The Official OC (OverClocking) Thread!

Wow, fascinating, not seen an X2 under the bonnet before.
I wonder why AMD still put the date of the copyright for the Athlon name, and not anything subsequently?

 

sam,
Don't know.

 

i think thats all they have registed as trademarks, the amd name and logo, and the proc types, athlon and sempron. i dont think they are that bothered about the number part of the name.

bigdk had it up to 3.8 or 3.9, i think, but couldnt keep it stable on air, so i guess he will hit that atleast now hes running the water cooling. does that new bios apply to his board crowy? if it does a combo of those two things should see him get above 4ghz.

crowy, your cpu looks mad without the heatspreader on, is that another smaller one thats just covers the 2 cores underneath? just out of curiosity what does/can your rig run at?

 

sammorris,

I think that's the regestration date for the present variation of the AMD logo. Before that, the old Athlons and Athlon XPs had 1999 on them, at least all mine do! The K-6 IIs are all 1998.

Happy Computering.
theone

 

Marse99,

No, thats the die with the two cores built into it.A lot of people expect to see two seperate cores but that isn't the case.I should have taken a shot after I removed the old thermal paste,the surface of the die is like a mirror!
I was stable @ 2840mhz with the IHS still on(that's on home made water cooling).Iv'e had it to 3033mhz,but it's not stable at that speed. I'm going to start testing soon with the cpu as it is now.
Nearly forgot,yes that bios is for his board I'm sure.

 

that makes me feel dumb. no i thought that the c2d were the first to put both core on one chip but hey you learn something every day.

crowy dont take this the wrong way but theres something about "home made water cooling" that just does not sound safe. could you elaborate on that please? if not here in the cooling thread.

some other questions occered to me before, now that it s off can you put it back on? will it cool any better with it off, and if so why do they put it on at all? will thermal paste damage any thing under the heat spreader?

 

Marsey99,

Yes you can put it back on.Just remove the old silicone,apply new thermal paste apply new bead of silicone and replace.
The whole idea of removing it is to get better temps.Your HSF/water block is concentrating it's thermal efficiency onto a smaller area.
To much thermal paste isn't good,you wouldn't want it oozing out and all over your motherboard!!LOL!!
Here's some pics I posted some time back of the stock heatsink with the cooling mod.Iv'e got a submersible pond pump in a 10 litre container of 70/30 mix of water / coolant working in a recirculating fashion.

 

those temps are amazing and the heat sink looks alot better/ more efficent than alot of the ones iv seen with pipes, one in, one out. sorry if i was being rude but the term homemade brought images of hose pipe and gaffer/duct tape to my mind.

so is the plan to machine a copper plate that will bridge the gap from the cores to the heat sink? in place of the heatspreader. will you attach it the same way as the original was or will you stick it to the hs?

 

marsey99,
What I'm hoping to do is make a copper water block that covers the die only.That way I'm getting maximum cooling onto a smaller area.
Should look something like this:

 

thats some proper engineering skills youve got there.

how much do you think it will drop your temp? or should i say, how much do you hope it drops your temps?

 

marsey99,
Not sure how much of a temp drop I will get but anything even 2 or 3 degrees will be a bonus!!Still trying to source the copper.Everywhere I look will only sell it by the length.I'm going to try some engineering shops and maybe try and get hold of an offcut.

 

crowy,

I don't mean to throw rocks at your design but the water should go in the bottom, or closest to the heat source and exit the furthest from the heat source as possible in order to work the most efficient. If you bring the water in the top, it will already be warm by the time it reaches the hottest point and will lose considerable efficiency. Water flow is another thing as to how fast the water travels. If the flow is too slow, it will get too warm as the heat isn't being carried away fast enough to be effective. If the water flow is too fast, it travels too quickly to absorb much heat and the CPU will overheat as well. A variable speed pump would be the best solution as you will be able to monitor the flow adjustments on the fly in real time as you change the flow for the best cooling characteristics for your particular system. It's not a question of volume, but a question of volume and time.

Do you plan on using a secondary radiator with your setup? If so, you may want to consider mounting it outboard to keep the heat liberated inside the case to a minimum. I've seen a few water coolers that have a case fan inside blowing out through a radiator on the outside with the case sandwiched in between. You want to go from the pump, to the CPU water block, to the radiator, and back to the pump. If you plan on using a GPU cooler, then put it in between the output from the cooler and the Radiator. GPUs generally run hotter than CPUs, so the reason is obvious. Also keep in mind that most systems have only one line in and one line out of the water block. Your pic shows 4 pipes on each side of your block. I would make a simple mod on your block by making one line in to the heat sink/water block, and interconnect the remaining paths with plastic tubing until you have one line in and one line out. This way you can keep the water flow as smooth as possible and increase the whole cooling system's efficiency! Basicly, you want the water flow to be as smooth as possible and at a good enough rate of speed to carry as much heat away from the source as possible. This is why I would recomend a variable speed pump so you can control the ammount of flow.

I actually did something along these lines when I was in College as an Engineering project (I studied Air, Water and Hydraulic systems). Our heat source was a Catalitic heater and we had to build everything from scratch. The goal was to design the most efficient heat removal process we could with the provided materials without boiling over. The radiator and fan were the same for everyone so our efforts were concentrated on making the water block as efficient as possible. I won the award for the most efficient design. I used the same basic idea I explained above with the tubing only I used copper pipes instead of plastic tubing to interconect the water paths I had drilled into the cooling block. Someone argued that I had cheated by adding the copper pipes, thereby increasing the surface area but the Professor pointed out that by drilling out the block that I actually had less because there was less copper in the block to absorb heat to begin with.

I wouldn't design something to just cover the die either, as you will lose efficiency and overheat the CPU. That's the reason no one builds a cooler that way to begin with. It leaves too much heat to escape rather than be absorbed by the cooler. In cooling a CPU, surface area is everything, not just the contact point!

It's amazing the things you learn that you think at the time, are useless only to find out many years later that they are relevent to something you need today! If I can be of any help with your design, just holler!

BTW! I used the Arctic Silver "Ceramique" when I put my system back together. I highly recomend it! Right now running at 3.81 I'm looking at 33C at idle in a 25C room. This new MB has better Throttling control, more like SpeedStep that lowers the CPU demands as needed, rather than wait for a certain temp before it throttles the CPU down. The highest temp I've seen in practical use is an occasional flicker to 52C while encoding with RB/CCE. 50C is about the norm! If I do a 10 pass burn-in, it will flicker to 57C in a 27C room, but will for the most part maintain 53 to 55C. MB never gets over 36C no matter what!

Happy Computering,
theone

 

theone,
Thanks for the comments,much appreciated.
Actually the inlet will be the same distance to the cpu core as the 2 outlets.The pic I drew isn't to scale.I figured with the water flow being directed straight onto the centre of the die then dissipating the heat in 2 directions,that should remove as much heat as possible.
As for a radiator,no plans there.I'm looking at getting a second hand bar fridge(can usually get them for around $50 aust) and placing the coolant reservoir inside and running the tubes through the door seal.
Obviously that will mean modifying(IE: cutting)a portion of the seal.
Then I will use some foam weather strip to reseal the modified seal.
The inside of the fridge should be around ~4degrees so the coolant should stay nice and cold.The pump I have has an adjustable flow rate so I can adjust for the best efficiency.

Nearly forgot,
remember,I've removed the IHS so the base of the cooler will only need to be the same size as the die.

 

Crowy,

It doesn't matter. In fact removing the IHS defeats it's purpose because you need the heat spread out as much as possible. That's why no one builds one that way. It takes a larger surface area to absorb all the heat made by the CPU. That's why all heat sinks cover the IHS. You can't get adequate cooling with such a narrow heat sink! Fact! The larger the surface contact area of the heat sink in relation to the IHS, the better it will cool. The IHS just helps improve the spread of the heat so the larger surface area of the heat sink can better deal with removing the heat! You will notice that even with a VapoChill, the heat sink completely covers the IHS. With water cooling and a heat sink that only covers the die, you could get some condensation problems as well, as heat from the die that hasn't been absorbed by the water block would go up the sides of the water block. The bar fridge woud make condensation problems even worse! It would also likely make the case and MB hotter!

Happy Computering,
theone

 

Russ,

The copper block on my big typhoon just protrudes beyond the size of the IHS.(By about 3mm all around).

The bar fridge will cool the water to around +4 degrees at best.

Iv'e had the coolant down to -10 straight out of the freezer.
Remember I don't run a side panel on my case and I have a 30cm pedastal fan blowing air straight in the side.This keeps condensation at bay.

 

Russ,

Do you think it's better than AS5?
If so I might try it.
Iv'e still got some AS5 so will use that to start with but may try ceramique and compare the difference.

 

It wouldn't be better than AS5, The Silver V is marketed at the full end stuff whereas the Ceramique is like the midrange stuff. It's probably good for the price, but I'd stick with Arctic Silver if I were you.

 

sammorris,
Thanks sam,you just saved me $10!!!!LOL!!!!

 

On a side note:
No post from BigDk since the 21st.
He makes me nervous when he does that.
I'm really curious to see how his intel is performing because I think he's gonna gonna come up something pretty damn good.
Iv'e decided my next upgrade has to include a vapochill case.
Then I know I never have to worry about getting things any cooler.
In winter I could Post @ 3ghz and run benches when I had the big typhoon hooked up(temps in the morning were around -1 ~ +4 degrees in the morning.
At lunchtime the temp would be 10 degrees and I was just getting into windows @2950mhz so 6 or 8 degrees makes all the difference.
@ -20 ~ -30 degrees,temps are the last of your worries!!!!

 

Well, water cooling, draw your own conclusions. he's either still messing with it, or he has messed it.....