Things we MUST cool are not cooled

Hello,
I am the gaming laptop arena for few years now, silently reading and making my own mods, and sometimes, rarely, asking for help because there are people a lot more skilled then me also more experienced than me. I’ve changed the video card in my Dell 6700 twice, from an low end ATI to an K5000M and then to a 780M and each and every time mounted my heatsink countless times. bending the heatsink to adjust, playing with the heat pads to make sure that every little pad make the right contact to cool each piece from the videocard or else I would buy a new one in a short while. I’ve read in countless threads where everyone agreed that you need to, MUST TO cool these parts if you want happy gaming with no additional costs (a new videocard).

But then, after reading a lot or reviews and threads, looking at a lot of pictures, found that are some laptop models that leave out some electronics that in most MXM cards MUST BE cooled properly. Can someone explain to me what is the difference from a MXM 980M and a binned 980M and why those parts from the pictures does not need the cooling?

 

There's actually thermal pads on the bottom panel to cool those things.

 

Thank you for your feedback !
I am sorry to say it but there is NONE, I've had the Clevo P650SE with the 970 and so was a friend of mine (I asked him too) and none of us had NO PADS whatsoever... That's one of the main reasons I started the thread... On the P650SE in theory they could be cooled with some pads because the button is made of aluminium but on P650SG is plastic and not cooling friendly.
I am curious how vital is to cool those parts... from the looks of things it's not on some but it is on others?! What is the difference? Why?

 

Knowing a little bit about electronics, those inductors don't technically have to be cooled. I couldn't find a datasheet for that specific part number, but looking at a similar high current power inductor's datasheet, it clearly doesn't have to be cooled. At full dissipation, that package will level out to about 80C (for a 40 amp rated inductor). And the maximum permitted temperature before failure is 260C.

The only reason for cooling is probably to reduce how much the inductance degrades with temperature. At 80C, the inductance may decrease by 15% or more. That might result in not as stable power delivery to the memory or gpu under load.

But if the drop in inductance is acceptable, there's no reason to cool it. And actually, cooling it with the same heatsink the gpu is attached to won't do much. The inductor will still get hot, and it's inductance will still degrade, but not as much. Probably only really important for extreme overclocking to keep those cool.

EDIT: Another reason why you might want to heatsink those, it too keep all your power phases at the same stability level (even if it's a bit degraded). Consistent stability across phases is desirable over different phases having different stabilities. However for normal use, it's really not that necessary . . .

 

An MXM board is isolated thermally from the rest of the system, when on the motherboard PCB the passive cooling from the larger board is greater, the fan will also be drawing air over them (note they are next to it).

 

Thank you guys !

Later edit : I have noticed that on some of the laptops the GPU heatsink (of MXM cards) does not cover some memory chips completely, only 25-40% at best, depending on how the memory thermal pads are placed and the shape of the heatsink. Some heatsinks there were completely lacking the precision to set the thermal pads correctly in place resulting in poor coverage of the chip (noticed on my Clevo P751DM-G). On the back of some the MXM video cards the video card memory chips are without a heatsink or heat spreader (Dell M6700 ATI cards)... at best they have some thermal pads that connect the MXM memory chips to the laptop motherboard (Dell M6700 K5000M).
Are the memory chips not cooled/covered properly in danger of premature failure?

Thank you !

 

IMO, I don't really think premature failure will be an issue as these chips are designed to take it and operate at what I imagine are acceptable temperatures. What will really be affected though is the overclocking ability of the VRAM, so you may find that the memory becomes unstable even with minor overclocking. What may also be an issue (not so much for normal use) is that some VRAM chips are cooled better than others on the same system, because that introduces variances, though I don't really now how much this will effect performance.