Mosfet cooling

[basserdan]

Some expert opinions please...

I'm about to put together a new system. I'll be cooling a clocked 2600k on a Z68 (P8Z68) with two RX240's (don't ask me why 2 x rx240, I'll show you eventually) and whichever next gen dual gpu is fastest (or available) with an RX360. Rads are all in the case and I will be pulling air inside as well. With all the rad capacity I will have, I intend to run my fans quietly, even when gaming (unless I can't), which of course, means there will be only a minimal amount of warmed air flow through the case. I will have a 120mm exhaust. I could mount my bracketed 80mm fan to blow over the vrm.

How warm is it going to get in the box? The southbridge should be ok but will the mosfets get to hot? I know they're supposed to be good to 120c but, in a low (warm) air flow environment, is this going to be enough?

[defect9]

consider that the air coming out of the rads into the case will be somewhere in the "just above room temp" to 50c range (depending on the water temp, naturally). cooling mosfest that are running at 70c (or whatever they get up to) with 50c air is still of benefit to them. the fact that there will be air movement is a benefit already, meaning you won't have a pocket of hot just sitting there growing.

* I honestly don't know what temp the heatsinks on the mosfets will get to, just using those temps as an example

[BeepBeep2]

Will it really be a benefit with 50c air?

Convection works on its own...

[defect9]

Will it really be a benefit with 50c air?

Convection works on its own...

if the heatsink is hotter than the air, yes. if not... no. temp isn't additive, it's relative.

[basserdan]

This is along the lines of my assumptions but, I would still like some more opinions.

I'd really rather not have to add more blocks to any of my loops. However, I don't want to keep anything operating at the edge of it's maximum tolerance long term.

[BeepBeep2]

I dunno...
Lets say you had a water loop with fans pushing through the radiator. The radiator, a heatsink, will displace so many watts of heat into the air passively at a certain ambient.

Lets say my CPU runs at 51c load with said water loop and fans turned on in a 20c ambient, and 75c without those fans.

Add 30c to that 20c ambient, for 50c and you end up with a CPU at 81c load compared to the load without any airflow.

Which is better?

[Utnorris]

The mosfets on a Z68 board will be fine as long as they have some sort of air flow. I run my Maximus IV with just the air from the rads flowing over the heat sinks and there is no issue whatsoever and that's with the added heat of the NF200. The SB boards do not get that warm, putting blocks on them for the mosfets and SB is strictly for looks unless you are trying to push insane amounts of voltages through them.

[CrazyNutz]

+1.

You need not be concerned.

[mav2000]

Well, just to add my two bits here. I used to run the loop that I have right now, with the top fans from the radiator pushing air in. I found that heating up the vrm and NB heatsinks a lot. I then turned the fans around to exhaust air off the top of the case and the temps on the mobo fell a lot.

[CrazyNutz]

Well, just to add my two bits here. I used to run the loop that I have right now, with the top fans from the radiator pushing air in. I found that heating up the vrm and NB heatsinks a lot. I then turned the fans around to exhaust air off the top of the case and the temps on the mobo fell a lot.

That would be expected. In your scenario, you had heat coming from every component trying to rise against your incorrect airflow.

I go by this rule of thumb for airflow:
Bottom & Front = Air In
Top & back = Exhaust

[basserdan]

Thanks everybody so far.

I wouldn't be against anyone else adding their opinion though. Simply, watercool the mosfets: Y (for yes) or N (for no).