Help!!! Are these temps high for my setup???

[Radi_SVK]

Edit.... OK Rado, I think I see what you mean. I suppose I could send an email to Corsair saying that the cooler head might not be totally flat as per the picture.

exactly my friend.you dont need to mention anything about using an OCed cpu and higher temps as consequences.build your RMA reasons on the uneven cold plate of the cooler.after all,you are not making it up are you?its a fact.

Of course. Hence my mention of IHS issues long before that

I personally dont know how rare this issue is,but dont you think that down at Silicon Valley they do the job a little bit more precisely than at Corsair?

 

[cadaveca]

I personally dont know how rare this issue is,but dont you think that down at Silicon Valley they do the job a little bit more precisely than at Corsair?

No, they do not. IHS's are almost never flat. It's more rare to find a flat one than to find one that isn't flat.


FYI cpus are assembled in Malaysia, not the US.

 

[Radi_SVK]

FYI cpus are assembled in Malaysia, not the US.

THe Silicon Valley was just a phrase

 

[cadaveca]

The IHS is almsot never perfectly flat. That is why so many people lap thier cpus...i have this same problem, but value my warranty more than a higher overclock.

Meanwhile, I am pushing 4.0ghz, and 2400mhz ram, with 4x 2GB sticks.

 

[PaulieG]

Every company will have a way out on the RMA thing. Corsair does not advertise these extremely low temps on overclocked chips. Therefore they won't RMA a cooler performing badly on any overclocked system and if you read the box it came in, the graph that they show is based on a test run on an i7 920 running the stock speed of 2.66 GHz.... I have an i7 930 which runs stock at 2.8 GHz. Their argument is that naturally the cooler won't give me the same temps as what's on the box because it's a different chip. So in my case an RMA is out of the question.



Edit.... OK Rado, I think I see what you mean. I suppose I could send an email to Corsair saying that the cooler head might not be totally flat as per the picture.

Corsair is an easy company to deal with. They would most likely accept the RMA with anything close to a valid reason. If you want to try for some lower temps with slightly less effort with mounting etc., I would try to RMA the cooler. It may or may not help since your IHS is concave, but in my mind, trying everything is worth a shot because I'm a little obsessive about low temps. If you want to try the best coolers available, here is a short list of the best coolers on the market. There are some others that are good, but I own 2 of the three, and temps for both are remarkable. From every review I've read, the Venomous-X and Noctua D14 are simply the best coolers on the market, short of a custom water loop. The Venomous-X is my personal favorite. All should give you better load temps than what you are seeing on that H70.

http://www.frozencpu.com/products/1..._Bracket_System_Socket_LGA_775_1156_1366.html

http://www.frozencpu.com/products/8...U_Heatsink_LGA_775_1156_1366_AM2_AM2_AM3.html

http://www.frozencpu.com/products/9..._Fans_-_Socket_775_1156_1366_AM2_AM2_AM3.html

 

[pabloottawa]

Hey guys,

Glad to see we're back at it and i have relatively good news.

I redid the entire thing and instead of following AC5 instructions I used my brain and limited knowledge of physics and how things spread and bond when they are pressed and spread a very thin layer on the CPU AND cooler head. My guess (and I think it's a better guess that place a 1mm line down the middle and hope it spreads evenly) is that both sides will make a far better bond and once cured will yield far better results. BTW, I also checked both CPU and cooler head and they are flat as a pancake!!! They really could not be any flatter. Furthermore I switched the connections around to a setup that makes more sense.

To rule out the possibility of the pump not running at full tilt, I connected it directly to the PSU and connected both fans to the CPU FAN header and set QV FAN CONTROL to "ENABLED". Now if the temps go up the fans will spin faster which to me makes more sense.

Now the test. Ran Linx and the highest temps are:

Core 0 88c
Core 1 85c
Core 2 82c
Core 3 79c

All around good news and definitely not an apples to apples comparison given that I switched the connections AND I removed the resistors that Corsair says to put on if the fans are too loud.

The downside is that my system is a little loud but I'll see if I can tweak that out.

Check pics

 

[PaulieG]

Hey guys,

Glad to see we're back at it and i have relatively good news.

I redid the entire thing and instead of following AC5 instructions I used my brain and limited knowledge of physics and how things spread and bond when they are pressed and spread a very thin layer on the CPU AND cooler head. My guess (and I think it's a better guess that place a 1mm line down the middle and hope it spreads evenly) is that both sides will make a far better bond and once cured will yield far better results. BTW, I also checked both CPU and cooler head and they are flat as a pancake!!! They really could not be any flatter. Furthermore I switched the connections around to a setup that makes more sense.

To rule out the possibility of the pump not running at full tilt, I connected it directly to the PSU and connected both fans to the CPU FAN header and set QV FAN CONTROL to "ENABLED". Now if the temps go up the fans will spin faster which to me makes more sense.

Now the test. Ran Linx and the highest temps are:

Core 0 88c
Core 1 85c
Core 2 82c
Core 3 79c

All around good news and definitely not an apples to apples comparison given that I switched the connections AND I removed the resistors that Corsair says to put on if the fans are too loud.

The downside is that my system is a little loud but I'll see if I can tweak that out.

Check pics

That is good news. If you decide you'd like to try for even lower temps, there is my advice above. In any event, enjoy that rig of yours!!!

 

[pabloottawa]

Corsair is an easy company to deal with. They would most likely accept the RMA with anything close to a valid reason. If you want to try for some lower temps with slightly less effort with mounting etc., I would try to RMA the cooler. It may or may not help since your IHS is concave, but in my mind, trying everything is worth a shot because I'm a little obsessive about low temps. If you want to try the best coolers available, here is a short list of the best coolers on the market. There are some others that are good, but I own 2 of the three, and temps for both are remarkable. From every review I've read, the Venomous-X and Noctua D14 are simply the best coolers on the market, short of a custom water loop. The Venomous-X is my personal favorite. All should give you better load temps than what you are seeing on that H70.

http://www.frozencpu.com/products/1..._Bracket_System_Socket_LGA_775_1156_1366.html

http://www.frozencpu.com/products/8...U_Heatsink_LGA_775_1156_1366_AM2_AM2_AM3.html

http://www.frozencpu.com/products/9..._Fans_-_Socket_775_1156_1366_AM2_AM2_AM3.html

In this case Paulieg I gotta disagree with you. You're not comparing apples to apples here. Your setup starts with an i5 whereas mine is an i7 930. I'm sure that I would not be able to overclock to 4 GHz on those massive air coolers you suggested and stay under 100. Especially on an i7 930. It seems that the 920's can remain cooler clocked higher than 4GHz under load but I'm not the guru on this stuff. Ask me about your Audi and I'll definitely be able to help lol.

I'm wondering if there's anyone out there having the same issues using an i7 930 and an H70 with the Asus P6T Cadaveca your setup is close but you're still running an 870. Seems the 930's are rare. Perhaps I chose the wrong model lol.

I just put the resistors back on to keep the noise level down and I'm topping at 91 according to Linx but I'll take it for now. In all honesty I really don't push my CPU to anything nearly as high as what these test programs do so I'm safe. AC5 hasn't fully cured yet and I'm convinced at this point that the method I used to apply the paste was the best. i guess we'll find out if a few days.

Cheers lads!!! If anyone of you are ever in Ottawa, the beers are on me!!!

Update in a few days

 

[Radi_SVK]

Hey guys,

Glad to see we're back at it and i have relatively good news.

I redid the entire thing and instead of following AC5 instructions I used my brain and limited knowledge of physics and how things spread and bond when they are pressed and spread a very thin layer on the CPU AND cooler head. My guess (and I think it's a better guess that place a 1mm line down the middle and hope it spreads evenly) is that both sides will make a far better bond and once cured will yield far better results. BTW, I also checked both CPU and cooler head and they are flat as a pancake!!! They really could not be any flatter. Furthermore I switched the connections around to a setup that makes more sense.

To rule out the possibility of the pump not running at full tilt, I connected it directly to the PSU and connected both fans to the CPU FAN header and set QV FAN CONTROL to "ENABLED". Now if the temps go up the fans will spin faster which to me makes more sense.

Now the test. Ran Linx and the highest temps are:

Core 0 88c
Core 1 85c
Core 2 82c
Core 3 79c

All around good news and definitely not an apples to apples comparison given that I switched the connections AND I removed the resistors that Corsair says to put on if the fans are too loud.

The downside is that my system is a little loud but I'll see if I can tweak that out.

Check pics

well done buddy,thats some real results in the temps decrease.

 

[OneMoar]

most 75-100 dollar air coolers are better then the h70 anyway =/
also make sure you have the fans in a push-pull setup it will help a bit

 

[JATownes]

I am afraid I have to agree with Paulieg here. There are DEFINATELY some air coolers that compete/beat the H70 regardless of what processor is being used.

Overclockers Review

Guru 3D Review

Overclock3D using i7 930

Legit Reviews

 

[PaulieG]

In this case Paulieg I gotta disagree with you. You're not comparing apples to apples here. Your setup starts with an i5 whereas mine is an i7 930. I'm sure that I would not be able to overclock to 4 GHz on those massive air coolers you suggested and stay under 100. Especially on an i7 930. It seems that the 920's can remain cooler clocked higher than 4GHz under load but I'm not the guru on this stuff. Ask me about your Audi and I'll definitely be able to help lol.

I'm wondering if there's anyone out there having the same issues using an i7 930 and an H70 with the Asus P6T Cadaveca your setup is close but you're still running an 870. Seems the 930's are rare. Perhaps I chose the wrong model lol.

I just put the resistors back on to keep the noise level down and I'm topping at 91 according to Linx but I'll take it for now. In all honesty I really don't push my CPU to anything nearly as high as what these test programs do so I'm safe. AC5 hasn't fully cured yet and I'm convinced at this point that the method I used to apply the paste was the best. i guess we'll find out if a few days.

Cheers lads!!! If anyone of you are ever in Ottawa, the beers are on me!!!

Update in a few days

My suggestions and opinions are based on 20 different i7 920's and 860's, most of which were cooled by the coolers that I mentioned above. The i5 is a new addition. Very much apples to apples.

 

[cadaveca]

I'm wondering if there's anyone out there having the same issues using an i7 930 and an H70 with the Asus P6T Cadaveca your setup is close but you're still running an 870. Seems the 930's are rare. Perhaps I chose the wrong model lol.

The setup that is in my "specs" list isn't even accurate.

With my attempts at being a reviewer, I've handled alot of parts that my own rig is not comprised of. I tend to not spend alot of time with nVidia cards, so very rarely do I ever comment on them...when I comment on nVidia, it's purely the business side of things.

I did start on i7 with a P6T, and a few other boards, as well as many cpus. Not with the H70, as it wasn't out back then...

I moved over to P55 MATX for my personal rigs just in November, 3 months ago. I'm now amassing parts for Sandy Bridge for future reviews.


As a reviewer, I need to know this stuff.

that said, I've got a H70 here, as well as this Noctua NH-C14, and the Noctua wins.

So I'm gonna hop on the bandwagon here. The H70 is good, but there are many other coolers that are fan-fin-based that compete with it, both in cost and performance, and truly, in the end, everyone's results are going to fluctuate due to problems that can arise, such as what you've just gone through.

Glad to see the the "thick tint" method worked well for you...

 

[cdawall]

In this case Paulieg I gotta disagree with you. You're not comparing apples to apples here. Your setup starts with an i5 whereas mine is an i7 930. I'm sure that I would not be able to overclock to 4 GHz on those massive air coolers you suggested and stay under 100. Especially on an i7 930. It seems that the 920's can remain cooler clocked higher than 4GHz under load but I'm not the guru on this stuff. Ask me about your Audi and I'll definitely be able to help lol.

I'm wondering if there's anyone out there having the same issues using an i7 930 and an H70 with the Asus P6T Cadaveca your setup is close but you're still running an 870. Seems the 930's are rare. Perhaps I chose the wrong model lol.

I just put the resistors back on to keep the noise level down and I'm topping at 91 according to Linx but I'll take it for now. In all honesty I really don't push my CPU to anything nearly as high as what these test programs do so I'm safe. AC5 hasn't fully cured yet and I'm convinced at this point that the method I used to apply the paste was the best. i guess we'll find out if a few days.

Cheers lads!!! If anyone of you are ever in Ottawa, the beers are on me!!!

Update in a few days

the old TRUE still beats the H70 its not great only reason i have it is i wanted a little more kick in my HTPC and H70 is the best cooler that fits

 

[pabloottawa]

Hey guys thanks for the support and feedback. I hope no one is taking offense as truly none is intended. After all you've all been there to help along the way and it is appreciated.

At this point I've already spent a ton of cash on cooling and fans and whatnot so I really want to focus on giving this some time to see if things improve. I'll go back to the clock settings and see if some voltages can be reduced and streamlined but I really want to try and make 4 gigs work. If ther are other people who can get their i7 930s up to 4ghz safely then I should be able to do it too right?

 

[PaulieG]

This is a screenie when I was testing memory recently. This was on a rather poor 920, with a 1.29v. WCG creates a significant amount of heat at 100%, just slightly below stress tests. Certainly more stressful than gaming. This was on a Megahalems, which is a step below the Venomous-X and Noctua I linked above.

Hey guys thanks for the support and feedback. I hope no one is taking offense as truly none is intended. After all you've all been there to help along the way and it is appreciated.

At this point I've already spent a ton of cash on cooling and fans and whatnot so I really want to focus on giving this some time to see if things improve. I'll go back to the clock settings and see if some voltages can be reduced and streamlined but I really want to try and make 4 gigs work. If ther are other people who can get their i7 930s up to 4ghz safely then I should be able to do it too right?

The one thing I learned about i7 is that chips are all over the place with potential. Some will do 4.0 easy, some won't do it at all within reasonable vcore. Generally the newer the batch, the worse the potential, though there are exceptions. Despite what you decide to do, good luck and we are here to help.

 

[pabloottawa]

I'm baaaaaack

Hey guys,

Update,

When we last found the fearless temp whore Pabloottawa he was whining like a child in a toy store but our noob has come a long way. Read on...

After my last post I had left the AS5 on using the "coat both CPU and cooler with thin layer using CC" method and got relatively good temps.

Since then I did some more reading and then tested some lower voltages. I had previously left everything on auto when it came to voltage but decided to see if I could run stable with it lowered just a bit more. With my 920 running at 4GHz the bios defaulted to an auto voltage of 1.425 (or something like that) and after some trials I determined that it ran stable at 1.4v. This brought my temps down a bit so I did some more research and moved to the TIM application

I tried the "pea dot" method which is just placing a pea sized amount of TIM on the CPU and carefully securing the cooler. I gotta say, this is the WORST method I've encountered. I ran a Linx test using just 1024 of my 12 GB memory and easily went into the 99 Celsius range at which point I stopped the test. I left it like this for 2 days and never reached normal load temps exceeding 82 degrees Celsius but I dared not run another load test!

This morning I bought some Arctic MX-2 Thermal paste and re-did the whole thing. I carefully cleaned out all the nooks and crannies of the cooler and CPU and primed the CPU with MX-2. Then using some tape I made a template for the new thermal paste (a rectangle oriented to the shape of how the paste should be spread over the CPU). I carefully secured the H70 cooler head to the CPU and put it all back together. After running some tests my results were FAR better than the pea-dot method and more comparable to the "coat both CPU and cooler with thin layer using CC" method. MX-2 really made a difference in this case and I will stay far away from AS5. After reading A LOT of reviews on it, it's garbage compared to other brands and it isn't even much cheaper if at all.

This said, I call upon the cooling and overclocking Gods to lend me their wisdom yet again and I ask you, with my settings, does this all look good enough to declare stable and safe? And by safe I mean, my CPU will last a good 2 years at these settings and not suffer any damage. See pic

Cheers and Thanks lads!!

EDIT!!! I FORGOT TO MENTION... These settings are all with the fan resistors in place. I just can't take the noise of it all and with them off, it only makes a difference of 2 or 3 degrees C.

 

[cadaveca]

Still a bit too hot for my personal liking, but that's due to the volts, methinks.

I like to use IBT for temp testing, myself.

 

[pabloottawa]

Hey Cadaveca,

I understand why you'd say something like that but the problem is when I begin to customize too much the system crashes, freezes up or BSOD's on me.

I know that BSOD's are usually due to hardware failure or problems but what I find so funny is that there isn't one overclocking guide that explains exactly how to overclock based on the frequency you want and the processor you have. It's very frustrating to try to learn this when the material isn't based on your setup. So i try to make sense of it but I always come up short unless I set most everything to "auto". If you have a chance, can you check out my previous BIOS screenies and see if I'm making a mistake somewhere and if-so, exactly where is it and what should it be set to keeping in mind that 4GHz is my goal.


Cheers

 

[cadaveca]

There is no formula for reaching a specific speed. There is merely a formula for testing stability.

 

[pabloottawa]

OK so what is it?

 

[cadaveca]

OK so what is it?

Depends on your definition of stability.

 

[pabloottawa]

Settings that won't crash under common load tests. Speaking of, I took your advice and changed some voltage settings around. What do you think of this?

 

[cadaveca]

I run prime until each iteration size has run, in the least. Put IBT on maximum, run for 20 runs.

 

[pabloottawa]

I just Googled it but I can't find a place to download if from.. Any ideas where I can get it?