OCZ copper ram sinks actually hurt my video card's performance

[Kosmic]

The copper ocz ram sinks that I bought ended up hurting my overclock instead of helping it. I think there are a few reasons:

I'm running my case without a side, and the "naked" memory could get a decent amount of fresh air on it. They would get HOT while gaming, but immediately cool off as soon as gaming was done.

With the ramsinks on (and no extra fans blowing on the card which is GPU watercooled) they seem to "hold in" the heat. They're still warm several minutes after running 3km2k1. Even at lower memory clock speeds than I had before, 3dm was getting horrid frame rates.

I wipped out the hair dryer, heated them up and removed them (plus a little bit of the sticky stuff that got left behind) and everything is back to normal.

The happy news out of all this is that I actually found a sweet spot for my card. I lowered the core and memory clocks trying to get something stable while the ram sinks were on there. When I took them off, I just ran 3dmark 01 at 421/388 and got about 150 points higher than I did at 425/391.

Maybe it's the thermal adhesive they come with, maybe it was that I needed to have a fan blowing directly on them. Maybe a little of both. Hopefully this info will help somebody else, too.

Kos

[texuspete00]

If you used a tape like substance thats no good. "Peel of the backing and stick" kind of stuff doesnt work. Need either thermal grease with some superglue in the corners or thermal epoxy.

I agree.... Sinking with tape is worse than not at all. An extreme example would be hyperX, they actually have a winter coat on.

[jamaljaco]

I use the ca in the corners trick with my favorite thermal compound , it works great.

[Soulburner]

I use the tape, but I put a very small dot of silver paste in the middle, so I get the best of both worlds.

They helped my 5950U from 965 to 985 on the mem, and my 9800XT from 780 to 820.

[celemine1Gig]

Well, what you just experienced is absolutely plausible!

Copper heatsinks are no good for passive cooling, that's what most people don't know. It's because copper can conduct the heat faster than for example aluminium, but it also keeps the heat longer inside than the just named aluminium. That's the problem with copper and passive cooling. Copper needs to be ventilated to have a better cooling effect than aluminium. Aluminium is far superior to copper when it comes to passive cooling, as the Alu can pass on the heat to the surrounding air much quicker than copper can. The absolute best solution for passive cooling would be a heatsink with copper base and large aluminium body (fins).

These passive copper RAMsinks are just some kind of marketing gag because everyone thinks that copper is better than aluminium if it comes to cooling. BUT, as I already said in the lines above, that's only true if you have a fan blowing over it.

So, you can either keep you RAM without any heatsink and have some fan blowing over it, which would be a good way to cool it. Or, you getsome alumium heatsinks (or as stated above copper aluminium hybrid) and cool it that way (a fan wouldn't hurt eitherway, logically).

Ah, and BTW, don't use conductive tape, as it conducts the heat like crap, in comparison to thermal epoxy for example. You could also use some thermal paste and fasten the heatsinks using hot glue.

[Soulburner]

Yeah I do have a fan blowing over mine.

I guess I could go xtreme and try the glue method and see if I get any gains from it, but the ramsinks as they are now get pretty hot after repeated 3D running at 400+ (800 DDR).

How much of a real gain would we be talking about though?

I use a small dot of silver paste in the middle and let the tape stick it down around the edges...

[celemine1Gig]

Originally posted by Soulburner
I use a small dot of silver paste in the middle and let the tape stick it down around the edges...

That's also one big problem!!! The tape is much to thick to allow a good thermal transfer in the middle. The layer of silver paste is as thick as the tape and that's way too thick to get optimal contact.

[naughty_guy]

celemine1Gig is right. Just want to add that the reason why copper heatsinks are worse than aluminum heatsink is because aluminum the element has a higher heat capacity. Which means a block of aluminum will need more heat to raise its temperature by 1C than copper.

[Soulburner]

Well i'm certainly not using epoxy after seeing what it can do.

Only 4 little dots, one on each corner of the GPU shim is holding a Vantec Aeroflow on my card, 100% firm and that thing isn't light by any means...

http://home.neb.rr.com/soulburner/Aeroflow%201.jpg

http://home.neb.rr.com/soulburner/Aeroflow%204.jpg

[Kosmic]

Thanks for the helpful replies, guys. I guess I'll try alumnium later with thermal epoxy. I hear that mixing some AS5 in with the epoxy makes it less permanent. Not sure what to do with these copper sinks though...I really don't want to mess with getting a fan blowing on the card.

[Minnyboy]

Got meself some of those a packet of thermaltake cooper hs as well as a oacket of the aluminium ones as well. They look pretty good.

Meant to use them on DDR Ram as they look like a perfect fit on the chips but I'm gonna put the sopper ones on my 9800XT (WC ASAP) & the aluminium ones on my IC7-G's mosfets.

There's 8 to a packet & I'd dare say that the copper ones are nearly on par with the OCZ Copper HS. Just a pity we poor Aussies never get anything good, wouldn't have minded the OCZ Copper HS.

Here's a pic
of the HS in their packets:

[Kosmic]

Be sure to use thermal epoxy and not "sticky tape" -- have a fan blowing on those copper sinks, too. Don't do what I did! (not a big deal...it was easy to undo)

[Liquid3D]

I haven't been posting for awhile but I must say this. Even though it's obviously belated, I want to welcome Celemine1Gig to the forum. I truly appreciate someone with a working knowledge of metalurgical/thermodynamics. The assessment was on the money in his/her replies. For a passive radiator copper isn't as effective as aluminum, and this where many of the "myth's" of aluminum's conductivity compared to copper become pertinent. For the longest time I simply thought mobo makers whom used passive NB/heatsinks, did so primarily for cost, and in many cases I'm sure they do.

But when two respectable Enthusiast companies such as Epox and Asus release their flagship (originally) Canterwood mobo's with passive aluminum heatsinks on the NB's it made me wonder? For a $185 mobo (Asus P4C800E-Deluxe) and the cost the Epox 4PCA3+ I'm sure they could have sprung for copper, if it would give they're platform the edge with Overclocker's. And this is where it becomes pertinent, that while copper has twicethe conductive properties of aluminum, based on mass and ambeint temps, that relationship can work against you. Where-as aluminum's perceived (or misperceived) propensity to dissipate heat "faster" then copper, is simply a result of it's condcutive properties.

[Soulburner]

Yes but those of us using ramsinks have fans blowing over them as well, and in this case copper wins.

[asw7576]

So should I take off the heat spreader??? mine is mushkin pc3200 with silver plates and black pcb. I feel it traps heat instead releasing it.

[enzoR]

Originally posted by naughty_guy
celemine1Gig is right. Just want to add that the reason why copper heatsinks are worse than aluminum heatsink is because aluminum the element has a higher heat capacity. Which means a block of aluminum will need more heat to raise its temperature by 1C than copper.

other way around. copper has a higher heat capacity than Al thats why copper is better. it takes more heat energy to heat it up by a certain degree as Al. with aluminum, given the same amount of heat "given" to it, the temperature would rise higher than with copper becasue copper "needs" more heat energy to raise its temps. Al doesnt give of its heat better, copper is a better conductor so it gives of its heat better to its surroundings. heat capacity has nothing to do with it. it is only the ammount of heat energy needed to raise a specific ammount of ammount of matter (say 1kg of copper) by a certain degrees (for 1kg its 1 C)

[pkrew]

Without looking up the heat capacity of copper vs Al I cant really say. I'm sure the info is out there and I would tend to think that enzoR is right. At any rate I've been very happy with the OCZ bga's and the OCZ flowers that I have. I have a mixture on my XT, it was what I had(4 flower gpu side, 4 bga other side) and a smart fan II blowing on them. My vdimm is set for 2.9 and I get a max oc of 435. I also used the bga's on my ram and with a 120mm smart cool blowing on them it gave me 5 extra 5fsb. I used Artic Alumina to attach all of them

[Kosmic]

I can't comment on specific properties of either copper or aluminum, but I'll be sure to have a fan blowing directly on my card and to use thermal epoxy instead of the "peel & stick" adhesive before putting the copper sinks back on my vid card's ram.

[texuspete00]

Yeah, I'd work on the TIM (thermal interface material) solution and not overthink the aluminum vrs. copper thing. A case still has airflow, or at least it should. At first it looked like you got scared out of using them, its good to see that you are. I think with a good TIM they'll be better than nothing even if you dont have a fan ziptied onto your card.

Not that the comments made werent good ones. But if you do want to get all crazy over temps then copper and direct fan would be the way to do it. If you bought the sinks for piece of mind, get rid of the TIM handicap and you're there. Again, an enthusiast box usually has a bit of flow anyways.

Once you pick up some thermal epoxy too, you'll end up sinking other stuff. I bought mine to put on ramsinks and just having it around has led to constant little improvements almost everytime I break down my system. It has turned into one of those things I dont go without. When I run out, I'll buy more promptly - kind of thing. Like beer and jack daniel's

and asw.... I dont know about mushkin but when I took the spreaders off my hyperX and had a look I was very sure I had done the right thing. Thick tape and even worse was the bad contact...

[buff]

Copper absorbes heat faster then aluminum. Aluminum dissipates heat faster. thats why we have copper core aluminum heatsinks, and thin fin designs for all copper.

[gkiing]

damn, I didnt want to hear this. I've got 8 ocz copper bga ramsinks epoxied on with arctic silver epoxy. Their never coming off. But I've got an 80mm fan blowing on them, so I think it's ok.. My mem does 375mhz on 9800pro

[texuspete00]

Uhm... McFly? Read the whole thread for more info. What you have is ideal.

Just kidding ya

[Liquid3D]

Originally posted by enzoR
other way around. copper has a higher heat capacity than Al thats why copper is better. it takes more heat energy to heat it up by a certain degree as Al. with aluminum, given the same amount of heat "given" to it, the temperature would rise higher than with copper becasue copper "needs" more heat energy to raise its temps. Al doesnt give of its heat better, copper is a better conductor so it gives of its heat better to its surroundings. heat capacity has nothing to do with it. it is only the ammount of heat energy needed to raise a specific ammount of ammount of matter (say 1kg of copper) by a certain degrees (for 1kg its 1 C)

Exactly higher heat capacity, and conductivity are synonomous in this circumstance. Copper has twice the conductivity as aluminum, as I said above, which means it "retains" whatever temperature, twice as long. Albeit heat or cold. Which is why again, copper requires a fan in this circumstance. HOWEVER if the ramsinks were made of extremely thin copper, it may be different. That's what I was trying to explain before. A material's conductivity is also relative to it's mass under these circumstances

For example: when comparing the fin thickness between Thermalright's all copper SLK-948-U fins, and the aluminum finned ALX800 as seen below. The ALX800 spec's much thicker fins, the added aluminum mass is more condusive to an environment where kinectic energy can propagate through it more thoroughly. In a passive design such fin thickness wouldn't be necessary;

[enzoR]

Originally posted by buff
Copper absorbes heat faster then aluminum. Aluminum dissipates heat faster. thats why we have copper core aluminum heatsinks, and thin fin designs for all copper.

yes thats true... but only after you turn your pc off that that counts. copper doesnt absorb it faster... it absorbs more and so the temperature or it wont go up as high. when the pc is turned off... it cools slower becasue it has more heat in it. but by that time.. the temp of the aluminum would be skyhigh becasue its temperature rises more than that of coppers becasue it has a lower heat capacity.

the reason why copper core and Al fins is for weight. nothing more

[enzoR]

Originally posted by Liquid3D
Exactly higher heat capacity, and conductivity are synonomous in this circumstance. Copper has twice the conductivity as aluminum, as I said above, which means it "retains" whatever temperature, twice as long. Albeit heat or cold. Which is why again, copper requires a fan in this circumstance. HOWEVER if the ramsinks were made of extremely thin copper, it may be different. That's what I was trying to explain before. A material's conductivity is also relative to it's mass under these circumstances

For example: when comparing the fin thickness between Thermalright's all copper SLK-948-U fins, and the aluminum finned ALX800 as seen below. The ALX800 spec's much thicker fins, the added aluminum mass is more condusive to an environment where kinectic energy can propagate through it more thoroughly. In a passive design such fin thickness wouldn't be necessary;

I'm sorry i'll just be like Kanavit on this one. However lets keep the discusion going in a matured manner.

Take a look at this linky

i would advise you physics lovers out there to read the whole thing becasue it really is educational. And our topic is on page 4 btw