It is currently Thu Oct 02, 2014 7:20 am

All times are UTC - 8 hours




Post new topic Reply to topic  [ 31 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sat Feb 15, 2014 12:04 pm 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
I was thinking to replace the Prolimatech cooler currently used on a Core i7 2600K as I might need it on another build.

So I looked to my spares collection to find out a suitable substitute: among them there's a Scythe Orochi Rev. B no longer used now, and I thought it would have been able to handle the Core i7, as it's so huge...

...well, long story short, I was plainly wrong: after a few seconds of 8 instances of Prime95 the system shut down.

I'm a bit dazed and confused: that monster cooler is a bit outdated but it's still pretty massive and apparently undamaged.
May a CPU heatsink suffer ageing, or second generation Intel Core is just too hot for that old giant?

_________________
Regards,
Luca


Last edited by quest_for_silence on Sun Feb 16, 2014 12:42 am, edited 2 times in total.

Top
 Profile  
 
 Post subject: Re: May a CPU heatsink suffer ageing? (A Scyte Orochi epic f
PostPosted: Sat Feb 15, 2014 5:21 pm 
Offline
Moderator

Joined: Thu Oct 06, 2005 4:36 am
Posts: 4638
Location: Monterey Bay, CA
there are more likely scenarios. Here's two:
- poor fit
- you installed in a case with rotated mobo and the the heatpipes are pointed the wrong way. Causing them to not work.

_________________
1080p Gaming build: i5-4670K, Mugen 4, MSI Z87-G45, MSI GTX 760 2GB Gaming, 8GB 1866 RAM, Samsung Evo 250GB, WD Red 2TB, Samsung DVD burner, Fractal Define R4, Antec True Quiet 140 (2 front + rear) case fans, Seasonic X-560. 35-40W idle, 45-55W video streaming, 170-200W WoW, 200-230W Rift, 318W stress test (Prime95 + Furmark)


Top
 Profile  
 
 Post subject: Re: May a CPU heatsink suffer ageing? (A Scyte Orochi epic f
PostPosted: Sun Feb 16, 2014 12:41 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
CA_Steve wrote:
there are more likely scenarios. Here's two:
- poor fit
- you installed in a case with rotated mobo and the the heatpipes are pointed the wrong way. Causing them to not work.


The latter reason can't be the one, as the mobo was on my test bench.

About the first scenario, actually it was one of my thought about, but then I has given the wrong subject to this thread (I correct it right now).
Now I've sent a message to Scythe technical support, but I hope there are other people here with an Orochi on a LGA1155 platform.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sun Feb 16, 2014 2:22 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
quest_for_silence wrote:
May a CPU heatsink suffer ageing, or second generation Intel Core is just too hot for that old giant?


According to you, folks, in case of erroneous mounting, shouldn't I have an even high idle temperature?

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sun Feb 16, 2014 6:34 am 
Offline
Moderator

Joined: Thu Oct 06, 2005 4:36 am
Posts: 4638
Location: Monterey Bay, CA
In general, a poorly mounted heatsink will exhibit higher idle temps, too...but, how much higher depends on a couple of factors...like how much power the cpu uses at idle and what's the C/W of the heatsink even with the poor fit?

In other words, a poorly mounted massive heatsink is still going to provide some level of thermal transfer and with a low idle power cpu might be good enough to compare to a less massive cooler.

Hey, it might be the heat transfer fluid in the heat pipes has leaked out/gone away while it was in storage...but the most likely scenario is poor mounting.

_________________
1080p Gaming build: i5-4670K, Mugen 4, MSI Z87-G45, MSI GTX 760 2GB Gaming, 8GB 1866 RAM, Samsung Evo 250GB, WD Red 2TB, Samsung DVD burner, Fractal Define R4, Antec True Quiet 140 (2 front + rear) case fans, Seasonic X-560. 35-40W idle, 45-55W video streaming, 170-200W WoW, 200-230W Rift, 318W stress test (Prime95 + Furmark)


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sun Feb 16, 2014 8:24 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
CA_Steve wrote:
Hey, it might be the heat transfer fluid in the heat pipes has leaked out/gone away while it was in storage...but the most likely scenario is poor mounting.


As said, the cooler looks undamaged: the heatpipes show some oxydation, but that's all (even the box is pristine, inside and out, there are no stains or any sign of leakages).

After some tests it looks like the Orochi can't handle the same heat of the Prolimatech, and if it were true, then it would be a bid odd IMHO (the Orochi Rev. B aimed at LGA1366 CPUs which were very hot).

At any rate, I've found that Orochi can handle the Core i7 2600K at stock settings, even if the relevant temps aren't spectacular ones: and so I guess it may rule out any poor mounting as the main reason of the initial shut down, but I need some more fine tuning.

I've also privately asked to the forumer boost the relevant load temp for his 3570K/Orochi combo, meanwhile I will do some testing disabling the Hyperthreading.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sun Feb 16, 2014 9:44 am 
Offline
*Lifetime Patron*

Joined: Sun May 21, 2006 9:09 am
Posts: 1866
Location: Northern California.
what temps do you get with just one or half the cores running?
even with all the heat pipes broken, and using peanut butter for tim, that cooler has enough mass to keep an i7 cool for more than 8 seconds.
-something is not making proper contact.-

post a picture of the pattern the tim left on the chip and heat sink.

_________________
“America, its a continent, not a county!”


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sun Feb 16, 2014 10:19 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
xan_user wrote:
even with all the heat pipes broken, and using peanut butter for tim, that cooler has enough mass to keep an i7 cool for more than 8 seconds.


No xan, it's (almost all about) a matter of heat: with 8 threads at 4.5GHz it took less than 4 seconds to skyrocket from 25°C to 98°C and shut down, while with the same 8 threads at 3.5GHz the Orochi worked well.

Even better, it also handled about fine several hours of Prime95 (at about 75-80°C with just one 500rpm fan): please take note I've done this further test without any TIM as I was very tired to apply and remove the Arctic Silver.

I've also managed to perform about 45 minutes with 8 threads at 4.2GHz, undervolting the CPU of about 110mV, but the test system isn't stable with such a massive undervolting (it freezed, while it works indefinitely disabling hyperthreading), and above all I cannot undervolt the CPU on the final rig.
Right before freezing max core temp was stabilized at about 85°C, still without any TIM.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Sun Feb 16, 2014 8:38 pm 
Offline
*Lifetime Patron*

Joined: Sun May 21, 2006 9:09 am
Posts: 1866
Location: Northern California.
I see.
-just a thought, maybe its a lack of airflow over the mobo/vrm ect.?

_________________
“America, its a continent, not a county!”


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Mon Feb 17, 2014 12:54 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
xan_user wrote:
-just a thought, maybe its a lack of airflow over the mobo/vrm ect.?


Well, xan, on the Intel mobo I have a dedicated sensor on VRMs, and it stayed well below 60°C when the system shut down, and below 55°C when testing at stock settings. So I don't think the main failing reason may be something related to the power circuitry.

I think it's just a matter of how the Orochi is designed, and how it works: it cannot benefit of any increase in airflow, it performs about the same with one 800rpm fan or with three 1900rpm fans, and lowering the fan speed to about 200rpm gives less than 10°C penalty. Moreover it has a big buffer heatsink over the 10 heatpipes which acts somehow as a damper, as opposed to any tower coolers, and this led to the consequence I cannot apply the same scheme of things.

Naively I've thought to swap the Prolimatech tower with it, close the enclosure, and then fire it up again indefinitely: it doesn't work so.
So I was forced to temporarily put a Core i5 on that rig, and testing for a couple of days the Core i7 with the Orochi on an ASUS Pro mobo on my test bench.

With any CPU voltage above 1,32V the Orochi reacts badly, and I definitely cannot lower the total CPU voltage under 1.4V with hyperthreading enabled and Turbo Boost OC (any level), as there's a steep step in the voltage set by Intel when stock settings are any way exceeded (so I should undervolt the system of about 160mV or more, but it cannot even boot with such undervolting, not to mention the final Intel mobo is not able to undervolt at all).

Summarizing, with the Prolimatech cooler (two fans, at 7-800rpm max) I reach the stability limit before incurring the overheating limit.
With the Orochi (one fan, 500rpm max) the opposite scenario is true, I reach the thermal limit before hitting any stability issue, and I'm prevented to add more airflow to the cooler, and/or enough undervolting to the CPU, by design.

Just a second thought: a 32% overclock squeeze a 21% of more processing power over stock settings, giving a 25°/30°C of temperature penalty (with the Orochi, while it is about a 18/20°C with the Prolimatech)... does it worth? I'm reconsidering this thing...

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Orochi and CPU TIM footprint
PostPosted: Mon Feb 17, 2014 6:01 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
Just to be complete: after the last two days testing, I removed the cooler and took a shot with the phone camera.
That's what I saw:

Attachment:
CPU.jpg
Attachment:
Orochi_base.jpg


Decidedly nothing good looking, but also nothing to be alarmed for: the paste is spread towards the border, but there's too much of it on more than half of the inner region of the CPU heatspreader.

It would seem the "rice grain" strategy doesn't work optimally with the Orochi base (or with this heatspreader), but that does not seem to affect (at least, too much) the stock performance as well: with 8 threads of Prime95 I managed to have the hottest core either around 50°C (undervolted) or below 60° (stock voltage), with just a single fan under 500rpm.


You do not have the required permissions to view the files attached to this post.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Mon Feb 17, 2014 11:59 am 
Offline

Joined: Mon Mar 17, 2008 1:49 pm
Posts: 1203
Location: UK
Consider the thermal capacity of the cooler and the power output of the CPU.

The Orochi is 1.1kg approx and aluminium has a heat capacity of around 0.9 J/g/k. This means it will take around 1100J to heat up the heat sink by 1K. So it would take around 80000J to heat up the CPU by the temp you suggest and in 4 seconds that would equate to a TDP of 20kW!

This thought experiment assumes perfect heat conduction which clearly isn't true but it proves that you have not got good conduction. Big heat sinks have high heat capacity and even if they are not well ventilated it will still take a long time to get up to temp. As the temp rises fast, you have another problem.

Have you checked that the CPU shim has not become dislodged from the core? That would give you catastrophically bad conduction. This can cause these problems and they're a real bugger to fix and in some cases result in binned processors...

_________________
Silverstone SG03B: E8200, Asus P5E-VM HDMI, 2Gb RAM, Leadtek 9600GT+S1 rev. 2, Samsung 500Gb, Seasonic X-400, 2x Akasa 120mm, Scythe Zipang 2 fanless


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Mon Feb 17, 2014 3:26 pm 
Offline
Moderator

Joined: Thu Oct 06, 2005 4:36 am
Posts: 4638
Location: Monterey Bay, CA
Those striations at the center of the cooler and IHS show that you aren't making decent contact. The paste should be squeezed down to a very thin layer - and it isn't in your case.

_________________
1080p Gaming build: i5-4670K, Mugen 4, MSI Z87-G45, MSI GTX 760 2GB Gaming, 8GB 1866 RAM, Samsung Evo 250GB, WD Red 2TB, Samsung DVD burner, Fractal Define R4, Antec True Quiet 140 (2 front + rear) case fans, Seasonic X-560. 35-40W idle, 45-55W video streaming, 170-200W WoW, 200-230W Rift, 318W stress test (Prime95 + Furmark)


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Mon Feb 17, 2014 5:48 pm 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
edh wrote:
Big heat sinks have high heat capacity and even if they are not well ventilated it will still take a long time to get up to temp. As the temp rises fast, you have another problem.


I'm not sure to understand what you said.


edh wrote:
Have you checked that the CPU shim has not become dislodged from the core? That would give you catastrophically bad conduction. This can cause these problems and they're a real bugger to fix and in some cases result in binned processors...


Providing that you call "shim" the CPU heatspreader, AFAIK in a Sandy Bridge CPU it is joint to the core with fluxless solder, so I argue it cannot be dislodged without destroying the CPU.


CA_Steve wrote:
Those striations at the center of the cooler and IHS show that you aren't making decent contact. The paste should be squeezed down to a very thin layer - and it isn't in your case.


Steve, I'm not an heatsink: *IT* isn't making a good contact. As said in my previous post, "nothing good lookng" under that base, but nothing to worry about also.

That is the original and undamaged, shiny Orochi base: more probably that not, it isn't a very good one, as already seen on several other coolers, like the Ninja 3, or the quoted Dark Rock, but at any rate I think I can not do anything to improve the relevant performance (the mounting mechanism has no adjustments/settings).

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Mon Feb 17, 2014 8:39 pm 
Offline
Moderator

Joined: Thu Oct 06, 2005 4:36 am
Posts: 4638
Location: Monterey Bay, CA
quest_for_silence wrote:
Steve, I'm not an heatsink

What about a teapot?

_________________
1080p Gaming build: i5-4670K, Mugen 4, MSI Z87-G45, MSI GTX 760 2GB Gaming, 8GB 1866 RAM, Samsung Evo 250GB, WD Red 2TB, Samsung DVD burner, Fractal Define R4, Antec True Quiet 140 (2 front + rear) case fans, Seasonic X-560. 35-40W idle, 45-55W video streaming, 170-200W WoW, 200-230W Rift, 318W stress test (Prime95 + Furmark)


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Mon Feb 17, 2014 11:44 pm 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
CA_Steve wrote:
quest_for_silence wrote:
Steve, I'm not an heatsink

What about a teapot?

Do I mumble/whistle too much?

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Tue Feb 18, 2014 9:50 am 
Offline
*Lifetime Patron*

Joined: Sun May 21, 2006 9:09 am
Posts: 1866
Location: Northern California.
- in picture # 2 i see that the socket adapter plates are bolted on so that they can flex more. is there room to flip those, so the hexnuts are on top, with the arms bent up, away from the mobo?
I assume youre also using a backplate?

_________________
“America, its a continent, not a county!”


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Tue Feb 18, 2014 10:37 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
xan_user wrote:
- in picture # 2 i see that the socket adapter plates are bolted on so that they can flex more. is there room to flip those, so the hexnuts are on top, with the arms bent up, away from the mobo?
I assume youre also using a backplate?


The mounting mechanism is the latest Scythe one (before the newer seen on the Mugen 4): http://www.scythe-eu.com/en/products/pc ... ate-4.html

I've already tried to flip the arms facing upwards, but I ended up with comparable results, temp-wise (though I've no shots of the relevant TIM footprint).

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Wed Feb 19, 2014 3:47 pm 
Offline

Joined: Mon Mar 17, 2008 1:49 pm
Posts: 1203
Location: UK
quest_for_silence wrote:
edh wrote:
Big heat sinks have high heat capacity and even if they are not well ventilated it will still take a long time to get up to temp. As the temp rises fast, you have another problem.


I'm not sure to understand what you said.


If you had a massive lump of metal (which the Orochi basically is) it would take a lot longer to get up to temperature than a small bit of metal. As your CPU temp shoots from 25C to 98C in 4 seconds, there is no way the Orochi can heat up that fast. Just put your hand on the Orochi and there's no way way that it is that hot, over 50C is painful for a few seconds so this is an easy way to test cooler temp at normal loads. Therefore the CPU is not in proper contact with the Orochi. It could be your thermal paste striations as Steve points out, it could be the heatspreader coming loose (this can happen after pulling a previous cooler off) or it could be some kind of physical compatibility problem. The Orochi however is easily big enough for that processor and you should give up all trials with fans, you're wasting your time if the cooler is not in good contact.

So, in summary, the laws of physics mean that your CPU can not possibly have good thermal conductivity to the CPU cooler.

_________________
Silverstone SG03B: E8200, Asus P5E-VM HDMI, 2Gb RAM, Leadtek 9600GT+S1 rev. 2, Samsung 500Gb, Seasonic X-400, 2x Akasa 120mm, Scythe Zipang 2 fanless


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 2:57 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
edh wrote:
So, in summary, the laws of physics mean that your CPU can not possibly have good thermal conductivity to the CPU cooler.


Well, I think that while the broad principle is right by general consensus, some over-simplifications you may do affect the relevant conclusion. I mean the adverb "possibly" is the key, but take also note that my english might not be up to the task.

This Orochi Rev. B is able to handle the 2600K at load (8 Prime95, small FFT instances), I think even passive even if I tried it just down to 200rpm, providing the CPU is clocked at its stock frequency (3.5GHz, Vore about 1.25V).

Whether the CPU is overclocked at 4.5GHz, Vcore about 1.5V, with the stock fan at about 500rpm (something lower actually) the Orochi can't handle the very same 8 Prime95, small FFT, instances, and things happen in a very quick way.

Which law of physics evidences that there is a contact "problem", or a not "good thermal conductivity to the CPU cooler"?

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 5:46 am 
Offline

Joined: Mon Mar 17, 2008 1:49 pm
Posts: 1203
Location: UK
quest_for_silence wrote:
Which law of physics evidences that there is a contact "problem", or a not "good thermal conductivity to the CPU cooler"?


As I already stated, to heat up the cooler from 25C to 98C will require a lot of energy. Here's the working out:

Energy = temperature difference x heat capacity x mass = (98-25) x 0.897 x 1155 = 75630.55 Joules

Power = Energy / Time = 18907 Watt

Now ask yourself: Does your CPU have a TDP of 18907 Watt?

No, it doesn't. It's 0.5% of that. Therefore it is physically impossible for your CPU to be heating up your cooler by that much in 4 seconds. Therefore heat conduction from the CPU through to the cooler is not working properly.

_________________
Silverstone SG03B: E8200, Asus P5E-VM HDMI, 2Gb RAM, Leadtek 9600GT+S1 rev. 2, Samsung 500Gb, Seasonic X-400, 2x Akasa 120mm, Scythe Zipang 2 fanless


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 5:53 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
edh wrote:
No, it doesn't. It's 0.5% of that. Therefore it is physically impossible for your CPU to be heating up your cooler by that much in 4 seconds. Therefore heat conduction from the CPU through to the cooler is not working properly.


As said, I think you make some unnecessary over simplifications (if you let me to tease, when you have to put a screw in a hole, then you need a screw driver, and you should not use an hammer): if I correctly understand your point, it's the CPU to heat up itself, and not the heatsink cooler.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 6:35 am 
Offline
*Lifetime Patron*

Joined: Thu Feb 13, 2003 6:53 am
Posts: 1525
Location: Sweden
Rapid temp increase despite a huge cooler (with or without fan) indicates that something is wrong. Bad contact, broken heatpipes, something. Because as edh points out, if the heat is transfered properly, physics tells us that the temperature rises rather slow.

Check if the base of the cooler is flat. Maybe some lapping can improve things.

_________________
Main: ASRock B85M-ITX | i3-4330 | 16GB DDR3 | Intel 730 240GB | HDPLEX H1-S | picoPSU | No moving parts | Idle 13.9W
HTPC: ASRock H81M-ITX | Pentium G3420 | 4GB DDR3 | Intel 520 120GB | HDPLEX H1-S | picoPSU | No moving parts | Idle 11.2W
Gaming: Intel DH77EB | i5-3570K | GTX 750 Ti | 16GB DDR3 | Intel 520 120GB | TJ08-E | G-360 360W
Server: Intel DH77DF | i3-2100T | 4TB+3x3TB | picoPSU | Idle 24W AC


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 6:49 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
Vicotnik wrote:
Rapid temp increase despite a huge cooler (with or without fan) indicates that something is wrong. Bad contact, broken heatpipes, something. Because as edh points out, if the heat is transfered properly, physics tells us that the temperature rises rather slow.


Vicotnik, you would seem to forget that any transfer speed is finite: so given an object you may simply overcome the inner transfer rate and result in a sort of thermal collapse.
And you forget that violent temp increase doesn't occur when the CPU package power is under 95W (actually under an about 115W mark), and the Vcore under the about 1.3V mark (actually under the 1.39V mark).
Moreover you also insist about the cooler size, but it's definitely less important than how a cooler work (and I would add how a specific CPU work): a truck can be a lot more powerful than a car, but it can't be as fast as that.
Last but not least, on the cooler there's no sign of poor/inefficient thermal contact, or of any mechanical damages: you can't have a murder without the murder weapon.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 7:53 am 
Offline
*Lifetime Patron*

Joined: Thu Feb 13, 2003 6:53 am
Posts: 1525
Location: Sweden
quest_for_silence wrote:
Vicotnik, you would seem to forget that any transfer speed is finite: so given an object you may simply overcome the inner transfer rate and result in a sort of thermal collapse.

Sure, but then you would have a scolding hot heatsink. And I think you do not. :) Since it doesn't seem to matter how much air you throw at it.
quest_for_silence wrote:
Last but not least, on the cooler there's no sign of poor/inefficient thermal contact, or of any mechanical damages: you can't have a murder without the murder weapon.

That imprint in the paste is an indication of sub-par contact. In the middle, where the core is. Could be a murder weapon, I'm not sure.

_________________
Main: ASRock B85M-ITX | i3-4330 | 16GB DDR3 | Intel 730 240GB | HDPLEX H1-S | picoPSU | No moving parts | Idle 13.9W
HTPC: ASRock H81M-ITX | Pentium G3420 | 4GB DDR3 | Intel 520 120GB | HDPLEX H1-S | picoPSU | No moving parts | Idle 11.2W
Gaming: Intel DH77EB | i5-3570K | GTX 750 Ti | 16GB DDR3 | Intel 520 120GB | TJ08-E | G-360 360W
Server: Intel DH77DF | i3-2100T | 4TB+3x3TB | picoPSU | Idle 24W AC


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 8:16 am 
Offline

Joined: Wed Jun 13, 2007 10:12 am
Posts: 3071
Location: ITALY
Vicotnik wrote:
quest_for_silence wrote:
Vicotnik, you would seem to forget that any transfer speed is finite: so given an object you may simply overcome the inner transfer rate and result in a sort of thermal collapse.

Sure, but then you would have a scolding hot heatsink. And I think you do not. :) Since it doesn't seem to matter how much air you throw at it.


I don't know what "scolding hot" means, so I don't understand your remark, sorry. :wink:

Broadly speaking, the Orochi is sort of two-stage cooler, where the smaller secondary heatsink is the one which take the heat from the pipes at first, and where just the half of those pipes are at direct contact to the heatsink base.

Long story short, currently I don't think the massive primary fins aren't enough to handle a 130-140W load, and as a matter of fact, they do that: they do that with the 2600K at the 4.2GHz level - something less, actually - and, before being "retired", I used the Orochi to cool an Intel X3220 (a Q6600 G0), oc'ed from 2.4 to 3.6GHz.

With reference to this latter, at those figures the CPU package power was comparable with the Core i7 one at load (just something above 130W, IIRC). But that Kentsfield have a power envelope at idle around 30W and more, not just 5W (like the Sandy), and I guess that far higher base-heat let the dual-stage ten heatpipes to reach the relevant "peak-torque point" (that doesn't happen with the Sandy Bridge above the 1.3V vcore).

I have no Nehalem to compare with, in order to look for some further (maybe more conclusive) evidence of my thoughts about, but up to now I have no clue to suspect something radically different.

_________________
Regards,
Luca


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 8:57 am 
Offline
*Lifetime Patron*

Joined: Sun May 21, 2006 9:09 am
Posts: 1866
Location: Northern California.
this has me stumped. i see no reason why orochi cant handle the heat.
going back and looking at the pics, one other thing strikes me as odd. in the first pic it looks to me like the cpu hold down tabs are not evenly centered on the heat spreader. the tab on the top looks to be barely making contact, while the bottom one looks to have a way better hold.

grasping at straws here... -could rapid heating of the spreader and cpu hold down allow the cpu to lose good contact with the heat sink, but return to normal after cooling off??

edit: and in pic two i swear i can see the indent of the hold down tab in the tim on the base of the heatsink.... i wont say on which side yet, so you all can see if you are seeing it too.

(clearer pics would help...)

_________________
“America, its a continent, not a county!”


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 9:18 am 
Offline

Joined: Mon Mar 17, 2008 1:49 pm
Posts: 1203
Location: UK
quest_for_silence wrote:
As said, I think you make some unnecessary over simplifications

I am a Physicist and I have done a lot of customised heatsink installations attaching old Socket A heatsinks to things they were never designed to be connected to. :twisted: So I do know my stuff and I can tell you right now that your CPU is not in good thermal contact with your CPU cooler. Please forget your preceonceptions.

quest_for_silence wrote:
it's the CPU to heat up itself, and not the heatsink cooler.

Correct. The CPU is heating up an awful lot in your system because the CPU cooler is not in good thermal contact.

quest_for_silence wrote:
Vicotnik, you would seem to forget that any transfer speed is finite

The transfer of heat varies with temperature. The higher the temperature differential, the higher the rate of heat transfer. This is why if there is good thermal contact you would see a slow temperature rise whioch starts tailing off.

quest_for_silence wrote:
Moreover you also insist about the cooler size, but it's definitely less important than how a cooler work (and I would add how a specific CPU work): a truck can be a lot more powerful than a car, but it can't be as fast as that.

Irrelevant anaology with a car and a truck. F=ma has nothing to do with heat.

I have already stated how you work out the heat capacity of an object, the mass of the object is a term in it. With the Orochi you have a massive heat capacity. If it was no good at radiating the heat, that would be a different thing, it would take ages to heat up but carry on getting hotter. That is not happening. The Orochi does not heat up. Your CPU is not in good thermal contact with it.

quest_for_silence wrote:
on the cooler there's no sign of poor/inefficient thermal contact, or of any mechanical damages: you can't have a murder without the murder weapon.

It's thermal contact, nothing else. Whether or not you can see it or not is not important. You could always try switching back to another cooler (do you still have the Intel cooler?) just to see what happens.

quest_for_silence wrote:
I don't know what "scolding hot" means

It means hot enough to burn your skin. If the cooler was in good thermal contact (which it isn't) then it would be at the same temperature as the CPU itself. If the CPU is at 98C then the cooler would also be at 98C. It won't be. In fact I'll bet you my house that it isn't. It will be hovering around the ambient temperature inside the case, maybe around 30C.

quest_for_silence wrote:
I used the Orochi to cool an Intel X3220 (a Q6600 G0), oc'ed from 2.4 to 3.6GHz.

That CPU has a TDP of 105W, more than the 95W of the 2600K. It is therefore more than good enough to cool the 2600K, if it is in good thermal contact.

So again, it's not in good thermal contact. Are we now clear on that?

In terms of cause, I have know coolers that have not been in good contact where it is OK at idle and terrible at load. Uneven pressure across the core could be a reason if the cooler is not properly level on the core.

Things to try:
1. Check the temperature of the cooler during operation. 50C feels hot to the touch so if the cooler stays cool then it is in bad thermal contact.
2. Try with the case lying down flat so that the cooler isn't putting excessive moment on the CPU socket. The Orochi is very heavy and if the weight of the Orochi is the problem, this will resolve it.
3. Try with another CPU cooler like the standard Intel cooler. If it is still bad then the cause could be the heatspreader coming off the CPU itself.

_________________
Silverstone SG03B: E8200, Asus P5E-VM HDMI, 2Gb RAM, Leadtek 9600GT+S1 rev. 2, Samsung 500Gb, Seasonic X-400, 2x Akasa 120mm, Scythe Zipang 2 fanless


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 10:15 am 
Offline
*Lifetime Patron*

Joined: Thu Feb 13, 2003 6:53 am
Posts: 1525
Location: Sweden
quest_for_silence wrote:
Broadly speaking, the Orochi is sort of two-stage cooler, where the smaller secondary heatsink is the one which take the heat from the pipes at first, and where just the half of those pipes are at direct contact to the heatsink base.

In the past heatsinks were just lumps of aluminum with fins to make the area in contact with air larger. Modern CPU coolers are better, thanks to copper bottoms, heatpipes etc. But when we do rough calculations we can assume that the cooler is a lump of metal.
I think you overthink the design of the Orochi. Even if it's not as good as the modern coolers of today, there is no way that it performs worse than a simple lump of metal, given that the heatpipes are working.

The less efficient the heatsink is, and the bigger the thermal load, the bigger the delta-T (difference in temperature between CPU and heatsink) will be. That is why you with bad thermal connection between CPU and heatsink still can have pretty good temperatures, as long as the thermal load is small. Pulling a few numbers out of my ass; 10W CPU (or powerful CPU at idle) gives us 40C CPU and 35C heatsink. Delta-T is 5C. Fan blows ambient air (say at 25C) at the heatsink, cooling it.
Then we start Prime95. We now have a 100W CPU and CPU temp is 90C with the heatsink rising only to 45C due to lousy contact. If you blow much more air onto the cooler you can get that down some, but since the delta-T is so high it doesn't really matter if you with push-pull Delta fans at 5k RPM can get heatsink temp down to very close to ambient. Delta-T is still huge and the CPU overheats.

Even if you want to look at the Orochi like a two-stage cooler, then the "base" heatsink close to the core should be very, very hot if contact is good. I don't think it is.

_________________
Main: ASRock B85M-ITX | i3-4330 | 16GB DDR3 | Intel 730 240GB | HDPLEX H1-S | picoPSU | No moving parts | Idle 13.9W
HTPC: ASRock H81M-ITX | Pentium G3420 | 4GB DDR3 | Intel 520 120GB | HDPLEX H1-S | picoPSU | No moving parts | Idle 11.2W
Gaming: Intel DH77EB | i5-3570K | GTX 750 Ti | 16GB DDR3 | Intel 520 120GB | TJ08-E | G-360 360W
Server: Intel DH77DF | i3-2100T | 4TB+3x3TB | picoPSU | Idle 24W AC


Top
 Profile  
 
 Post subject: Re: Can a Scythe Orochi be mounted on LGA1155/1156/1150?
PostPosted: Thu Feb 20, 2014 11:43 am 
Offline
*Lifetime Patron*

Joined: Sun May 21, 2006 9:09 am
Posts: 1866
Location: Northern California.
Quote:
then the "base" heatsink close to the core should be very, very hot if contact is good. I don't think it is.

if the heatpipes were broken, pointing a high speed/cfm fan at the bottom heatsink, should let the pc run longer than 4 seconds. im thinking with the bottom heatsink only and a high output fan, it should be pretty close to a stock intel cooler.

_________________
“America, its a continent, not a county!”


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 31 posts ]  Go to page 1, 2  Next

All times are UTC - 8 hours


Who is online

Users browsing this forum: Google [Bot], Yahoo [Bot] and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group