it means original x48 dfi kit from mips should function with some skills
Printable View
it means original x48 dfi kit from mips should function with some skills
Performance wise what's the best nb block that can fit on here?
Anyone use EK Water Blocks EK-Mosfet DFI UT on DFI UT X58 T3eH8 ??
Quote:
Originally Posted by westsider
I was just looking around the mips site and they have a native X58 PWM waterblock, that looks like it should work.
OK I am trying to use same setup with only MCW30 on the NBQuote:
Originally Posted by Praz
http://img196.imageshack.us/img196/3336/img0305xcp.jpg
All fans set to exsaust, loop is PA120.3+Fuzion v1+MCW30+DDC10W. And as you see I have a little airflow near PWM area.
But running max bclk 180 I may get on default voltages - CPU and VTT 1.2 - I have much more higher system and chipset temps :confused:
Ambient 30C, water 34C.
http://img196.imageshack.us/img196/4102/61600524.jpg
What's wrong with my setup?
May anyone share his temps?
It says 51C. I think your pwm temp is fine for your 30C ambient.
My ambient temp for that screenshot was probably somewhere around 25C. Try getting some amount of direct airflow over the PWM area if you want to lower temps further.
Is the heatspreader on the NB removable or would that mean jerking the NB die off with this heatspreader? :D
I'd hate to kill it lol... but if it can be done you can remove the PWM cooler and put a separate waterblock on it :up:
And stand a really good chance of killing the board.Quote:
Originally Posted by Garrett
I run a quick test with cpu 200x20 vid 1.36v vtt 1.36v and got ugliest results ever
http://img19.imageshack.us/img19/8732/83844235.jpg
cpu core 79C!
this is with 31C ambient and water temp jumped to 38C
so it's not a problem of mosfet heatpipes but my whole cooling system weak? PA120-3 and simple DDC not enough?
thx in advance
Odd. Same setup as I was running.Quote:
Originally Posted by rioja
He probably had girly fans :p:
I'm plagued with high PWM temperatures as well. I use the same setup as above, with an MCW30 on the northbridge, but I had to put a couple of fans on the PWM heatsink to keep temps down anyway. That lowered the load temps by around 8c, but they're still pretty high. I wish it was possible to put a waterblock on the PWM area without having to kill the board! My CPU temps are high as well, and that's on a dedicated loop. I'm contemplating buying a HK 3.0 to see if that improves anything... I guess some CPU's just are hotter by nature though.
Of course you can buy a wb. I had a heat killer block on my mosfets and temps were better but I just don't see any performance benefit doing so.
Yeah but apparently it's extremely easy to kill this board by altering the PWM cooling. I read somewhere that Oscar Wu strongly adviced against taking off the stock heatsink... and I've already seen many cases of dead boards where people took off the heatsink. Something about the slave chips not making proper contact, and it's apparently very hard to get proper contact all over. There is of course always a slight risk in altering stock hardware, but it seems the risk is much higher on this board than usual. So therefore I'm not sure if it's worth trying. I also read somewhere that the PWM circuitry on this board is supposed to be ok up-to 130c. I've no idea if that holds any truth though. Sounds like alot..
60W 40mmx 40mm TEC module + almost any NB Square block.Quote:
Originally Posted by I34z1k
:yepp:
Yes parts of the pwm [volterra chips] are rated over 100C. I just bought some really thick thermal pads to ensure contact.
I'm tired of people whining about digital pwms lol. All the ATI guys :banana::banana::banana::banana::banana: endlessly about it on the 4870/4890 reference design and the same with the DFI mobos. It's suppose to get hot. I would just put some fans on the stock heat sink [let's face it, it is MASSIVE] and be done with :)
I did full board cooling with my x48 and atm I am just wc my nb. While a mosfet block looks awesome, it's so not worth it.
Momi drop a TEC on it!!!
common!!
:rofl:
Well, I have put fans on mine (2x 70mm) and it still reaches around 80c at 1.25 vcore. Without the fans it would get in the territory of 90c... and if I push more vcore temps get completely out of hand. So yeah... it is more than whining, it's a bit of a problem on my board. I live in the cold north too so ambient temp is not the source of the problem either.Quote:
Originally Posted by MomijiTMO
Quote:
Originally Posted by jagt
I gotta agree. The whole PWM cooling on DFI boards suck. The heatpipe doesn't really do much at all. with temps of the northbrige less then 50 with a MCW30 on it, and stressing the CPU... temps of the pwm reach like 85 on my setup. and the heatsink on the PWMs certainly isn't even close to 85.. so the heat transfer from the fets to heatsink isn't good.
I'm thinking of taking the hinksink off, doing away with the thermal pads (if all the fets are planarized . Since my board has analog pwm, i think it might be less prone to blowing up like the boards that use the volterra and the tiny glasslike fets.
or i might just leave it, since it's never gonna get that high under normal conditions (only stress testing).
So I have results from 2 independent systems, using only MCW30 at NB and running i7-920 190x20@1.35v
Factotum's system
ambient 26, nb 37, pwm 51, cpu 36, cores 48 idle
ambient 26, nb 40, pwm 74, cpu 54, cores 66 load
ambient 26, nb 38, pwm 65, cpu 51, cores 73 load + direct airblowing
My system
ambient 27, nb 37, pwm 47, cpu 36, cores 43, water 30 idle
ambient 28, nb 46, pwm 72, cpu 58, cores 76, water 38 load
ambient 28, nb 43, pwm 60, cpu 55, cores 78, water 38 load + direct airblowing
By touch, while running with system temp 72C mobo heatsinks and heatpipes are barely warm.
So I have theory that some mobos are not as others :)
Some have good heatsink contacts in PWM area and run with 50-60C system temp like Praz
But others (revision dependent, poor quality or smth) do not have good contacts and run 70-80 as with me, Factotum, jagd, Diverge etc
Correct?
Alittle update:
It seems Linpack thru the OCCCT stresses CPU a higher than Prime
So under Prime95 I got these results:
ambient 27, nb 44, pwm 67, cpu 55, cores 71, water 36 - load with no direct airflow
Spec in sig. 21x191 HT on, 1.33vcore, 1.33vtt. No pwm airflow. Prime blend.
ambient 26C, nb 38, pwm 48, cpu 38, cores 42 idle
ambient 26C, nb 42, pwm 66, cpu 50, cores 66 load
good news - SW-16 FULL compatible with MOBO :)
and simple testing:
my first setUp:
Laing 3.2 -> XSPC RX360 -> Fusion v2 -> MCW30 on stock heatsink without a direct blow of air in the area of PWM + FLAME FREEZER which attaches to I/O connector area.
http://images.people.overclockers.ru/182385.gif
my new setup:
Laing 3.2 -> XSPC RX360 -> Sapphire (CPU) -> MCW30 (NB) -> SW-16 (PWM) -> MCW60v2 (GTX285), - no FUN in case :)
http://images.people.overclockers.ru/182540.gif
SB - 4xMC14 (50C)
WC PWM with SW-16 = 61C vs 74C with stock heatsink + MCW30
WC NB with MCW30 = 37C vs 40 with stock heatsink+MCR30
Cool I have the SW16. No one had tried it until you came along!
Cheers.
i did a little research and found out, that pwm heatsink has a gap between main base and part, which contacts with mosfets
http://img26.imageshack.us/img26/3892/img0395r.jpg
so i think this is one of reason of high temps (thermalright, eh?)
next i may confirm that mips v1 will suite here after 10 mins of modifying :)
http://img26.imageshack.us/img26/4897/img0387paq.jpg
when i used only one mcw30 on nb temps jumped over 84C :eek:
now pwm temps is 44C in idle and near 66C at load
http://img26.imageshack.us/img26/7947/13467057.jpg