temps fans push pull who would have guessed ?

[freeboy]

OK, long title for this..
But really, I have been using my badass Skythe .6 amp fans that push a published 165cf.min at 3300 rpms as my main " these are awesome" fans I bought another rad and figured I could add some more Skythe fans, but they where hard to find at resonable pricing, so I did some digging and came up with some monster Delta fans... on a deal. They look actually identical to the Skythe fans, other than they are way heavier, and slightly larger diameter moter housing, but they look like clones otherwise....
I maxed out my single 360 SR-1 black ice and have OC room but temps, they got hot using only the Skythe fans.. so I switched to the Deltas, and other than when at max power sounding like Delta Airlines jets, work fine. I saw a measurable decrease in temps...OVER 15%.. from the 3 Skythe fans to the three Delta fans. (delta rate at 256cf/min and 3.1 amps... monsters).
So people have used a push pull, and I figured my former BA Skythe fans should at least help...
I hooked them up to push air hopefully reducing the workload and increasing actual push through the rad...... as well as giving me more options for sound management, after all who wants to have a desk shaking experience with their computer? ok dont answer that one!
So I add three Skyth fans, now we have six fans , three Skythe pushing and three Deltas pulling ...Next
So, I boot up, load Real Temp... run Intel Burn Test and ...
I gained ONE little degree on one core with the 3 Skythe added to the system's existing three Deltas... I think I could swap fans around with 2 Deltas on one side and a Skythe and one on the opposite side, but Im shocked .... wow...
Im getting more Deltas... and then we will see if PUSH pull is that much better than PUSH or Pull by itself.. I know these fans are irregular, and most folks push pull with normal fans and it helps... but wow was I shocked to get essentially NO GAINS !!!!

 

[MxPhenom 216]

Okay.

1. Its Scythe, not Skythe
2. Take a chill pill
3. You are doing it wrong.
4. Do some research on your radiator
5. Your radiator is not optimized for high speed fans. the SR-1 radiator are specifically designed for low speed fans.(Only has 9fpi) It is a capable radiator at higher speeds don't get me wrong. However, that is why you are not seeing the scaling you seem to expect from adding push/pull or crazy high cfm/high speed fans.

You will see on the following review, that the SR-1 pretty much dominates at fan speeds at ~600-1000rpm, then it drops down through the list as the fan speeds get higher.

http://martinsliquidlab.org/2012/04/08/hardware-labs-sr1-360/

Also, pay more attention to fan specs. CFM does not matter when it comes to radiators and heatsinks. You want to pay attention to static pressure. "The ability a fan has at pushing air through something restricting like fins." The only reason I can suspect that the Delta fans did any improvement is that they had a bit higher static pressure. They were able to force air all the way through the radiator.

 

[freeboy]

my point is adding the Scythe fans added no help whatsoever..AND I do realize its about static presure, just they don't publish those numbers... Ill switch over to my more restrictive rad, a GTX xtreme 360, similar to the SG-1 but optimized for greater fan metrics(pressure). I did read up on Martins testing.. but really, are you saying you would not expect a change, even a few degrees? Its not the rad dude, its the fans... the Deltas so dominate the equation...
Im pulling and switching rads as the next step, and buying three more Deltas, too bad the deal I had went away, price jumped up...

Note that when I changed from the Scythe to the Deltas I had a big drop..better fan more static pressure.... I am not totally sold on the idea of six fans in push pull.... does anyone have some actual data? from there own system.. I was not and am not angry, its the limitation of the medium, More like I was lampooning myself and posting something which surprised me....

 

[MxPhenom 216]

my point is adding the Scythe fans added no help whatsoever..AND I do realize its about static presure, just they don't publish those numbers... Ill switch over to my more restrictive rad, a GTX xtreme 360, similar to the SG-1 but optimized for greater fan metrics(pressure). I did read up on Martins testing.. but really, are you saying you would not expect a change, even a few degrees? Its not the rad dude, its the fans... the Deltas so dominate the equation...
Im pulling and switching rads as the next step, and buying three more Deltas, too bad the deal I had went away, price jumped up...

Note that when I changed from the Scythe to the Deltas I had a big drop..better fan more static pressure.... I am not totally sold on the idea of six fans in push pull.... does anyone have some actual data? from there own system.. I was not and am not angry, its the limitation of the medium, More like I was lampooning myself and posting something which surprised me....

you don't need the push pull because the deltas are already pushing enough air for the other 3 fans that are weaker to be useless.

 

[d1nky]

I got my black ice xtreme 360 in push pull with corsair sp120s highperformance fans.

they do pretty well on it, I did try just push but wasn't as good as push/pull

the rad is restrictive and martins lab even noted that the higher the rpm fans the better, whereas their site says otherwise!

 

[freeboy]

awesome, I love Martins testing.. truly detailed and thorough, I also like real world applications...
P/P with six deltas would help.. Im going to add the rad asap.. amybe just do two rads as I have lots of fans...
thanks again guys...
The deltas wher 3 shipped for 46, now its 3 shipped no discount free shipping same seller for 75... ouch
Not that its that critical... just playing with the oc cooling equation.... thanks again
So you like the Xtreme? I am thinking the deltas on the Xtreme and the Scythe on the SG-1...

 

[erocker]

You're not going to see a change in your max temperatures due to the "poor" contact between the CPU and the IHS. No amount of ambient cooling will fix that. Maybe you should look into delidding.

 

[freeboy]

What are you saying? I already have made changes and seen great decreases.... I gain some room, than oc some more..
I understand there is some concern with the glueing and deliddign as AN option, but really saying; "You're not going to see a change in your max temperatures " is not acurate.. I will continue to see both changes and higher clocks, and if I chose to de lid, then Im sure that I will again see changes..
That was an odd post!
Im running over 1.425 and have highs in the upper 70's... at 4.9.. those really are not bad temps and lead me to think I may be the top 1/3 of chips....

Delidding is an option, I agree with that, just not a MUST DO...
I understand you are a mod.. but I think you should reconsider the TONE of your post!

 

[RCoon]

What are you saying? I already have made changes and seen great decreases.... I gain some room, than oc some more..
I understand there is some concern with the glueing and deliddign as AN option, but really saying; "You're not going to see a change in your max temperatures " is not acurate.. I will continue to see both changes and higher clocks, and if I chose to de lid, then Im sure that I will again see changes..
That was an odd post!
Im running over 1.425 and have highs in the upper 70's... at 4.9.. those really are not bad temps and lead me to think I may be the top 1/3 of chips....

Delidding is an option, I agree with that, just not a MUST DO...
I understand you are a mod.. but I think you should reconsider the TONE of your post!

Adding more powerful fans and more radiators is not necessarily going to cool it better than it already is. You cant cool something lower than ambient with radiators and fans. Eventually you reach close to ambient, you're still pumping that hot air into the room, which then gets sucked back in eventually. No fans or radiators will fix the crap contact between IHS and die.
What I'm saying is you're going to hit a wall of the thermal dissipation possible with fans. Trust me, I added Nidec Beta V's to a 240 rad, had them running at 4500RPM, and they made NO DIFFERENCE to dissipation over generic Corsair stock fans.
If you want improvements, you're going to need to delid, or lap, or use coolab. After that, you're still not going to get temps close to ambient unless you go extreme cooling.

tl;dr - adding more dissipation wont fix poor contact. Those fans arent going to make the IHS conduct better.

 

[freeboy]

not looking at ambient, using for OC, Im in the 80's at 5k.. validated but not stable, high 70s C at 4.9...
Im trying to gain some headroom, not bring below ambient.... obviously you cannot go below the temp you are using to cool with...
Im moving 17C air through.. and Im gaining ground with each step, not deliddingyet, Im going to see the edge of the envelope then I may delid if it looks like Im stuck... But honestly I think you are missing the point...

By adding a better fan rad setup I can get a lower onder heavy load max temp, lower temp then higher oc...
Not 100% sure where yo are going with the delidding, its not about lowering the temp for gaming, my temps are low when not under artificial testing loads and I run gaming at 4.8 or lower OC and a little higher mem state... thats fine

The goal is to get into the mid 5's in the OC, and I will get there... on water ...
so, I may have to de lid, Im not saying its not a good idea... but even delidding and getting some extra room is leaving a rad fan window, I am also using AN OLD Danger Den block.... this will go in time as well... but for now its what I have.....

interesting the Apagee HD is out of stock everywhere.... hummm.. new block coming perhaps?
So, Im going to add the other rad soon, and will keep all you fun loving folks outthere informed of the progress...

Heck Im just happy I hit 5 and 4.9 stable...

 

[RCoon]

360 rad on just the CPU? I dont think adding another rad is going to make any difference at all.

 

[freeboy]

ok, thanks for the input..

 

[RCoon]

ok, thanks for the input..

You first need to improve on the thermal conductivity of the components before you start adding area for dissipation. This includes delidding, or change of paste(coolab pro), or having the system pump water over the CPU block faster than your current pump, or lapping the block/IHS. When theres more heat being conducted from the cpu to the block, and from the block to the liquid, THEN adding more dissipation might be useful, but not necessary. You might get a couple of degrees improvement at best.
Direct die contact is also a better solution, but very dangerous.

 

[freeboy]

I completely dis agree
But
Thank you

 

[RCoon]

I completely dis agree
But
Thank you

You can disagree with physics if you want.

"A: The more radiators you add the lower RPM your fans can run to achieve the same heat dissipation. This is also true the other way around. With sufficient airflow you could cool a whjole system with one 120mm radiator. The typical recommendation is 240mm of rad space per main component, however, after two components you can start adding 1 radiator per component. By main component I of course mean GPU or CPU, thing like RAM and motherboard cooling should not be considered in these calculations as they can withstand a lot of heat and don't produce much. Hence why they are normally cooled passively.

When adding rads after a certain point will you get to a point of diminished returns. Becase of the extra loop length your coolant will flow slower around your loop which may actually increase your temps. However, you should only worry about this if you are going overkill with your set-up. (Having more than 10 rad slots would be about that point). You could add another pump to counter this, however, once your coolant is a room temp it cant go lower so adding 100 rads would have the same effect as adding 10 if the coolant is a room temp with both solutions."

 

[fullinfusion]

I'm against delid period, money out the window imo.. just play with different fans. I used to run a 30fpi black ice 240 and it worked ok but I needed more sp. My noctuas worked fine so after getting some corsair sp120's there was a difference but not as much as I xpected.
My acool 9 fpi 45mm loved the corsairs and really cooled ofer the higher fpi rad.. weird but rpm and sp work hand in hand for my cooling. I can run my 3770k all day n night @ 5guz and I top the mid 80's. I'm not a corsair fan boy but they do make some great fans. Get a two pack of the sp120's for cheap and give us hem a go in push pull. Some may argue how loud hey are but I think they are not loud as some make them out to be.

 

[freeboy]

"You can disagree with physics if you want."

ok..
read the fine print and CHECK THE ATTITUDE
the error noted is about pumping sped or coolant flow rate... IF my flow rate remains constant, above the "Magic" one gallon per minute Martin of ML refers too, the PHYSICS as you so rightly state cannot be argued.. that is if flow rate is constant or above the magic "one" above then the heat, lets say adjusted or lowered after first rad heat goes into tubing then out to rad two.
In rad two, we get cooling.... the cooling is just a product of physics, so my 88C goes down if Ambient is steady.... I have a 16 C ambient in a room that is 1150 Sq Feet...
So, I could do rad rad pump or any combination as long as I get good flow rate....
then we see decrease as ambient is 16 or some other much lower temp... and the coolant flows through, each successive radiator would dissapate heat, simple physics.... if you could keep flow rate up and do any number of rads..
RESPECTFULLY dissagree....

 

[RCoon]

Refer to my previous post about improving flow rate to improve temps by a fractional amount(by improving conductivity), which you seemingly disagreed with en masse.
I'm out. Best of luck with your ventures.

 

[freeboy]

You could add another pump to counter this

for flow, and again
" once your coolant is a room temp it cant go lower "
So how am I not paying attention to physics?
I have great flow, almost two GPM, and adding one rad and small amount of tubing is not a big add on to ressitance, I calculate a 7-12% decrease in temps... from a 65C de=difference from ambient, or about 6 -8 degrees.. lets see..

 

[d1nky]

what rcoon is saying is partly true (theres too much to read tho)

I was told;

lets say the cpu to block heat exchange is 70% effective, the water conducts x% of that heat, and dissipation is x%

more rads will cool better but not at a scale that is effective enough to see or counter the costs, if 40% of the heat is getting dissipated by x amount of rad the scale would be very slight thereafter.

those are guessed numbers, im not into thermal physics but some people on this site are.

 

[freeboy]

ok.. if we look at what is posted
ML refers to terms like ambient .. cannot go below ambinet, check, and flow rate..
So lets stay on focus here...
If flow rate stays constant... then additional rads will continue to push temps towards ambient.. this is a true logic statement.. so the question that should be asked is cna the rads work at a decent flow rate?
AND
can my fans keep up...
Fans and rads I own... tubing as well, I need to get a 1.00 clamp. So, I also have about 31 OZ of my primochill rad fluid.. so adding another radiator and going to 720 from 360.. or 6 X 120 from a good manufacturer... one is more constrictive and will get the delta fans, the Xtreme...

So, if my pump can handle to extra resistance then I should be ok... If the pump cannot handle the extra resistance I will add my second pump....
go figure...

This is not really that complicated...

 

[d1nky]

This is not really that complicated...

overkill on cooling is the best kind of overkill, and has its benefits.

I would just do it tbh, I want to add more rads/thicker rads just for the looks and added benefits.

(plus its waiting for some decent gfx)

 

[MxPhenom 216]

I'm against delid period, money out the window imo.. just play with different fans. I used to run a 30fpi black ice 240 and it worked ok but I needed more sp. My noctuas worked fine so after getting some corsair sp120's there was a difference but not as much as I xpected.
My acool 9 fpi 45mm loved the corsairs and really cooled ofer the higher fpi rad.. weird but rpm and sp work hand in hand for my cooling. I can run my 3770k all day n night @ 5guz and I top the mid 80's. I'm not a corsair fan boy but they do make some great fans. Get a two pack of the sp120's for cheap and give us hem a go in push pull. Some may argue how loud hey are but I think they are not loud as some make them out to be.

Yep, Intel did what they did with the IHS and TIM for good reason, just like NVidia has their Green program. It is so that the chips heat throttle before actually being killed, from people who get all voltage happy and kill the chips.

Even Dave does not recommend delidding for his own various reasons, and I trust his word for it more then anyone else.

overkill on cooling is the best kind of overkill, and has its benefits.

I would just do it tbh, I want to add more rads/thicker rads just for the looks and added benefits.

(plus its waiting for some decent gfx)

Its wasted money IMO. If you are already getting load temps close to ambients, spending $50+ on another radiator will not justify the cost or effort.

@Freeboy, you are running an 84w TDP chip, with a radiator capable of dissipating ~200w of heat. Even with your chip overclocked, your radiator has no problem dissipating the heat. Unless you are running a high TDP GPU in the same loop, another radiator could help, but if your just running CPU. your not going to see enough benefit that makes the hassle worth it.

 

[freeboy]

ok, we will see wont we... I really think we are beating a dead horse, the physics of cooling just are not this complicated... hot coolant into second rad, ambinet air cools coolant in second rad..
I would have to say that you may be too close to the issue, step back and think about the last post....
We are talking about simply the function of a rad...
It does not matter to our second rad that it is a second... it might as well be a first or a hundredth.. the only thing it does is take colder ambient air and cool the coolant...
Weather I am wasting time and effort is my concern, its simple physics that IF WE CAN PUMP the coolant additions rads will lower high temps, given they are fanned correctly..

This is really self evident and your post refers to a judgement about how much cooling i need...
Im at 5k on my overclock.. I want more.. and really the only think holding any HAswell Over clocker back is temperature and learning there new z87 bios.

So, I respect you have an opinion.. and I respectfully suggest stepping back and looking at Martins Lab quote from above warning taken in the positive that says if you can keep coolant flow at a proper level addition fans will push temps toward ambient..
How much toward ambient is all about the functions of the components.. how much heat in watts are we dealing with, how well do the rads work, the fans the pumps etc....

So.. Im testing asap and will let everyone know how another rad does and how the pump does...
thanks for yoru interest, it should be fun

 

[MxPhenom 216]

if you want to be overclocking that far. Should really consider some sort of phase or LN2 cooling. Id like to see your validation of a 5ghz Haswell chip too.