Water vs. TRUE

[GMdoubleG]

Hello all,

I have a friend who is running an i7 with a Thermalright TRUE heatsink and he is thinking about going to water cooling so he can venture a little higher with his overclocking. My question is, will this loop (below) outperform his TRUE?

Loop (CPU only): Res->MCP355->Swiftech MCR220->GTZ->Res

Any and all comments and concerns will be appreciated.

Thanks

[SuperMiguel]

looks good but for a bit more extra if you want to use swiftech u can use the http://www.xoxide.com/swiftech-quiet-power-3x120mm-radiator.html

[rainless]

Hello all,

I have a friend who is running an i7 with a Thermalright TRUE heatsink and he is thinking about going to water cooling so he can venture a little higher with his overclocking. My question is, will this loop (below) outperform his TRUE?

Loop (CPU only): Res->MCP355->Swiftech MCR220->GTZ->Res

Any and all comments and concerns will be appreciated.

Thanks

That would beat the holy hell out of TRUE.

They just started selling water cooling stuff at the microcenter around here.

I've been thinking about changing out my pondmaster pump for one of those.

[hokiealumnus]

That would likely beat a TRUE, yes. If the price of a triple isn't too much more than he's planning, I'd go ahead and go with that. More rad never hurts (within reason of course) and will give him that much better temps. Those I7's are furnaces; water seems to be the way to go for decent clocks. :)

[GMdoubleG]

I want him to get the MCR320 but it's more about space for him. He purchased a really nice mid-tower Lian Li and we are trying to figure out how to "slightly" mod the case for a 2x120 rad.

[EarthDog]

That would likely beat a TRUE, yes. If the price of a triple isn't too much more than he's planning, I'd go ahead and go with that. More rad never hurts (within reason of course) and will give him that much better temps. Those I7's are furnaces; water seems to be the way to go for decent clocks. :)Yeah a dual rad is hardly enough for a an i7 over 4Ghz. I wouldnt go less than triple rad to get the best temps to be honest.

[CharlieCS]

Still its going to be better then a true even with 240 rad .

[baditude_df]

That would beat the holy hell out of TRUE.

I have to disagree with the "holy hell" part. More like, It will likely out perform if certain conditions exist.

Still its going to be better then a true even with 240 rad .

No, quite frankly that's not true. Water is not magic, and there are air cooling configurations out there(a good TRUE is one of them) that will whip a 240mm Rad Dedicated loop.

In fact, I'm going to go ahead and disagree with everyone who has posted in this thread so far, and I know there are other air coolers in this forum who will agree with me. There are loops that are appreciably superior to an air cooler, that is a certainty. But if the aim is to stay near silent, and to significantly best the TRUE, I pretty much guarantee you that you'll be disappointed no matter how much you spend.

Take the loop in my sig. It is for the most part better than what your friend proposes, and my TRUE keeps up with it. If I can do it, then your friend can too.

B4 I go on lets assume a few simple things:

1. That your friend knows how to properly gauge the flatness of both IHS and Heatsink base and the quality of the TIM application, to make sure he's got a good seat.
2. That he has chosen some good high pressure fans to remove the heat from in between those closely packed fins. And also that he has compared performance of a single fan to a push pull configuration.

It is not in any way worth the money to switch away from a TRUE to go to water if you're intent is to see noticeably lower temps and lower noise. If your friend is expecting either, then he's going to be disappointed. If you want noticably lower temperatures with a water loop you are at the very least going to need some high pressure/med-high speed fans which are going to be louder than you want. And on top of that,shell out for a quality loop. We're talking 4 or 5 times the price of the TRUE, and that's just a kit.

Any other air cooler for the i7 and I may not be singing this tune, but I'm speaking from experience since I've personally compared the two on the same system with the same settings. I was pleasantly surprised to see that I got the same if not better temperatures with my push pull TRUE, as I did with the high end $300.00+ quiet water loop in my sig. The only way the water beats with any significance, it is if I crank the fans up on the controller and that is loud as hell.
So will the loop win out? YES. Will it win quietly? Most definitely NO.
The max setting for the antec tricools on the TRUE is definitely bearable, and they are on a bench not 3 feet from my head.

Check this out:
http://img248.imageshack.us/img248/2893/i79204ghz1232vloadlinxt.th.png (http://img248.imageshack.us/i/i79204ghz1232vloadlinxt.png/)http://img21.imageshack.us/img21/3711/imgp0529r.th.jpg (http://img21.imageshack.us/i/imgp0529r.jpg/)http://img411.imageshack.us/img411/4476/imgp0501.th.jpg (http://img411.imageshack.us/i/imgp0501.jpg/)

This is from my testing with the TRUE the other day. Full load under LinX, HT enabled, 22degC ambient, cooling config in sig.
Sub 75 with those settings under air? Why would I go to water? Don't get me wrong, I'll keep the loop and try to improve it's performance however I can because water cooling is a hobby for me. And if your friend wants to get into WC for that reason, power to him. But if it's solely for the reasons in your OP, tell your friend to save his money and work the kinks out of his air cooling.

[ChanceCoats123]

^^^ Some people use cases you know :screwy:. Hahaha, but with that said, I think water would be the best way to go while running a case and should help to get higher clocks. I do think that an MCR320 is a necessity though. I mean I am even going to running a dedicated cpu loop with one of those and my processor doesn't pump out half as much heat.

[baditude_df]

I hate cases. ;)
But seriously, other than the slight increase in ambient within a case, there's no difference. Or, maybe there's alot of increase in ambient. Maybe airflow within the case is where the OP's friend needs to start to help lower temperatures? We don't know.
Add to that, since the TRUE is within the case you can use higher cfm fans with little to no noise increase proportionally when compared to an externally mounted rad. And you WILL need at least medium speed fans to make that rad work the way it needs to.
All of my points still stand. I'm just trying to save the guy some money.

[Metallica]

I agree with baditude, as I've never used any sort of water cooling, I sure as hell have seen lots and lots of various temps with a similar loop and I've personally used the TRUE .. when done right, the TRUE can keep up no doubt. Now, I've seen w/c setups that spank the TRUE also, but for the most part, you would be disappointed in the money spent/temp drop ratio O.o

I'm considering a nice w/c loop sometime in the future for multiple reasons

1. I want the slight decrease in temps
2. I want to look like a bad ass with tubes of glowing liquid going through my case ;d
3. The experience would be fun I think. I hear its quite addicting :p

[Daddyjaxx]

The problem with water cooling is that you either:

1. Pull cool air in and exhaust the warmer air into the case which will raise the ambient case temps.

2. Exhaust out of the case but then you aren't using cooler air running through the rad.

I have noticed that switching to the Skeleton case has improved my temps pretty good. That and moving the setup to a cooler room.

Using air cooling doesn't affect the case's internal ambient as much as watercooling IMO.

[baditude_df]

Now, I've seen w/c setups that spank the TRUE also, but for the most part, you would be disappointed in the money spent/temp drop ratio O.o

QFT. There's the ratio, but also to "spank", you would likely not have a quiet loop either. One of the requirements in the OP.


I'm considering a nice w/c loop sometime in the future for multiple reasons

1. I want the slight decrease in temps
2. I want to look like a bad ass with tubes of glowing liquid going through my case ;d
3. The experience would be fun I think. I hear its quite addicting :p

I say do it man. For all of those reasons. You won't regret it. Unless you leak into your PSU while your machine is on. Then you'll hate it. ;)

[Daddyjaxx]

Don't do it...because it IS addicting. I'm not happy unless I'm tearing it apart and redoing it somehow.

[Metallica]

QFT. There's the ratio, but also to "spank", you would likely not have a quiet loop either. One of the requirements in the OP.




I say do it man. For all of those reasons. You won't regret it. Unless you leak into your PSU while your machine is on. Then you'll hate it. ;)

Gimme like a month, and I'll be all over that **** haha.

Just need to get the funds and what not. I also want to try some DICE runs in the near future .. now which to try first :O

[CharlieCS]

baditude

240 rad will outperform true in almost all scenarios , it will be alot more expensive yes, it will take more space yes .But saying that 240 rad doesn't perform better the a true i dont agree with that at all .

[baditude_df]

I didn't lie when i said I did a direct comparison in a controlled environment with an i7 920 with a loop including a high end, dedicated TRIPLE RAD much less a double, and my TRUE kept within a degree. I didn't just bring that out of thin air. And I'm not the only one who will agree that a properly configured TRUE will out perform a piddly double rad dedicated loop. Particularly because i7's seem have been humbling even experienced watercoolers like me, with higher end loops than that since they came on the scene.

Regardless, I was surprised to say the least, but there it is for all to see.

You really need to go back and read that post carefully, then you'll notice that I did give full acknowledgment to the fact that given enough airflow and someone who knows how to build a loop properly, WC will likely win out. But if you're going to go down that road, then you can put a Delta on your TRUE as well and likely get a comparable temperature drop.

The OP asked if it would decrease temps enough to make it worth the money. I said no. He already got arguably the best i7 air cooler on the market. Many will agree that this is saying ALOT.
The OP asked if it would be quieter while giving a significant enough temperature drop to make it worth the money. Given my findings, I again said no.

I've done the legwork, I've analyzed the findings and now I'm giving the advice based on that.

What's the basis for your comments?

[CharlieCS]

I never sad u lied .U also never sad you did a direct comparison nor have you provided much data. I never put the two against each other , but the fact that the rad has bigger heat dissipation capacity then the true and that's my basis .I have no problem with you trying to save someone some money and i pretty much agreed its going to be more expensive but it will out perform a TRUE .

I love True and yeah properly configured one will go a long way , but so will a properly configured loop , maybe if you share some of your data with us and your testing methods so we can have a better insight and perhaps even change my mind .


You really need to go back and read that post carefully, then you'll notice that I did give full acknowledgment to the fact that given enough airflow and someone who knows how to build a loop properly, WC will likely win out.
I read your post carefully u started by saying
No, quite frankly that's not true. Water is not magic, and there are air cooling configurations out there(a good TRUE is one of them) that will whip a 240mm Rad Dedicated loop.

So now am not sure what you saying will a properly configured loop win out or will TRUE ?

[baditude_df]

Sure I did.
Me:Any other air cooler for the i7 and I may not be singing this tune, but I'm speaking from experience since I've personally compared the two on the same system with the same settings.
I'm afraid you're not going to get a lab setting. I have a condo. I waited until I had the same ambient room temperature, then I put the TRUE on the same board and CPU that the loop was on, with the same TIM and used the exact same BIOS settings with fans of similar db, cfm and pressure.
That's it. Simple, but controlled and effective. At the end of the day you can get everything you need by watching and listening. I'll leave the data capture and reporting to the experts. This ain't Martin's Liquid Lab I'm afraid.
I provided pictures of the rigs and a screenshot of the Air cooled system under load. Specs of each cooling system are in my sig.
I'm here to tell you that this air cooler is as good as my quiet loop in every way. In fact it's better because it's far more compact and doesn't require near as much power.
If you're as skeptical as I was about air cooling an i7, then the only way to open your eyes is to buy one and compare for yourself. I'm not interested in changing your mind.

You: I read your post carefully u started by saying
That was in response to your post. The rest of my reply was to address the OP's two main concerns(Cost/performance, Noise). Please don't mix the two, because they don't address the same topic.
You said a dual rad is better than a TRUE, I wholeheartedly disagree. I've used a BIXII 240 in a good loop with YL mediums and quite frankly it was unimpressive at best. And that was on a Q6600 GO. It would be completely overwhelmed on this i7 machine, where as my TRUE does a splendid job. That's where I'm coming from on that one.

You:So now am not sure what you saying will a properly configured loop win out or will TRUE ?
I did state it in what I thought was a clear manner.
Me:There are loops that are appreciably superior to an air cooler, that is a certainty.
If you take that quote from context it means that, some loops will be superior from a raw thermal efficiency standpoint, but you'd be hard pressed to do it with a setup rigged for silent running. Or with a dual rad setup for that matter.

I just don't agree with you, and you don't agree with me. But I've used all the gear we're talking about here and I simply have no choice but to hold on to my opinions until my own testing and comparisons prove them wrong.

I'll tell you what. As soon as my new pump, res, fan controller and Denkis arrive, I'll convert my air bench to a H2O/AIR bench and I'll put a fully documented comparison thread together with all of my findings. I'm sure you'll be quite interested.

It should be clear to anyone that I am lacking a fully documented comparison. I did not collect any data from the H2O bench because the need never arose until I had it disassembled and the TRUE installed. I do however have the idle and load temperatures from that rig burned into my mind, and I am constantly reminded of them, because the AIR rig is showing virtually the same numbers.
All that being said, I will stand by my previous comments 100%, because those little yellow numbers in the task bar speak for themselves.

OP, apologies for taking your thread off track.

[BioTuned]

I have to agree with baditude_df.

A properly configured TRUE can perform neck to neck with a water-loop. Swiftech loop , being the case, is a mid-range loop. Low end loops (thermaltake kits, gigabyte kits, etc) can be outperformed by a TRUE. However, in an closed case water-loop would be better than TRUE, assuming the rad is installed outside of the case or where fresh air is easily reached to the rad, and the rad is a 120x2 or better. It ,sometimes, can be hard to provide fresh air to the heatsink. I too have had share of water loops and TRUE experiments of my own.

However, as for your friend, if he is looking to go water, then he should look into high end 120x2 or 120x3. I would recommend a 120x3, if not he should stay with a TRUE, and tweak TRUE, i.e. flatten the surface of it maybe even the cpu, do a washer mod.

[baditude_df]

And make damned sure the airflow within the case and the fan choice are optimized, or he's doomed to failure with AIR. These are both factors which need to be addressed right away if they haven't already. If not, tell your friend that proper airflow and a good high pressure/high cfm fan will noticably reduce his CPU's core temperatures, and show him the difference he's looking for in order to venture towards 4GHz or above.

And yes, my results will differ slightly since both rad and heatsink are in open air, whereas in an enclosure, the heatsink would be interior ambient and rad would likely be exterior ambient. That's a given, assuming you don't mount your rads internally.

[EarthDog]

No way would that true be able to compete with my loop. I had the Tuniq120 in a push/pull and my loop spanks it. The TRUE was only 1-3C better than Tuniq, and temps at 4.698Ghz are barely breaking 60C. The Tuniq couldnt hold a quad past 4Ghz in most ambient situations and keep it under 70C.

[aphellyon]

I'll leave the data capture and reporting to the experts. This ain't Martin's Liquid Lab I'm afraid.

Good point. Has anyone done air vs water in a decently constrained experiment? Or is cobbling together various independent experiments and performing some kind of pseudo-meta analysis whilst leafing through pages of anecdotes all we have? Like the OP, I also have a friend having the same i7 920 true vs wc issue... I thought I had a pretty firm understanding on this, but it seems the wc and air consensuses may not be lining up.

[baditude_df]

Good point. Has anyone done air vs water in a decently constrained experiment? Or is cobbling together various independent experiments and performing some kind of pseudo-meta analysis whilst leafing through pages of anecdotes all we have? Like the OP, I also have a friend having the same i7 920 true vs wc issue... I thought I had a pretty firm understanding on this, but it seems the wc and air consensuses may not be lining up.

My test methodology was simple as outlined above. The results are more than enough for me to make an informed decision, and should be for anyone else. Open air bench envirionment only though.

As for AIR vs H2O testing done within an enclosure? I'm afraid I'm in the same boat as you. Have you searched XS?
One has to assume that a dual or triple rad exchanging heat outside the case will be more efficient than a TRUE inside a case, give that the same fan is used.
Increase case airflow however, and mount a higher pressure/higher cfm fan on the TRUE, and the story will be quite different. I think, to a point, you could get away with putting a better fan on the TRUE as opposed to the Rad due to the fact that the Case will muffle the sounds somewhat.

No way would that true be able to compete with my loop. I had the Tuniq120 in a push/pull and my loop spanks it. The TRUE was only 1-3C better than Tuniq, and temps at 4.698Ghz are barely breaking 60C. The Tuniq couldnt hold a quad past 4Ghz in most ambient situations and keep it under 70C.

As you can see, my loop is not too shabby either. In fact it is comparable to yours in every way.
I'm not speaking for everyone when I post, only for me. I am presenting what I have found on MY i7 bench.
Given all of the findings of all the different cooling solutions I have seen across the web(including OCF) and it becomes obvious that, given the temperatures I am seeing, both of my configurations are configured and operating optimally.
Lord knows what was going on with your setup. But I will agree with you that on 775 with the 9550, the TRUE couldn't touch my loop in terms of heat dissipation efficeincy. On this i7 setup however, WOW is all I can say. But TBH I've been seeing alot of this lately, especially from members here. Absolutely insane overclocks and temperatures on their 920's with AIR(TRUE usually).

I'll be reconfiguring this loop and adding it to the same bench as the TRUE as soon as Petra's can get my new parts to me. I'll really document everything from start to finish and see if we can't come to some conclusions as to why AIR seems to be so comparable to H2O on some i7 rigs.

[aphellyon]

Yea, I've been crawling XS for a little while now... haven't come across any decent water vs. air tests yet.... just various water vs. water and air vs. air ones. I'm leaning towards advising my friend to get a true, just so I could do a little experiment myself... is that wrong? :D

[baditude_df]

Lol, no. If he figures at the end that he doesn't like it you can always sell it to me for a discount. ;)

[EarthDog]

I assume with your testing you are using HT?

[BioTuned]

Yea, I've been crawling XS for a little while now... haven't come across any decent water vs. air tests yet.... just various water vs. water and air vs. air ones. I'm leaning towards advising my friend to get a true, just so I could do a little experiment myself... is that wrong? :D

That would be a good start, indeed.

@ GMdoubleG
just out of curiosity, which one of the Lian Li cases does your friend have?

[GMdoubleG]

My friend has the Lian Li PC65. I have no clue how good the air flow is inside of the case. I do know that he only has one fan on the TRUE so adding another one should help him a bit.

[baditude_df]

I assume with your testing you are using HT?

Always. I never turn it off.

[ascl]

No way would that true be able to compete with my loop. I had the Tuniq120 in a push/pull and my loop spanks it. The TRUE was only 1-3C better than Tuniq, and temps at 4.698Ghz are barely breaking 60C. The Tuniq couldnt hold a quad past 4Ghz in most ambient situations and keep it under 70C.

I have experienced something very similar. 20 degree improvements in my CPU temp going from Tuniq Tower to an MCR220... but this is on a Q6600, and in a case (Cosmos 1000) that has pretty poor air flow.

For me, even if TRUE vs Water was equal in terms of cooling, if water was quieter it would still be worth choosing. With my setup, there is no way a TRUE would win however (not that I am saying baditude is wrong, I have no experience with i7's, and he has no case to hinder airflow).

[baditude_df]

I experienced nothing near 20deg improvements going from a TRUE to a dedicated Dual Rad setup on my Q6600. I went from ~25deg-TJMax to 35Deg-TJMax under small ffts, so ~10deg.
That being said, 10degC reduction = well worth the money IMO.
Both setups: XFX 680i and 1.2625VID Q6600.
http://img174.imageshack.us/img174/9074/img0208kr8.th.jpg (http://img174.imageshack.us/i/img0208kr8.jpg/)http://img241.imageshack.us/img241/9418/img0372ku7bo9.th.jpg (http://img241.imageshack.us/i/img0372ku7bo9.jpg/)
Both using YateLoon 1600RPMs.
These were some of my results (http://www.ocforums.com/showthread.php?t=558492&page=2) after I installed the loop. I'll have to dig around to see if I recorded anything with the Q6600 and the TRUE though.

I know Tuniq towers don't suck that bad. Did you have fans on it? ;)

[CharlieCS]

I just don't agree with you, and you don't agree with me.

Yes.

[ascl]

I know Tuniq towers don't suck that bad. Did you have fans on it? ;)

Haha, of course. Stock TT fan (I tried a YL highspeed, but the stock TT fan was better). Given the lack of airflow in this case, I can easily see that causing a 10c difference between our results. TT got warmed case air, rad gets fresh air.

Anyway, just a data point, not really useful since the discussion is about i7's... but I am happy with the results (and the quiet!) :)

(And can't wait to hook up this 140x3 I have sitting here).

[ChanceCoats123]

I think I may have a point that water should be able to pull away. Due to the amount of water in the loop, basically any loop should have a decently large amount of space for heat energy from the cpu to move to, now with decent air flow, a 320 or 220 should be able to keep up with an i7. But what I see: Since there is a much larger area for heat energy in a loop compared to a TRUE, as you increase oc's and temps start to raise and the TDP increases and more and more heat it pumped out, the TRUE will reach its limits relatively quickly due to the smaller area for heat energy where as the loop, having more space will (should) still be able to handle the extra heat relatively well.


"Real" type example, moving from an oc' of 4ghz the TRUE and the loop both sit at lets say 60* under load. Now you move the oc' to 4.5ghz and the volts go up, the TDP increases and more heat is put into both cooling solutions. In this case, the TRUE, having a lower amount of wattage it can cool is now at it's limit with the fans on it and the temps pop up to 90* under load where as the loop can take a larger amount of input watts and still stay close to 80* with the same fans. Obviously temps would get better for either solution with higher CFM or higher pressure fans, but that isn't the point.

[baditude_df]

I think everything you said holds merit.
I can put it to the test and actually document it when I get the loop put back together if you'd like. I'm gonna be testing anyway.

[ChanceCoats123]

Okay sounds good. I mean, obviously since I don't own neither a TRUE nor a loop, I am basing this off basically nothing, but I do think that in some conditions it would be doable.

[nzaneb]

What will the computer be doing? Is he going to be running DC? If so, it makes sense to go water. If not, TRUE all the way.

[baditude_df]

I've been crunching with the TRUE for a few weeks now and it has cooled every bit as effectively as the water loop I had on there for the last few months.

What's the basis for your comment nzaneb? Honest question :)

[EarthDog]

Chance nailed it on the head (as did I with a similar statement earlier in the thread, post #22)

At "normal" o/c levels they should perform similarly with the water still coming out on top a couple of C. But a water loop like we have Bad, has a lot more ability to dissapate heat then the TRUE. Where the TRUE falls off acceptable temps, the water continues with lower temps at higher clocks.

And please trust me when I say a 240 sized rad will be ok for an i7, but I would without a doubt get a 360 for best results.

[baditude_df]

As I said, that theory holds merit. But I don't know that anything has been hit on the head, unless you mean being very similar with what you believe.
I'm simply not seeing results that parallel that in any way. The TRUE is keeping up at every stage, at a 4GHz overclock from idle right up to the max temperature reached during a 20 pass LinX stress test. The temperatures are just the same after a 20 minute test. That is enough time for a loop to equalize. I just don't know what to tell you.

Again, I'm going to rebuild this loop with some better/scalable fans on a controller and put both cooling solutions a good head to head test. I'm currently waiting for my petra's order, but my Denki's arrived last night. :)
I already have one on my TRUE and it will also be thrown on the controller when it gets here.

And please trust me when I say a 240 sized rad will be ok for an i7
Yeah it will likely be sufficient to disipate heat but what's the point? Again we have the price performance factor as compared to air, and the fact that you'd likely need loud fans to have anywhere near the efficiency to justify the cost difference.

but I would without a doubt get a 360 for best results.
That's hitting it on the head IMO.

[ascl]

But is the noise the same bad?

[baditude_df]

The max setting for the antec tricools on the TRUE is definitely bearable, and they are on a bench not 3 feet from my head.

From post 8.
A touch louder than the YL Low speeds that were on the rad. But those were shrouded, which will quiet them down a touch even further. More comparable to a YL Medium(~30dB Manuf. rating)
Most of the time I keep the antecs on medium, which is more of a whisper quiet, but definitely not inaudible.
To me it would not be worth it to run an air cooler with a jet engine fan(like the stock Denki I have on there now), just to say I could keep up to a WC loop. I had to keep the fans somewhat similar, and the Antecs were the closest thing I had lying around. The other alternative was a YL high speed and that's just ridiculous.

[ascl]

So a 2000rpm fan vs a 1350rpm fan? Thats quite a difference! If you do get around to doing this test again, please try and make it as fair as possible. I personally find the TriCools on high much louder than a YL low speed fan (or even a medium speed YL). (Although 3 fans is going to be louder than 2 equal fans).

Those Sanyo Denki fans are amazing....ly loud! Heh, I have 3 of the 3600rpm version, and they are LOUD. I haven't even bothered hooking them up to a rad yet (and they are not all that great at under-volting).

[Spawn-Inc]

18 high speed yates for the win. man i feel as though i repeat myself all the time. cheers once again.

[MongGrel]

I keep playing with the idea of W/C, and Bad has outlined most of the reasons I guess I haven't.

I like not having a freakishly loud system and have been trying to quiet mine down and just keep tweaking my TRUE I guess lately, have a few more things I want to try soon.

I leave this one 24/7, leaks make me nervous when I'm not around possibly, and the TRUE seems to work very well as is.

And things are pretty quiet atm as I've changed a few fans around and experimenting with things I guess, watch movies a lot with the computer in the same room as the large screen.

I think I can actually do a few other things soon and push this TRUE a bit more, the thermal paste I'm using at the moment not even very good, I didn't realise I had run out of MX-2 and had to turn things down.

Have some MX-3 on the way to try out.

Have been trying to keep my fans around 24 dba tops at least.

:beer:

[Spawn-Inc]

I keep playing with the idea of W/C, and Bad has outlined most of the reasons I guess I haven't.

I like not having a freakishly loud system and have been trying to quiet mine down and just keep tweaking my TRUE I guess lately, have a few more things I want to try soon.

I leave this one 24/7, leaks make me nervous when I'm not around possibly, and the TRUE seems to work very well as is.

And things are pretty quiet atm as I've changed a few fans around and experimenting with things I guess, watch movies a lot with the computer in the same room as the large screen.

I think I can actually do a few other things soon and push this TRUE a lot more, the thermal paste I'm using at the moment not even very good, I didn't realise I had run out of MX-2 and had to turn things down.

Have some MX-3 on the way to try out.

Have been trying to keep my fans around 24 dba tops at least.

:beer:

for me it's the pump dying not leaks that gets me worried. but i have left my pc running for 2 days once.

[ChanceCoats123]

^^^ Woah woah woah. Are you not allowed to have a 24/7 pc under water? If so this might be an issue if I go water before my new board (probably won't happen but #$%& happens.)

[hokiealumnus]

Sure you can have a PC under water for 24/7 use. Mine is folding; CPU and GPU. BIOS monitored thermal protection will kill the system if it reaches 90C, so I'm not worried. This CPU has been running water cooled 24/7 since February. The GPU almost as long (except for a stint on air when I was waiting to install my new block but after I took the old one out of the loop), with the NB now having a few months under its belt. If I couldn't fold while water cooling, I'd never have gone to water.

[baditude_df]

So a 2000rpm fan vs a 1350rpm fan? Thats quite a difference! If you do get around to doing this test again, please try and make it as fair as possible.
I disclosed that in my first post, this is no secret. You missed the point of the comparison. The point was not to use the same speed fans, it was to see if I could get similar cooling to my expensive loop with a noise level that was in the ballpark(infield). RPM did not enter into the test. Pressure and cfm maybe, there are fans that do both well with fairly low noise.
I sat at the my desk and had a close listen. It was just on the low end of bearable just like the 3 YL lows on the Rad were, and there you have it.
If you could spend $60 on a TRUE and get performance and noise level that rivals water, would you care what RPM/Pressure/CFM the fans were? No.
As it happens though, I will be using the exact same fans in my upcoming comparison just to be "fair".
But I'm going to tell you right now, the goal of all this work is to come out with a combination of fan and TRUE that comes as close as possible to the WC at low noise (ie: the noise people would expect from a high end, near silent loop). Otherwise its a complete waste of time.

I personally find the TriCools on high much louder than a YL low speed fan (or even a medium speed YL). (Although 3 fans is going to be louder than 2 equal fans).
I find no audible difference between YL mediums and Tricools on high. You must have very sensitive ears. And even at low setting in push pull I was getting similar results. Maybe 5 degrees higher which meant that I was 5-10 degrees away from the WC. Pretty damn close.

Those Sanyo Denki fans are amazing....ly loud! Heh, I have 3 of the 3600rpm version, and they are LOUD. I haven't even bothered hooking them up to a rad yet (and they are not all that great at under-volting).


^^^ Woah woah woah. Are you not allowed to have a 24/7 pc under water? If so this might be an issue if I go water before my new board (probably won't happen but #$%& happens.)
You most definitely CAN have a 24/7 WC'd PC. I've crunched months worth of Rosetta straight under water.
You know when people tell you not to chince out on the pump??? That's why. Get yourself a nice D5 or DDC, and you can just let it go without worrying about it.

[ascl]

You are right I did miss that, sorry. And yes, I have spent far too much time reading SPCR forums, and hence am very sensitive to noise! I find my D5 noisy... even fully suspended! People's definitions of 'quiet' vary widely -- of course with the a/c on I can't hear jack anyway :(

Anyway, I am not trying to discredit your statement at all, I am just trying to figure out the variables. If the TRUE works better for you (were better is a combination of quieter/cooler/cheaper/less hassle) then that is great. 5-10c difference with similar noise levels is pretty damn good!

[baditude_df]

Yeah, 5-10 is a range that I recall, it may be slightly more or it may fall right within. I just didn't make a screencap of the lowspeed runs. What I do remember is that I never once lost that feeling of wow, I can't believe how well this is working.
If you find your pump noisy, uhhhh wow. Your ears are sensitive! Do you have to wear earplugs when you sleep? :p
I always put mine right next to the rad, which is probably why I can never hear it. Except when I'm bleeding air of course.

[ascl]

I couldn't sleep with my computer in the same room! Fortunately I have a study to keep it in :)

EDIT: And I have my vario turned down to 3 or 4 to try and cut the noise a bit...

[MongGrel]

Frig, I just like air to be safe I guess, went on a cruise a year and a half ago I guess and had a Q6600 running at 3.6 then and came back a week later and was still running solid like when I left. I just left it that way then because it had been anyways.

Sure if you get a leak and it runs out and things get too hot the comp will shut down and then all the liquid would be laying all over for a week :)

I'd rather not I guess.

At least at this time, Ive plenty of aquariums I've had set up over the years and used to have a whole basement set up with probably 20 fish tanks 30 years ago, ranging from 4X55 gallons, several 40's, god knows how many with Killie fish and various specialty tanks, pretty much had my own mini sea-world going att. Were many pumps etc involved to say the least. Actually, a lot of it was on one big PVC rigged main pump system, and even had a few tanks doing separate filtering tasks.

I could probably run a water loop in here NP in my sleep, I still don't really feel comfortable with one :)

I guess was into fish then heh.

So I'm a bit familiar with water set-ups, I just still don't really trust it on a PC I guess.

Just me :)

:beer:

[Dapman02]

Frig, I just like air to be safe I guess, went on a cruise a year and a half ago I guess and had a Q6600 running at 3.6 then and came back a week later and was still running solid like when I left. I just left it that way then because it had been anyways.

Sure if you get a leak and it runs out and things get too hot the comp will shut down and then all the liquid would be laying all over for a week :)

I'd rather not I guess.

At least at this time, Ive plenty of aquariums I've had set up over the years and used to have a whole basement set up with probably 20 fish tanks 30 years ago, ranging from 4X55 gallons, several 40's, god knows how many with Killie fish and various specialty tanks, pretty much had my own mini sea-world going att. Were many pumps etc involved to say the least. I could probably run a water loop in here NP in my sleep, I still don't really feel comfortable with one :)

I guess was into fish then heh.

So I'm a bit familiar with water set-ups, I just still don't really trust it on a PC I guess.

Just me :)

:beer:
If done right, a water loop will NEVER leak. I don't get why people think that these things just magically leak at the most inopportune time. The main reason why I wouldn't go water (I have anyway) is the fact that the complexity increases the chance of breaking down. I would be more worried about the pump going out than leaks.

[baditude_df]

Yes, that fear gradually faded away after I saw my loop could operate for a month or two without so much as a drop outside the loop. And I use zipties to fix connections. :) There was always the fear that the pump might fail too, but that also just sort of went away.

[fsarfino]

I used to have a WC setup when I had my 2500 barton

I had blocks on evrything cpu, NB and on the video card. While I liked having the setup and it kept everything cool. Overall IMO it was a pain, from changing out the water every 3 months, bleeding the system and having 2 pumps die on me. Never had any leaks though :)

Overall the worst part of it was having the pumps die and either waiting for the replacement or searching for one around town.

If I was to do it again it again I would need to have a 2nd comp to do it on since I hated the failure rate.

[Randyman...]

KISS (Keep It Simple, Silly ;) ). For me, the complexity, reliability, and cost of water just doesn't justify the performance gains I *might* see (I think I'm limited by my CPU and the risk of killing it by overvolting it more than I am by cooling with a TRUE on a 9550 @ 3.98GHz). So, I might be able to get 4.0GHz - 4.2GHz on water. Big whoop, and hardly worth the cost, time, and reliability concerns ;). My Desktop runs 24/7, and I NEVER have to worry...

Never even had a hankering to try Water. I'd rather lap a TRUE, slap it on, and be done with it ;)

:cool:

[jr1]

+1 to the tought of not worrying about leaks or pump dying anymore
My pc have been on 24/7 for 3 - 4 months stright now exept for 3 hrs here the other day when we had some thunder
Never had one issue with any leaks or pumpfailure with the setup i have now
I had one leak long time ago with a tt block that cracked on me but that's all

[IcedOutCpu]

Based on the initial discussion about TRUE. I just would like to add that I'm using one right now and oc'ed @ 4.43ghz HT-on on my 920 and currently using only 1 fan on my TRUE but I have not seen my load temp more than 75c on all cores.. I'll soon setup another fan for push and pull..

[baditude_df]

Do you have some screenshots to show us? Can you share your BIOS settings, fan type etc. Thx.
I'm not sure that I saw much improvement with the pull fan, but it does seem to help pull the air in a more linear fashion throught the fins. If you put your hand beside a TRUE it seems like alot comes out the sides(not that it really matters too much I think). This is especially noticable when I put the San Ace on there. Wow! Hardly anything out the back, but most out the sides, above and below. During a LinX run, the base of the heatsink and all of the heatpipes feel cold to the touch!!! What the hell's with that? It kind of begs the question, if the CPU cores are supposedly at 70-80 deg then why the heck am I not feeing it in that chunk of copper right on top of it?
I see that you have a TRUE black like me, and I'm beginning to think that it may(and that's a big may) have something to do with it. Maybe a slight process change when they put out the series, or something. I know they're virtually the same cooler as the silver model, but I don't remember performance like this on my 775 systems with an original TRUE.

[EarthDog]

Based on the initial discussion about TRUE. I just would like to add that I'm using one right now and oc'ed @ 4.43ghz HT-on on my 920 and currently using only 1 fan on my TRUE but I have not seen my load temp more than 75c on all cores.. I'll soon setup another fan for push and pull..Define load please and post your voltages and ambients. Those temps are INCREDIBLE for a TRUE with that chip and HT on. Never have I seen that. :eek:

[ChanceCoats123]

Bahaha, the loading is probably something like opening IE 8, or the ambients are around 50*F. :shrug:

[BioTuned]

My friend has the Lian Li PC65. I have no clue how good the air flow is inside of the case. I do know that he only has one fan on the TRUE so adding another one should help him a bit.

is it the 80mm version or the 120mm version. The older one is the 80mm version, which has fan placement in the front, the back and the top, while the 120mm version has a fan placement in the front and the back. The 120mm version is PC-60 but with a window. On either one, you can mod it to fit a 120x3 on the top assuming you are using a regular size PSU and not the longer one, and you won't be using the top two three 5.25" bay slots.

[ascl]

Define load please and post your voltages and ambients. Those temps are INCREDIBLE for a TRUE with that chip and HT on. Never have I seen that. :eek:

And ambient temps! He is probably sitting in a freezer! :D

[freeagent]

With my old Q6600, I was able to run it at 3700mhz, 1.5375, and it would load at 60-65c with 18c ambients, loaded with IBT. With this cpu I can load it with IBT and 30c ambients, and get away with 65c loads. Bluescreens come for me at around 75c :(

I have another Q6600 on the way to me, it has a slightly lower vid then my old one :thup:

Ill be looking for 4ghz again with it, but more then likely 3.7-3.8 is all id be able to get, due to the limitations of air cooling in general, heh. I dont care about voltage, but I do care about temps :D

This took 1.65v I believe :soda:

Maybe 1.7, the most Ive givein it was 1.75v :screwy:

http://valid.canardpc.com/show_oc.php?id=598279

[ChanceCoats123]

^^^:whistle:! I like the Extra Spicey.

[freeagent]

Thanks! :D

[rge]

I have never seen a true with normal ambients run i7 at 4.5ghz with 1.4 vcore with decent temps, or even 4.4ghz. But I would love to see one, if someone is willing to post a screenshot of one. Would get one for the wife if I could overclock hers above her current 4 with scythe.

Ambient in my run below is 23.5C, but you can tell that from looking at average of minimum, ie, idle realtemp core temps. At 4.5ghz and 1.4vcore there will be about 13-15C diff between ambient and idle core temps.

I have gtx 295 and combined loops with 2x360 rads, but basically get about same when running single 360 rad with 6 scythe GT at 1500 rpms in push/pull.

Just ran prime load for ten minutes so can see temps, water reaches a steady state on mine after ~8 minutes. Temps in low 70's max.

http://img377.imageshack.us/img377/4536/4500temps.jpg

[baditude_df]

Is this the RGE from over at XS that I've seen in uncle webb's Realtemp thread?

If so, welcome. :welcome:

As for a screen shot of a TRUE providing excellent cooling for a 4GHz(+) i7 @ normal ambients and low Vcore, look no further than my first post in this thread.

This thing's phenomenal.

[rge]

Thanks for welcome, yep same on xtreme.

But I would need your core i7 as well, that is a nice chip, do they come in a package:bday:

My old i940, which wife has, needs 1.38vcore for 4.2ghz, and with scythe it was throttling at 100C, so she is at 3.8 with 1.2vcore and reasonable temps. Could use stronger fans, but she is not into noise.

[MongGrel]

Based on the initial discussion about TRUE. I just would like to add that I'm using one right now and oc'ed @ 4.43ghz HT-on on my 920 and currently using only 1 fan on my TRUE but I have not seen my load temp more than 75c on all cores.. I'll soon setup another fan for push and pull..

I seriously find this hard to believe at all without something to back it up myself :)

Much as I like my TRUE.

Unless thats just to get it to boot, heh.

[baditude_df]

Based on my observations with this cooler so far, I have confidence he could cool that chip at those settings without hitting TJMax. However, he's seeing the same results as I am at 4GHz, only with higher Vcore and much higher frequency. That part, I am also finding that a bit hard to swallow. And nothing was said about noise level. I would imagine it's not too pleasant.
The part I would like to know even more is how he manages to get such a high Bclock as to allow a stable 4.4GHz overclock out of a 920 with HT enabled.

[IcedOutCpu]

Bahaha, the loading is probably something like opening IE 8, or the ambients are around 50*F. :shrug:

Lol. When I say load that's when I'm running Prime95 blend tests..:screwy:

[IcedOutCpu]

Do you have some screenshots to show us? Can you share your BIOS settings, fan type etc. Thx.
I'm not sure that I saw much improvement with the pull fan, but it does seem to help pull the air in a more linear fashion throught the fins. If you put your hand beside a TRUE it seems like alot comes out the sides(not that it really matters too much I think). This is especially noticable when I put the San Ace on there. Wow! Hardly anything out the back, but most out the sides, above and below. During a LinX run, the base of the heatsink and all of the heatpipes feel cold to the touch!!! What the hell's with that? It kind of begs the question, if the CPU cores are supposedly at 70-80 deg then why the heck am I not feeing it in that chunk of copper right on top of it?
I see that you have a TRUE black like me, and I'm beginning to think that it may(and that's a big may) have something to do with it. Maybe a slight process change when they put out the series, or something. I know they're virtually the same cooler as the silver model, but I don't remember performance like this on my 775 systems with an original TRUE.

here's link for some screens shots..http://www.ocforums.com/showthread.php?t=616689

[baditude_df]

Lol. When I say load that's when I'm running Prime95 blend tests..:screwy:

Use the same as everyone else. P95 Small ffts or LinX for CPU stressing. Prime95 Blend put significantly less stress on the CPU.

But still, the cooler is clearly working for you. Again though, what is your ambient temperature while running those tests?

[EarthDog]

It sure is. Ambient must be decently low (20-22C) to hit 75C with HT on at that speed and voltage. Its almost like Tjmax is off or something.

[baditude_df]

20-22degC is room temperature. That's about as normal as you can get. Maybe not during summer, but hardly low. If ambient temperatures allow(33degC in my condo yesterday :(), I'll post some shots tonight of my new 950 at 4.5Ghz, under the TRUE and a 20 Pass LinX test.

[IcedOutCpu]

Use the same as everyone else. P95 Small ffts or LinX for CPU stressing. Prime95 Blend put significantly less stress on the CPU.

But still, the cooler is clearly working for you. Again though, what is your ambient temperature while running those tests?

My ambient temp since the my pc is on the basement is 18c - 19c ..

[baditude_df]

Whoa! That's low.
That definitely helps alot.
I was sitting at 28degC last night and my idle reference temperatures were sitting between 41 and 47degC. If I could knock 10deg off the ambient, I'll bet any money I could keep sub 70's under LinX with this overclock.

[IcedOutCpu]

I ran Linx for a sample tests of 10000 and able to finish the tests w/ out errors.. What does everybody is set to when running Linx?

[EarthDog]

My ambient temp since the my pc is on the basement is 18c - 19c ..I guessed the abmient temp wrong, but was correct in the idea. That is a cool area the PC is in!

[baditude_df]

Yeah, anyone that tests their core temperatures should try to do it on a day when they have ~normal room temperature of 22DegC or 72DegF to keep things halfass standardized. If you are seeing inordinatly low temperatures you should first and foremost list the ambient, becuase more often than not, that is what is making the difference between your high end cooling results and someone elses.
In this case, not only does he have a good cooler, but the significant benefit of very low ambients.
Whatever.

[EarthDog]

Whatever.:thup:

[rge]

I ran Linx for a sample tests of 10000 and able to finish the tests w/ out errors.. What does everybody is set to when running Linx?

Can you post a pic running linx or prime with temps. The pictures you link have prime without temps, and temps without prime, but not together.

Just curious as to your temps at 4.43 and 4.5 running prime/linx. I get your ambient is 18-19C, but have not seen any load temps except spi which only loads 1 core.

Edit: For example my load temps only increase 10C over idle running spi 32m, running prime they increase 30C over idle, and my C/W is significantly higher than yours. Running spi32m your load temps increased about 15C to 55-60C, but have not seen prime or linx temps yet but I could guess at them.

Nevertheless nice O/C for air, but temps dont make sense even with low ambients unless your cpu is a rarity in terms of leakage, but your spi32m run would suggest otherwise.

[astrodanco]

People who don't post their ambient temperatures when they post their core temperatures should be ignored. Without knowing the ambient temperatures, their reported core temperatures are worthless information.

Personally I also think reporting a temperature for anything less than 100% load on all processors, after less than a full hour of said loading, with HT off, etc. is also worthless, but that's just me. Reports from people with freak chips that can do extreme overclocks at low voltages are also worthless to anyone except other people with freak chips.

[baditude_df]

Love the avatar. Is that a capture from The Prisoner?

[aphellyon]

Love the avatar. Is that a capture from The Prisoner?

Monty Python... I think?

Reports from people with freak chips that can do extreme overclocks at low voltages are also worthless to anyone except other people with freak chips.

It is nice to know the upper limits though... but you're right, outliers should probably be culled from any decent data set.

[freeagent]

That i7 950 @ 4508mhz above actually runs cooler then this super easy clocking Q6600. This cpu is too much for this cooler in 23c ambients running IBT. It takes an amaseing 1.3375v to run this cpu at 3600mhz. It idles hot, and loads well into the 80s, and promptly reboots at 85c lol.

This cpu has no buisness being that hot, for such a low voltage! The IHS is dam near perfect too! Contact is great as well. The only other cpu that Ive had for s775 that ran this hot, was an original E6300, the first gen with the high power smd's heheh.

At least I was able to run the 6300 at 3700mhz :D

The cool thing about this cpu is that it runs at 3400mhz with 1.2875v, wich is the vid of my first Q6600. I needed 1.3875 to run 3.4 with it, and 1.48 for 3.6, and 1.58 for 3.8 :thup:

This is the same setup Ive been running for the last 3 Cpus Ive owned, wich has been a Q6600, E8600, and another Q6600. Nothing has changed except the ram, and vid cards.

http://img512.imageshack.us/img512/1138/img2349s.jpg (http://img512.imageshack.us/i/img2349s.jpg/)
http://img44.imageshack.us/img44/7977/img2349l.jpg

This is the first cpu that Ive had that Im actually scared to apply voltage to! I havent givein it more then 1.36v :attn:

[baditude_df]

Here's some initial AIR testing with this new i7 950. Ambient temperature 20.5degC.

LinX 11:17mins in:
http://img142.imageshack.us/img142/2417/capture11.png
Setup and Thermometer, taken a few minutes later:
http://img142.imageshack.us/img142/6421/imgp0608h.jpg

Overall I would say it's running slightly cooler than the 920 D0, due mostly I'd imagine to the fact that it requires significantly less voltage to run a 4GHz overclock(1.168/1.200v as opposed to 1.264/1.296v for the 920 D0). Air testing with the 920 D0 at 20deg ambient showed 70-73 degC across the cores in Everest.

Parts from Petra's should be here tomorrow so hopefully we can get some head to head water/TRUE action going. Stay tuned.

[rge]

Parts from Petra's should be here tomorrow so hopefully we can get some head to head water/TRUE action going. Stay tuned.

That would be interesting to see. If you can show few different mhz/volts that is most interesting part.

High end air keeps up with high end water pretty well, until certain TDP (vcore/mhz), then air trails off once cooling capacity is exceeded.

May only see 5C difference at 1.2vcore and 15+C difference at 1.35-1.4 vcore. My old i940 temps at 3.6ghz, just 5C difference on load and idle same. But running at 1.4 vcore 4.3ghz water loaded in mid to high 70's, scythe approaches 100C, little over 20C difference.

Be interesting to see True vs water at couple different settings.

[baditude_df]

I'll be ranging from 4GHz to 4.6GHz in 200 MHz increments. I already know the TRUE can keep up at low voltage, that I have seen. My curiosity, as well, lies in the higher voltage range, to the point where the TRUE is just overwhelmed. TBH, I'm confident that it will hold up all the way to 4.6GHz, but we'll see. Depends on how much voltage I have to feed this thing.

If there's room left on the Bclock, I may put a 4th set of results in on the top end.

[astrodanco]

I'll be ranging from 4GHz to 4.6GHz in 200 MHz increments. I already know the TRUE can keep up at low voltage, that I have seen. My curiosity, as well, lies in the higher voltage range, to the point where the TRUE is just overwhelmed. TBH, I'm confident that it will hold up all the way to 4.6GHz, but we'll see. Depends on how much voltage I have to feed this thing.

If there's room left on the Bclock, I may put a 4th set of results in on the top end.
If at each step up in GHz you can run OCCT in cpu:linpack mode for four hours straight without any trouble, that will be interesting to see.

Whatever you do, please be sure to record ambient air temperatures throughout your testing.

Personally I'm not impressed with tests done at low (less than 23C) ambient air temperatures, but whatever, that's just me.

[baditude_df]

Well, this is the best I can do without a lab. Testing done at Room Temperature(22degC), as I live in Canada. 4Ghz results up. http://www.ocforums.com/showthread.php?t=618279