste heat isn't as hot as the flash, but it is much warmer then the rest of the air in the case. So the flash will be warmer when under a GPU than on its own elsewhere in the system. Thus, keeping the warmer air fom blowing directly on it will shield it from the direct heat.

Adding the thermal pad and aluminum plate increases the mass in direct contact with the heat generated by the SSD, thus it becomes a thermal dump mass lowering overall heat density. The airflow from the GPU will then hopefully take some of the heat away.

You're trying to be cute by setting up contradictions to then break down to make yourself seem smart or Chris look stupid. However, there's nothing incorrect about calling this a heat shield for an M.2, even though the shielding system can also act as a heat sink.


Did your 3.5" or 2.5" drives come with mounting screws? No, they came with your case or drive enclosure. Not to be harsh, but you're blaming a company for your own fault. Whether the screw came with the mboard or the drive, you'd still only have one, which in your case would be lost. If you're saying they both should come with one, then how many screws would get thrown away because people don't need two? Kinda wasteful.
 

bit_user

Polypheme
Ambassador
So, you're arguing that because the pad is not perfectly heat conductive, it acts as an insulator. Well, I agree with this, but the problem is that whole point of the top plate is to increase the rate of heat exchange between the SSD and the air. There would be no point in doing it as a heat sink, unless the rate of heat loss for the top plate + pad were better than the SSD by itself. And if it's not a net improvement in SSD cooling, then it's going to trap more heat in the SSD and we should see hotter temps. Otherwise, it's increasing overall thermal conductivity, and that's a 2-way street.

I don't even know what you're on about, here. What I was trying to say is that if they wanted to shield it from GPU heat, then there'd have to be ducting from all the way in front of the GPU, and a hole in back. Given that this isn't the case, sticking an insulator atop the SSD would simply trap the SSD's heat. Ergo, this top plate + pad is not a net insulator.


Thermal mass only matters when looking at transients. In cases where GPU heat would be a significant concern, the steady state is much more relevant.

Wow, that's a first. I'm trying to point out a contradiction in what you wrote, and you accuse me of somehow introducing it? By now, I know that you don't concede. And I'm not going to engage in a "he-said, she-said" with you. This could all be settled, with a bit of testbech analysis. So, cool your jets and let's agree to disagree until then.

I've gotten several retail-boxed 2.5" SSDs with screws (both Crucial and Intel).

Relax, man. @ah is just venting. A bit of sympathy or practical advice would be a more constructive response.
 

bit_user

Polypheme
Ambassador
I'll go on record with the following predictions. Assuming it's possible to test with & without the plate in place, here's what I expect the data will show:

Scenario 1: SSD is idle; GPU under load. The SSD should be hotter with the plate than without. The temperature of the GPU's exhaust should be hotter than the SSD, so the plate will act draw the heat into the SSD.

Scenario 2: SSD under load; GPU idle. The SSD should be cooler, longer with the plate than without. If the top plate can't even accomplish this, MSI's credibility will surely take a black eye.

Scenario 3: SSD under load; GPU under load. We hope the SSD will be cooler, longer with the plate than without. With M.2 SSD temps reaching above 70 C, the test will show whether the thermal resistance offered by everything between the silicon and the air is enough that the top plate's temperature is still above the temperature of the GPU's exhaust. If it is, then there'll be a net movement of heat out of the SSD. If it's not, then it's possible that running without the top plate will actually improve SSD temperatures, in this scenario.

If the top plate fails #3, then I wouldn't say it's a disaster. I think thermal throttling tends to occur during heavy writes, and I don't know how common that would be when the GPU is churning away. Maybe professional 3D & motion graphic artists will hit this case, but probably not gamers.

I still think the best solution for M.2 drives would be if there's a way to mount them perpendicular to the mobo, in place of one of the PCIe cards. I'm not talking about a carrier, here. I think the drives should still plug directly into the mobo and attach to a bracket shipped with the mobo that slots into the back of the case and holds the other end of the SSD. Holes in the bracket would allow air to pass over the drives, and exit the rear of the case. This should be in place of the last PCIe slot, so that it sits below any GPUs, in tower-style cases.
 
Status
Not open for further replies.