New round of testing coming.

[muddocktor]

Well, since things are slow and I had a little money left over to mess with, I decided to do some more heatsink testing. I am also trying to roll in a new system to the test box too; a 980X system is presently mounted in the test case which I will try to use for this series of testing. The only problem I'm having is that when I have it overclocked over 4.0 I am seeing clock throttling even when Real Temp and Core Temp both show temps well below the point where throttling should take place and I will start another thread on that issue in the Intel cpu forum. The heatsinks I plan to test are as follows:

Copper TRUE, which has a lapped base due to the incredibly poor base finish on it when I bought it a year or so ago.

TRUE Rev. C Retail

Prolimatech - Megahalems Rev. A

Thermalright Venomous X

And the new additions to my test heatsinks are:

Cogage True Spirit

Noctua NH-D14

I am really curious to see how the 2 new additions will stack up against the heatsinks I already own. The Noctua should be a real no-brainer for a cooling choice but it is very expensive also. The real sleeper might very well be the Cogage, which is based off the Ultra 120 design and is very affordable at under $40 including a fan.

Anyways, over the weekend I plan to try to solve my throttling problem on the 980X system and get started testing my present heatsinks while waiting for the Noctua and the Cogage to come in. If I can't solve the throttling problem I will go back to the old Q6600 system to test with, since i can get a thermal load in excess of 200 watts with it if needed for testing by pushing the snot out of it.

 

[m0r7if3r]

Good to see the true spirit in there, its a favorite recommendation of mine for budget projects...what're you using for paste?

 

[muddocktor]

I'll be using Ceramique for all the tests since it's inexpensive, works well and I have a bunch already. As long as my TIM remains the same for all the testing, it should give a good comparison between the heatsinks. I've also tested with it before and Ceramique has given quite consistent results. It might not be the best TIM out there but it's close enough, especially in the bang/buck department.

 

[bing]

Great news Mud !

Just don't forget to test them with "The Beast" to see how good they scale.

 

[hokiealumnus]

*Cough* Gather data & write it for the front page. *Cough.*

Make it one-per phase or all at once, your choice. It would be great content for the FP.

p.s. I think there's an H50 around you can test, PM me if you're interested.

 

[SecrtAgentMan]

When you're using the Noctua could you run a test with the thermal paste that comes with it?

 

[muddocktor]

I'll probably do it after the testing is complete, just to see how good their paste is. But I want to gather my data first before going off on a tangent. All this testing takes a large amount of time and it's too easy to go off on a tangent and then have some kind of hardware failure that resets you back to zero.

 

[ehume]

One thing I discovered when I did my testing on my D14: a good push fan like a San Ace 9G1212H101 does a great deal of cooling. I know folks usually test the D14 with its stock fans, which are really good, but this cooler really benefits from major airflow. While you get a 1c improvement in cooling with two San Aces or two UK3K's on the Megahalems, you get a more pronounced benefit with the D14. Just try it with nothing in the middle - a 140mm fan can't keep up. Heck, I want to see you put your 9CR1212P0G03 on the thing.

 

[burebista]

[...] try to solve my throttling problem on the 980X system [...]

I'm curious, what throttling problem do you have on your Gulftown?

 

[muddocktor]

Yeah, I plan to do some playing with it after I get all my main testing done. I just modded the flange with my dremel this morning on The Beast so the wire clips will grab the flange area properly. I had to thin the conrners out to a normal sized 120 mm flange area as The Beast has much more meat in the flange area on both sides of the fan. I will also have to pull the rear case fans off and leave the case side off to keep from having too much backpressure too.

burebista, it's a P6T problem and not a Gulftown problem per se. If you clock with the multi like I was doing, it would drop the multi down in various stages once the cores hit around 65 C, which really sucks. So I went the other way with it and overclocked by raising the bclock instead, like you would do with a non-extreme processor. Kind of defeats the purpose of spending a grand on an unlocked proc though. This bios is only about half baked as far as Gulftown support goes. I hope Asus comes up with an updated bios soon.

 

[burebista]

I dunno if it helps you but take a look at unclewebb's latest Throttlestop. It's a tool designed for laptops but for me is the perfect replacement for RMClock on my E8400 desktop.
Quoting from Kevin

I added a new feature to ThrottleStop that lets some Core i users adjust their turbo boost TDP/TDC power limits. The result is more turbo boost before throttling.

 

[muddocktor]

Heh, got finished with making runs with the heatsinks I have on hand, so I decided to mount The Beast on the Venomous X, which was the last heatsink I had tested. I removed the 2 exhaust fans on the case and had to swap from my Kingston DDR3-2000 ram to some Corsair DDR3-1600 ram because of the tall heatspreaders on the Kingston stuff and then mounted The Beast on the heatsink. I fired it up to make sure everything is solidly mounted and left the case side off because of the ridiculously high air movement of The Beast. I'm now waiting for my computer room to cool back down to testing temps so I can make a run with this configuration.

 

[hokiealumnus]

Methinks I see one of those 76mm thick deltas.

 

[muddocktor]

Methinks I see one of those 76mm thick deltas.

Nope, Deltas are for wimps. This happens to be the most powerful 120 mm DC case fan I know of, the Sanyo Denki 9CR1212P0G03, which moves around 80 more cfm than the Delta GFB and at a much higher static pressure too.

 

[Prunecandy]

im a little dissapointed..i expected two more of those fans on the top of the heatsink from the doctor.

 

[Dolk]

If he did a push pull, his system would take flight


Or you could get what this guy has:

http://www.xtremesystems.org/forums/showthread.php?t=254326

 

[Prunecandy]

If he did a push pull, his system would take flight

yeah but the temps would be great
and it would be awesome to see a flying pc

 

[muddocktor]

If he did a push pull, his system would take flight


Or you could get what this guy has:

http://www.xtremesystems.org/forums/showthread.php?t=254326

Actually, that one doesn't move much more air than The Beast; about 40 cfm more. And the beast doesn't need to use an adapter to fit a normal sized heatsink either. The Beast also has almost 2 inches of static pressure too; higher than any fan I've messed with and the highest I've seen on a 120mm fan (including the Delta fans). And this Denki pulls a steady 75 watts of power while running, with a startup load of around 100 watts (these are measured values, BTW). You have to have a good psu just to play with this bugger.

yeah but the temps would be great
and it would be awesome to see a flying pc

Yeah, but over $100 by the time you get it delivered and put heads on the wires is a bit too much for me. It hurt bad enough buying just one of them.

 

[ehume]

One thing I forgot: when adding a third fan on the D14, use 25mm Megahalems clips and a fan gasket for vibration damping. You can get a third set of pads, clips and pegs from Noctua, but using the Noctua clips for the third fan will interfere with the clips for the middle fan.

 

[muddocktor]

ehume, thank you for the info on using the mega fan clips. They worked very well for me. I was able to use 3 Yate loon High fans on it for testing, using the 1 Noctua 120 mm fan clip (in the middle) and 2 sets of the mega fan clips on both ends. And that configuration worked very well too, if I must say so.