Danger Den Cpu Block

[monk3y]

I have a tdx water block for my loop, and the temps are awful 45c idle 55-60c under load. Should I ditch it and switch to something else?

 

[p3sty]

wc specs? please + ambient temp

 

[Dew itt right]

Not sure how different the TDX is from the MPC but my Danger Den block keeps my Ci7 920 at 3.6GHz (1.325 Vcore) at 68C under full load. P3sty is correct, we need more info because it's probably the pump or radiator (or maybe even the TIM) that is your limiting factor. If those temps are for your overclock then they may be normal given your clock and voltage...

 

[monk3y]

Black Ice Pro 120.2
Black Ice Xtreme 120.2
Laing D5 fixed speed
Swiftech MCW-60
Unknown Res
3/8" Tubing

ambient temp 22C

 

[Zoson]

3/8" Tubing

Well thar's yer problem.
Go 1/2" for better flowrate and better c/w.

 

[Dew itt right]

Well thar's yer problem.
Go 1/2" for better flowrate and better c/w.

No offense but I highly doubt he'd see much difference from uping his tubing size from 3/8" to 1/2". The total loop restriction in his case is almost entirely in his components not to mention the fact that he's got a very strong pump. My whole loop is 3/8" plus I have a much weaker pump. Not to mention the fact that I have an external system that sits more than 5ft from the PC so I've got over 12ft of 3/8" tubing in my loop and I still haven't found the ceiling of my Core i7 920 (currrently at 3.8GHz 1.325Vcore).

Are the temps of your overclock? The Q6600 @ 3.6GHz in your sig? If so, what's the Vcore?

Looks like you're cooling the GTX285 as well? Is the CPU the first cooled component after the radiators? Are you WC'ing the NB, SB, or anything else?

 

[Zoson]

No offense but I highly doubt he'd see much difference from uping his tubing size from 3/8" to 1/2". The total loop restriction in his case is almost entirely in his components not to mention the fact that he's got a very strong pump. My whole loop is 3/8" plus I have a much weaker pump. Not to mention the fact that I have an external system that sits more than 5ft from the PC so I've got over 12ft of 3/8" tubing in my loop and I still haven't found the ceiling of my Core i7 920 (currrently at 3.8GHz 1.325Vcore).

Are the temps of your overclock? The Q6600 @ 3.6GHz in your sig? If so, what's the Vcore?

Looks like you're cooling the GTX285 as well? Is the CPU the first cooled component after the radiators? Are you WC'ing the NB, SB, or anything else?

http://www.procooling.com/index.php?func=articles&disp=38&pg=1
Flowrate suffers about a 50% drop in 3/8" vs 1/2" systems. Strong pump or not, that's a significant difference. The only reason there is to EVER use 3/8" is if you are in a space limited situation. Learn yourself and then come talk to us again.

 

[Bigbacon]

http://www.procooling.com/index.php?func=articles&disp=38&pg=1
Flowrate suffers about a 50% drop in 3/8" vs 1/2" systems. Strong pump or not, that's a significant difference. The only reason there is to EVER use 3/8" is if you are in a space limited situation. Learn yourself and then come talk to us again.

How does this effect temps?

I am using 3/8" as well and it still does an excellent job at removing heat and keeping temps much lower than Air.

I am running a hotter system than his and my temps idle around 32-34c. (room temp is about 77F)

Changing out the tubing may help, but there has to be another explanation to the high temps. I don't see the 1/2" dropping those temps signifigantly.

 

[Zoson]

How does this effect temps?

I am using 3/8" as well and it still does an excellent job at removing heat and keeping temps much lower than Air.

I am running a hotter system than his and my temps idle around 32-34c. (room temp is about 77F)

Changing out the tubing may help, but there has to be another explanation to the high temps. I don't see the 1/2" dropping those temps signifigantly.

First, nobody cares about your system. You don't have the same parts, now do you? No comparison can be made.

What really matter is his fans. You can have a single 120mm rad with a 250cfm fan on it and do as well as a 120.3 with only 40cfm fans.

The question becomes: "with his current radiator, how can he get his temps down"
1. switch to 1/2 so he has more water absorbing heat per unit time
2. switch to high speed fans so he has more air absorbing heat per unit time.

A change in waterblock won't do anything if the radiator can't dump the heat.

Switching to 1/2 doesn't cause an increase in noise like the fans, so naturally it is the best choice.

 

[monk3y]

I used to have 1/2" tubing in my rig, I really didn't see much of a difference. Right now I am only cooling my cpu and gpu. My current cpu speed is 3.48ghz, and since I am using a 780i I had to crank up the voltage to 1.42 because of the vdrop. even when I run my cpu @ stock speed/ voltage I see only slightly lower temps.

 

[Zoson]

even when I run my cpu @ stock speed/ voltage I see only slightly lower temps.

This tells the whole story. Your radiator can't deal with the heat load. put higher cfm fans on, or get more/bigger rads.

 

[Dew itt right]

First, nobody cares about your system.

Quiet down everybody, GOD has spoken! I didn't realize Zason the English version of Zeus?

I loved the link you sent me, did you read it yourself? Your proving my point. The lowest flow on that chart was 2.57L/min which inside a 3/8" tube comes to 24 inches per second! Is that too slow for your super-duper cooling loop? It's a closed system so flow rates at that level are not that critical. I'm a mechanical engineer so I have "learned myself". Lower the flow rate and the temp will rise more in the blocks because it spends more time in the block but it'll also cool more in the radiator because it's spending more time there as well. On an extremely smaller level, the flow rate will only control the rise and fall in the loop but when you're you're comparing the cooling difference between 24 in/sec and 48 in/sec you might as well be debating the scope of Zoson's intellect vs. that of a monkey! At both flow rates the water spending insanely small fractions on a second in the block and very little time in the radiator as well. It's negligible...

This tells the whole story. Your radiator can't deal with the heat load. put higher cfm fans on, or get more/bigger rads.

This I can agree with. Either not enough air thru the rad or possibly the mounting of the block. Have you tried reseating the wb or using better TIM? My apparently ignorant opinion would be that the 2 2x120mm rads are just fine.

 

[Erasmus354]

Please stop posting if you don't know what you are talking about. Nobody here has really said anything that is completely correct or useful.

@Zoson:
Get a more up to date comparison between different tubing sizes. Cathar did a good one on this forum about a year or two ago. There are various versions of Martin's flow rate estimator spreadsheet floating around that are very good as well. Bottom line, tubing size makes much less impact on flow rates these days. Watercooling pumps are beefier than back in the stone ages of the link you provided. Also most waterblocks are more restrictive than back then so the tubing plays a proportionally smaller roll in overall flow rates.

@Dew itt right:
The slower the flow is, the worse the transfer of heat between components becomes. Stop thinking about temperature differences and start thinking about energy per unit time, or in other words Watts. At a lower flow rate you decrease the turbulence and therefore the thermal efficiency. This equates to a higher temperature gradient and a poorer overall cooling system. More flow rate is always better if you can ignore the heat dump from the pump. There are diminishing returns on the effect of flow rate on cooling and therefore in the real world there is a point when the pump puts out more heat than the gains in flow rate achieve so you end up with a worse overall performance.

 

[Zoson]

Quiet down everybody, GOD has spoken! I didn't realize Zason the English version of Zeus?

I loved the link you sent me, did you read it yourself? Your proving my point. The lowest flow on that chart was 2.57L/min which inside a 3/8" tube comes to 24 inches per second! Is that too slow for your super-duper cooling loop? It's a closed system so flow rates at that level are not that critical. I'm a mechanical engineer so I have "learned myself". Lower the flow rate and the temp will rise more in the blocks because it spends more time in the block but it'll also cool more in the radiator because it's spending more time there as well. On an extremely smaller level, the flow rate will only control the rise and fall in the loop but when you're you're comparing the cooling difference between 24 in/sec and 48 in/sec you might as well be debating the scope of Zoson's intellect vs. that of a monkey! At both flow rates the water spending insanely small fractions on a second in the block and very little time in the radiator as well. It's negligible...

While we're throwing credentials around, I'm an electrical engineer, from an ivy league college far better than whatever school you went to.

The very simple matter of fact is that if you're moving twice the amount of water through there, you have twice the ability to pick up heat(well, not really but to illustrate the point), depending on your temperature delta. There's literally twice as much material going through to pick up heat.

The ability of water to pick up heat decreases as the delta between water temp and element temp shrinks. In otherwords, you pick up heat faster with cooler water. You get a higher volume of cooler water over your equipment with higher flowrate. So no matter how long your water 'heats up' in your block you will never be able to match high flow performance simply because high flow has water at a greater dT constantly cooling.

Anyway, you clearly did NOT read the article, as the entire POINT is that high flow is significantly better.

@Erasmus: The additional restriction of new blocks does in fact make tubing size that much more important. Isn't the point to eliminate any avoidable limitations? I still have an AthlonXP true 1/2" system with one of those old pumps, and it gets considerably higher flowrate than my new system for my system in my sig. Every little bit matters. The point I was making is that 1/2" is flat out better when it comes to performance, period. Also, pumps are NOT much better. The pump people use most commonly nowdays is that wimpy MCP 350 or whatever, which is definitely less powerful than the Hydor L30, Eheim 1250, or the original swiftech pump from the article. Hell, even the MCP655 isn't much better than the pumps I just named.

 

[Dew itt right]

While we're throwing credentials around, I'm an electrical engineer, from an ivy league college far better than whatever school you went to.

...if you're moving twice the amount of water through there, you have twice the ability to pick up heat...

...There's literally twice as much material going through to pick up heat...

...The ability of water to pick up heat decreases as the delta between water temp and element temp shrinks. In otherwords, you pick up heat faster with cooler water. You get a higher volume of cooler water over your equipment with higher flowrate...

Apparently they don't teach tact or thermodynamics in ivy league schools anymore. I bet you're just a gem to work with...

Look, Erasmus is right, turbulence causes greater conductive heat transfer between the fluid and the capillary tubes in the radiator thus a higher flow rate is better, but only minimally at these flow rates. You've already reach the point of diminishing returns. Ask anyone with a variable MCP655 pump to turn the speed down to 1 record temp, crank up to 5 and record temps again. I'd almost guarantee the flowrate at least doubles between 1 and 5 and they'll only see a 1-3C temp drop (I've done this myself).

Or, if you want to follow Zoson here, just put a second D5 in parallel to double your system flow rate and you'll magically double your temperature differences!

Anyway, OP, sorry for bickering in your thread. I hope you found something helpful here. Best of luck...

 

[Erasmus354]

@Erasmus: The additional restriction of new blocks does in fact make tubing size that much more important. Isn't the point to eliminate any avoidable limitations? I still have an AthlonXP true 1/2" system with one of those old pumps, and it gets considerably higher flowrate than my new system for my system in my sig. Every little bit matters. The point I was making is that 1/2" is flat out better when it comes to performance, period. Also, pumps are NOT much better. The pump people use most commonly nowdays is that wimpy MCP 350 or whatever, which is definitely less powerful than the Hydor L30, Eheim 1250, or the original swiftech pump from the article. Hell, even the MCP655 isn't much better than the pumps I just named.

More flow rate is always better, I already stated that in my previous post. My point was that tubing size actually plays a pretty minor factor in today's systems in overall system flow rates. I told you to go find more up to date information but because you refuse to do that I will post it for you.

Cathar's analysis of tubing sizes on flow rate: http://www.hardforum.com/showthread.php?t=1200495

Also the MCP 350 is not the most common pump these days, that would be either the D5 (MCP655) or the DDC3.2 (MCP355). And both of those pumps are much better than the 1250. The problem with the hydor and the eheim pumps are their poor head pressure. They aren't as suited to pumping through restriction like the high head pressure DC pumps are.

You seem to be pretty long out of the loop Zoson. Your information would have been more correct about two to three years ago.

Using the final version of Martin's Flowrate Estimator spreadsheet I come up with this for the OP's system:

D5 Pump
5 feet of 3/8" tubing : ~1.61GPM
5 feet of 1/2" tubing : ~1.71GPM

DDC 18W Pump
5 feet of 3/8" tubing : ~1.51GPM
5 feet of 1/2" tubing : ~1.56GPM

Eheim 1250 Pump
5 feet of 3/8" tubing : ~1.25GPM
5 feet of 1/2" tubing : ~1.33GPM


As you can see the tubing size is only a small factor and the Eheim 1250 lags significantly behind. There is a reason why people stopped using aquarium pumps and started using DC pumps and it isn't just convenience.



Back to the matter at hand...... is the block a classic TDX or the newer MC-TDX? If it is the classic TDX that is a VERY old block and you would certainly see strong gains by switching to a newer block.

 

[LanPirate1]

Erasmus354.. I have to agree with you.
Two months ago I switched all my fittings from 3/8" to 1/2 and all the tubing in the same exact bends and lengths. My blocks, rad, pump and res all the same units. I did testing and found I might have dropped maybe 1-2C overall CPU temp loaded, not what I was expecting. No degree, or knowledge can replace real world testing. If there is anything I've learned in this world, it is that.

 

[RabidCicada]

Erasmus and Dew are two very well informed guys.

For summary of this entire thread:

Flow Rate-->Higher flow rate is always technically better but, assuming you have "reasonable" flow rate to begin with, the gain is only marginal at cpu watercooling speeds (not enough of a change is achievable to really matter).

TubeSize-->At cpu watercooling flow rates, assuming you have "reasonable" flow rate to begin with, tube size change only affects you marginally (favoring 1/2 inch tubing-->slightly higher flow rate).

monk3y-->Your problem lies in one of your watercooling COMPONENTS.
Any of the following could be your issueump broken, rad can't handle your heat dissipation to begin with(for whatever reason), waterblock not diong job, or thermal paste applied incorrectly/outdated thermal paste.

 

[game14]

First try reseating your block and for second ,i see you are using Black Ice radiators.Those dont work well with low speed fans, so i dont know what kind of fans you are using but it might be that your current fans dont push/pull enough air through your radiators.
On the other hand the TDX is not really the best block for cooling a quadcore.
just my 2 cents,

 

[E.E.L. Ambiense]

I have a tdx water block for my loop, and the temps are awful 45c idle 55-60c under load. Should I ditch it and switch to something else?

Is this for the system in your sig currently?

I confess I'm not much of a DD product user, but am familiar with their products that I used in commissioned builds previously. The TDX is an older block designed for dual-core CPU's; the MC-TDX updates the design with quad support. The older block doesn't cover all the cores properly versus the new block. That may have a little something to do with your temps.

The main difference between the new MC-TDX and the old TDX is that the new MC-TDX has 240 pins over the CPU core area. The MC-TDX was designed especially for quad core processors that carry much more heat than the dual cores when they are pushed and that is where the MC-TDX excels.

 

[Zoson]

You seem to be pretty long out of the loop Zoson. Your information would have been more correct about two to three years ago.

Using the final version of Martin's Flowrate Estimator spreadsheet I come up with this for the OP's system:

D5 Pump
5 feet of 3/8" tubing : ~1.61GPM
5 feet of 1/2" tubing : ~1.71GPM

DDC 18W Pump
5 feet of 3/8" tubing : ~1.51GPM
5 feet of 1/2" tubing : ~1.56GPM

Eheim 1250 Pump
5 feet of 3/8" tubing : ~1.25GPM
5 feet of 1/2" tubing : ~1.33GPM


As you can see the tubing size is only a small factor and the Eheim 1250 lags significantly behind. There is a reason why people stopped using aquarium pumps and started using DC pumps and it isn't just convenience.

As I can see, it's only a small factor, which I said was small. Honestly though you just disproved what you were saying all along. The difference is GREATER with the D5, which you say is the most common pump.

Way to repeat what I said about his radiator, everyone.

 

[monk3y]

Alright I think as soon as I get some extra cash I will pick up some new fans, and maybe a new water block.

 

[Erasmus354]

As I can see, it's only a small factor, which I said was small. Honestly though you just disproved what you were saying all along. The difference is GREATER with the D5, which you say is the most common pump.

Way to repeat what I said about his radiator, everyone.

Not if you look at it as a percentage change it isn't....

 

[Zoson]

d5 = 6% change
ddc = 3% change
eheim = 6% change

Did you even run the numbers? The second pump you gave as a recent pump is significantly worse. lol.
And the d5 is no better than the eheim. Which is exactly what I said in the first place.

 

[Hidetaka]

Flowrate suffers about a 50% drop in 3/8" vs 1/2" systems. Strong pump or not, that's a significant difference.

Proven wrong, tubing's got close to no influence on flow and therefore temperatures.

This tells the whole story. Your radiator can't deal with the heat load. put higher cfm fans on, or get more/bigger rads.

If his temps are bad no matter the heatload [OC vs. stock] it would suggest bad IHS-waterblock contact more than anythging else... If it was all about insufficient rads he'd notice a difference when lowering clocks and voltage.

Also, pumps are NOT much better. The pump people use most commonly nowdays is that wimpy MCP 350 or whatever, which is definitely less powerful than the Hydor L30, Eheim 1250, or the original swiftech pump from the article. Hell, even the MCP655 isn't much better than the pumps I just named.

Proven wrong again.

d5 = 6% change
ddc = 3% change
eheim = 6% change

Did you even run the numbers? The second pump you gave as a recent pump is significantly worse. lol.
And the d5 is no better than the eheim. Which is exactly what I said in the first place.

You've just preven that D5 and Eheim's flow drop more than DDC's which is hardly relevant... What it shows:
- Eheim is ~20% slower than the "wimpy" DDC
- 3/8 tubing gets almost as much flow as 1/2 - so there's no point in replacing it [and the barbs] thus wasting money.

Maybe it's time you stop trolling and do some reading on more recent watercooling hardware?

The first advice he should get is: 1. reseat your WB - not enough pressure or too much TIM can increase temps even a couple degrees. 2 Report back.

PS. Is 60* really that bad? If it's the highest it reaches I would just let it be.

 

[Zoson]

Proven wrong, tubing's got close to no influence on flow and therefore temperatures.

If his temps are bad no matter the heatload [OC vs. stock] it would suggest bad IHS-waterblock contact more than anythging else... If it was all about insufficient rads he'd notice a difference when lowering clocks and voltage.

Proven wrong again.

You've just preven that D5 and Eheim's flow drop more than DDC's which is hardly relevant... What it shows:
- Eheim is ~20% slower than the "wimpy" DDC
- 3/8 tubing gets almost as much flow as 1/2 - so there's no point in replacing it [and the barbs] thus wasting money.

Maybe it's time you stop trolling and do some reading on more recent watercooling hardware?

The first advice he should get is: 1. reseat your WB - not enough pressure or too much TIM can increase temps even a couple degrees. 2 Report back.

PS. Is 60* really that bad? If it's the highest it reaches I would just let it be.

You're bashing flowrate and arguing tim joint? lol. Any appreciable amount of pressure forces out most tim problems unless the person REALLY gooped it on. Not only that, if he had poor seating on his block - his water wouldn't be getting hot and stabilizing.

And actually, the difference between the new ddc and the old eheim IS 50% not 20%. So what was your argument again? half of 6% is 3%. Almost as much flow doesn't matter because flowrate is not linearly proportional to temperature drops. Not only that I proved that the eheim 1250 has the exact same performance as the D5. In fact, this information came out of Erasmus, confirming what I said. So no, I wasn't wrong, I was RIGHT.

So why don't YOU stop trolling, and learn yourself.

 

[filthysanchez]

Flowrate suffers about a 50% drop in 3/8" vs 1/2" systems

d5 = 6% change
ddc = 3% change
eheim = 6% change

The difference between 6 percent and 3 percent isn't 50 percent


Read this