Discussion in 'TechPowerUp Club Forum' started by jbunch07, Sep 22, 2008.
Beautiful shots HammerON
yeah its extra TIM AS5 to be exact
Nothing fancy here just a shot of 550BE (which is now X4 at 3.7GHz) socketed on my TA790GX 128m mobo.
Sunbeam Core-Contact Freezer dominating my mobo and case. The case is a CM-690 (or damned similar).
Finally a close up of the MOS-C1's which cool the CPU's mosfets. Without them I couldn't OC much.
beware, if you unlocked it to a quad it probably pulls enough to heat up the pwm severly under load conditions... you see the little electrolytic cap, near the solid caps of the pwm? (the one located towards the I/O area of the board, centered between the NB and the PWM?
this one blew on me after a couple of months on a 955... you may want to temperature probe that area, because since that cap blew, the board lost about 20% of its overclocking capability i would recommend cutting up a leftover sink from an old/broken board, cut to size. i did that on the board, and it fits me well, and leaves the pwm only slightly handwarm, on a 90nm X2!
I'm not sure I follow you. I see the cap you're talking about and I heed your warning but I can't figure out where I would put a heatsink (or heatsinks) that isn't already covered.
The mostfets involved in the pwm for the CPU have sinks on them so the only thing I could think to do is put heatsinks on some of the other mosfets or even the chokes. There's a chip just to the right of the P4 connector that looks like it's involved in the pwm circuit that I could put a heatsink on. Could you specify?
lets say... that small sinks dont cut it, they will be oversaturated with heat in minutes. you need more thermal dissipation area for the mosfets, and it would indeed be not an unwise idea to cool the chokes additionally, if possible (they get really hot)
Those MOS-C1's are plenty for his purposes. I had them on a vmodded 4870, and they did fine on mosfets that generated a shitload more heat than his will at his settings. They are copper, and do better than the crappy leftover aluminum sinks you find on most boards.
Clearly this issue needs some hard data and I'm going to collect it using two thermal probes placed in various combinations on the mosfets, the chokes and on both. I'll report back with the results. (I sure wish I had a thermal camera!)
A side note: I think it's also important to mention that my case has no shortage of air flow. There are six 87cfm 120mm fans, three intake and three exhaust, in addition to a 120mm fan on the cpu which exhausts onto the pwm to aid in cooling it. (You can see how the fins of the CPU HS nearest the pwm are flared down in my pictures.)
I believe in what you say, but i can tell you, that a hd4870 has a much,much stronger PWM,even if its PWM heats up more severely. (its not even a mosfet, but digital, so i think you can hardly compare that )
The Biostar is meant for 140w heat dissipation, but just sports a weak 4+1 PWM for doing that. It wouldnt be the first Time, that one of exactly these Boards Mosfets simply explodes thru too much heat.
But,as streetfighter suggested, only a temp probing will bring him knowledge about whats really going on, so i think i wait for that before i can bring my final statement
My 4870 was a non reference powercolor. It most certainly did have mosfets. With the massive overvolting I was doing, those exact heatsinks were fine. I guarantee I worked those mosfets harder than he is working his at his relatively mild settings.
I bet i dont have to tell you, that most VGAs tend to run 70+ on load, just on the core. hell, i think most could even survive a "life" (3 Years? maybe more) at about 100c, if thats the absolute Max, and they are specced 120c for shorttime.Show me one board which is SO stable, that Mosfets that literally burn, do not affect system stability much,nor the Lifetime of the Board! (in fact they did affect stability, especially above 1.45v on a quad, and yeah.... these things really go "POP", on quite a few owners of the biostar 790gx... they use crappy mosfets... one guy even stated, that when his Biostar 790gx a2+ blew, there were Flames shooting out to the open sidedoor
i owned that board, paired with a 955, and crunching that nearly set that damn thing on fire! i wasnt able to touch the mosfets for even the fraction of a second, because it was severely painful. i also remember seeing a pic of this board, blewn, with exactly these sinks on what was left of the nearly molten mosfet area^^ therefore i dont recommend them for that critical task.
i may be wrong, but i think i know this burning bitch a little, so im really interested in the results of streetfigher
something a little more along these lines then
sorry guys, but please make a thread about this or do it via PM
this is for Pictures!
just before a mod says it
shit where'd you get that avatar pic Don that's frigging funny as hell
The core has nothing to do with Mosfets. I guarantee running 1.7 vGPU thru the 4870's mosfets put more stress on them than his mosfets are seeing.
Here is my small contribution. HUGE PIC WARNING.
nice pics, but that mobo if bending like a mofo haha
Yea, I know. That board really sucked ass too. A DFI 790FXb-M2RSH. It would randomly crash, then finally crapped out. I now have a new one from RMA waiting.
of course it hasnt^^ its just, that on hd4870 and 4850, usually the Mosfets (if they are used) get much hotter than the core itself. and of course would that amount of load, bring a much greater stress. I just cant tell you how long you can do that on air, before it gets toasty.
can it be, that you just misunderstood me, due to my just average english?
I never doubt that your card can take a lot heavier punch, voltage and currentwise, than even the most mainboard PWM can. i have seen my 4850 cook at temperatures, that i wasnt able to reach with, even if i combined most of my sensor readings. i believe, the max i provoked was 125c before it would shut down. Try it with any board.... i think it will die very fast at that temp
Here are some close up smexy shots for all of you.
Sorry for the large post, but I thought you all would like to see these.
Love the VF-1000!
I was lookin at getting a VF-1000 for my HD5770 as the fans started getting noisy and it's an sealed unit so i can't re-lube the bearing sod it and i can't find a fan to fit
so what's it like to use good temps
Here you go. This is after 6 3DMark06 runs. Notice that I have the VF1000 on my top card. It stays the same temp as the bottom card. After a couple of hours of JC2, it tops out at about 62 or so. I would definately recommend it.
Separate names with a comma.