Are my watercooling expections to high?

[amora]

So yeah, I understand the watercooling thing is pure hobby and doesn't necessarily mean super duper high overclocks in comparison to high end air cooling these days.

With that said, depending who you are and the reason you partake in WCing(very low noise to cooling performance, super low temps miligated by ambient and with ultra high speed fans) I'm starting to get the impression that per dollar spent vs gain in performance(of the acctual WC loop, NOT OC potential) is rather fruitless.

We'll start with the WC loop specs:

Swiftech Apogee HD
Swiftech MCP35X
1/2" ID tubes

...as for radiators(and this is where my gripe is)...
- 2x legacy 120mm HWlabs Black Ice Pro's
and
-1x XSPC RX120

Reading reviews and such, per 120mm general area of rad space gains equates to about 5c drop from a single 120mm rad to a dual 120mm rad

-and-

a 2-3c drop again from a dual 120.. to a triple 120mm


Story:
I started with a single BlackIce pro(very old that was reactivated from being in my closet since the days of OCing P4 1.6A northwoods), then added a brand new BlackIce pro and just yesterday added a XSPC RX120.

Now one would expect a very distinguished drop in temps right? Proly in total about 8c drop(low balling) from a single 120mm to triple 120mm rads(with the XSPC rad being double the thickness of my BlackIce pro's).

I have documented several tests of each addition of 1-2-3 rads and I show little to know improvement in temps either at 1200rpm P/P fans or 1700rpm p/p fans + a 38mm 80cfm single fan on full blast on the XSPC RX120.

Pretty much each addition of rad equated out to be only about a 1c drop in temps. This has been a LONGGGGGGGG process as well. It started with me thinking my old *** BlackIce Pro was all gunked up and needed a higher power pump...So I brought the MCP35X. Yeah well that didn't do anything...and after reading and reading and reading sites all over the place about about flowrates and pressure drops and all sort of nonsense....flowrates dont' mean crap!

In practice for ****s and giggles i turned my pump down to 10% for a whole day burn of Prime95 and my temps only rose 2c over my best temps. and at 100% my temps rose 1c over my best temps. My best temps being achieved at a pumps speed of 40%.

Moving along, Ok so the pump didn't help anything, lets see if adding a rad helps! Hey, look at that a single BlackIce pro is only $30, maybe that'll knock 2-3c off my temps. NOPE temps stayed just about the same maybe .5c lower. I also tested this dual setup with several different pump speeds and fan speeds...barely anything.

Moving along again...at this point I'm just thinking the BlackIce pros are super thin and outdated legacy, first generation, watercooling rads(which they very much are, they came out circa 2003ish) so i was like OK! Lets buy a awesome new big old super thick XSPC RX120, this thing should do wonders!!!!!!! NOPE added it to the loop to create 3 rads in series...I only achieved a 1c drop!!!!!!! WTFFFFFFFFFF!!!!

So all in all at MOST from 1-3 rads i dropped 3c...Holy mother F*ck what a waste of money.

Now I'm thinking, OK something is SERIOUSLY EFFED. I go back to thinking, maybe it's that old *** Blackice pro from back in 2004 thats absolutely destroying my loop.

NOPE!

Last night I pulled it out of my loop and ran dual rads(brand new Black ice Pro + XSPC RX120)... This setup was NO better than my old dual setup with 2x black ice pros...in fact I'm pretty sure it actually hurt my temps.

As it stands right now I have a complete brand new EVERYTHING loop with results BARELY different than my initial single rad loop.

In case you were wondering how I'm able confirm this. This tests were standardized:

======================
OCCT 1h
small data set
High priority

tests were completed using fans at 100%(~1700) and at 1200rpm
- they are 2x GT-AP15's p/p plus 2x Zalman ZM-F3(used only bc ap-15's are never friggin stocked anywhere) and a single 38mm sunon @ around 80+cfm so dont' even try to blame the fans
78f ambient ~25.5C
======================

I've even gone so far as swapping rads, putting all rads in series, turning on/of fans on certain rads to see what performance impact it would have.

The only think i can think of now is that I could change is to swap all 3 single 120's and buy a single 360rad... I'm not doing this though bc through all my tests and by my count...
Pump $85
Block $75
BlcIce $30
XSPC Rad $55
Res ~$30
Tubing $30
Fans $ $40

.... $345 with of crap with little to NO results

Please don't' mention mounts, the Swiftech hardware is stupid proof You just tighten the damn things down and go, there is literally no way to screw up that mounting.

I'll always be a fan or watercooling bc it's extremely fun and addicting to tinker with things...buy man o man I really with this tickering would show some results that would reflect the amount of money that i'v e dumped into this loop!!!!

I would start posting ALLL my output graphs from each of the tests but theres not point as the variation between the tests is 1-2 c, which isn't even enough to account for ambient temp fluctuations.

Very disappointed with my tests....and my biggest let down is that there was little to NO increase in performance between my skinny little blackicepro and my THICK XSPCRX120....I mean come one, I woulda though that thing woulda KILLED the skinng blackice pro, even though it has a lesser fin count. Well that and I noticed the stupid think has made my damn fans louder.

bahhhhh Someone help I'm confuseddddd Why isn't all this hardware showing significant results!!!!!!!

 

[manu2b]

Well, the water in the loop will never go below ambient... And current high end hsf are very efficient. And new chips are rather voltage than temp limited due to die shrink and lower tdp. So yes, water cooling doesn't bring as much extra performance as it used to.

But it is stil a lot of fun!

EDIT: What cpu do you use?

 

[amora]

Well, the water in the loop will never go below ambient... And current high end hsf are very efficient. And new chips are rather voltage than temp limited due to die shrink and lower tdp. So yes, water cooling doesn't bring as much extra performance as it used to.

But it is stil a lot of fun!

EDIT: What cpu do you use?

Core2Duo E5300 @ 4ghz @ 1.36v...yeah yeah yeah, I know, a generation back, but it was cheap and if i fry It I wont' care lol I other words is a playtoy for amusment.

This guy prefers not to buy a $300 chip + a $200 mobo...for hardware that lasts little over a year before it becomes outdated again.

I spend $0 on the CPU(was laying around unused) and $40 on the mobo thats it... Its faster than i need since i dont' "game" anymore on PC, that aspect is handled by my PS3 and Xbox....So yeah, I have no use for the thing except for tinkering aka bench testing my WC setup which I'm still kinda hurting inside about

 

[Conumdrum]

Yea, your heatload is so low to start with, so your DT is pretty low to begin with. We have a DT reference in the sticky up top.

If you are at 5C DT already and you get 15% better performance with more rad, that's less than 1C DT improvement. Your at the point of diminishing returns with your loop, I don't even see WC as needed on that chip with such good air coolers out now.

I'd say the RX120 is all you need for that CPU, trash the old rads....................

 

[Diggrr]

Well, where to start...

Firstly, you've got a 65W processor running inside it's factory VID of 1.365 Volts.
Had you noticed that most radiator tests are at 150 watts (like an OC'd 130 watt processor would do)? With that kind of heatload you most definitely would see the advertised difference in temps.
You're running less than half that, so you should expect to see less than half the temp improvement that it sounds like you were expecting.

Radiators are all about temp differential, and in a non-linear fashion, meaning when you're using it in a room at 78°F, your Delta between room temp and cpu temp will be smaller than if that room were 72°F. Slight yes, but different.

Flowrates will impact temps more than your tests will show because you are using 3 series rads .vs a single triple rad. Each rad you're adding will slow the flow rates due to fittings and tubing restrictions, not to mention changes in direction within the endtanks of the radaitors (ie, you have 3 times the normal restriction due to a radiator).
Flow absolutely does impact temperatures, but you need to keep in mind that you are also adding heat from the pump itself as you crank it up and down. @40% it's 7.2 watts, @ 100% it's 18 watts though the flow is only increasing by about 40% (less for you because you're using a more restrictive system than what the pump curves were determined with).
In short, adding 60% to the PWM does not give you 60% more flow, but it does increase your wattage dump by 10.8 watts. Doesn't sound like much, but that's 1/6th again what your processor is putting out!

With a little more thought given, I'd say you are seeing exactly what you should see. You've got more restriction in a warmer room, with a low wattage processor. It's not watercooling's fault...you'd have the same disappointing results moving from a $40 HSF to and $80 HSF (minus the pump heatdump difference).

Ah, Conundrum chimed in while I was answering the phone...anyway. I watercool my 65 water for silence and cause I'd already had it. It's nice though that it transfers easily to a new(ish) 95 watter easily, and will stay just as silent.

 

[amora]

Yea, your heatload is so low to start with, so your DT is pretty low to begin with. We have a DT reference in the sticky up top.

If you are at 5C DT already and you get 15% better performance with more rad, that's less than 1C DT improvement. Your at the point of diminishing returns with your loop, I don't even see WC as needed on that chip with such good air coolers out now.

I'd say the RX120 is all you need for that CPU, trash the old rads....................

5c DT!? Try 30c delta. At stock vcore and 3.4ghz it does 43c Max, when I bumped to 1,34vcore and 4ghz it jumped to 55c with prime95 small ffts/ occt and bursts to 60c with IBT/linpak.

And trust me when I say ive bashed my head against the walls trying every config loop.

Right now it's running a brand new xspc rx120 and a brand new black ice pro and it his 55c in a 1hr burn using small data sets in In occt.

It's not a flow rate problem, that can be ruled out by turning my pump down to a trickle then comparing it to full blast, and same goes with fans. I have 2 sets of good fans.

Also I'm pretty sure my tubing is leaking out plasticizer clogging my block a bit, but last time I checked, I cleaned it out and even with all the junk out my temps didn't change.

Ps its not the mobo Mia reading temps, this proc has been placed on 3 diff monks now lol

 

[fall]

You are misunderstanding what DT really is...

DT is the difference in the ambient air and the temp of the water in the loop.

the water in the loop is always within a few *C of the ambient air (assuming enough raddage).

You are thinking of the difference between ambient air and the chip temp.

 

[Diggrr]

You just completely ignored what I said about using your PWM as a flowrate judge. Changing via PWM tells you absolutely nothing about your flowrate, it only tells you how your system reacts to a 10+ watt heat dump.
If you want to judge flowrate get a meter and a gate valve. That's the only way to check without skewing the results via fluctuating pump power.
How do you think the reviewers like Martin and Skinee are doing it??

 

[spitter3]

You should check your block, take it off and make sure you are making good contact with the cpu!!!