X1900 with MCW60

[scavengerCO]

Have an X1900XT.. Receive my mostly swiftech setup tomorrow (details in sig). Curious if anyone is using an MCW60 successfully with their x1900xt? Gabe @ swiftech said it'll fit. So I'm wondering about performance as well. I'm going to use the ramsinks with arctic alumina epoxy (non-conductive)..

Any input?

Since I'm using the 655 pump I was kind of thinking I should've gotten a storm, but Gabe defends the Apogee as the better route, so I'm game.

 

[Steeeeve]

the Apogee is highly underrated after they have a few control mishaps. I am seriously thinking about selling my Storm for the better flow Apogee.

Your set up is fine and the mcw60 will work great.

 

[ProHandyman]

the Apogee is highly underrated after they have a few control mishaps. I am seriously thinking about selling my Storm for the better flow Apogee.

Your set up is fine and the mcw60 will work great.

Swiftech will ship you the new spacer plate (rev 2.) that increases the flow and performance; at no or minimal cost. The changes are based on design changes Cathar has made for the "G7".

 

[Budwise]

im using the MCW55 with my X1900 and it works great. Its rare to see it above 39C...

 

[Erasmus354]

Swiftech will ship you the new spacer plate (rev 2.) that increases the flow and performance; at no or minimal cost. The changes are based on design changes Cathar has made for the "G7".

Since when was 53$ no or minimal cost?

 

[DFI Daishi]

Since I'm using the 655 pump I was kind of thinking I should've gotten a storm, but Gabe defends the Apogee as the better route, so I'm game.

of course he defends it........of course cathar defends the storm.......and billA refers to gabe as a lying pig........and it's all a lovely little game of who is biased towards who and what.

that's what third party reviews are for, and systemcooling's review showed something some pretty distinct results.

enter steeeeve to explain how that's only becaus it was a heat die........and site other reviews which show the two performing the same, when tested without proper equipment.

as for what GPU waterblock to use: it's not going to make a measurable difference regardless of what you use.

 

[Demon_of_The_Fall]

Swiftech will ship you the new spacer plate (rev 2.) that increases the flow and performance; at no or minimal cost. The changes are based on design changes Cathar has made for the "G7".

Yeah, it definitely is not no or minimal cost. It's practically the cost of an entire new storm.

 

[scavengerCO]

that's what third party reviews are for, and systemcooling's review showed something some pretty distinct results.

thanks for directing me to that review.. I assume you mean http://www.systemcooling.com/swiftech_apogee-01.html it was very thorough but I'm not sure what you mean by "distinct results". I gathered that they pretty much perform on par with each other, and the updates at the end make it sound like they have addressed any quality control concerns.

Basically my interpretation thus far is that the apogee performs on par with the storm for about half the price.

And as gabe pointed out to me, the MCW60 utilizes the apogee base and has been receiving very positive reviews.

So could someone who is using a swiftech mcw with an x1900 (budwise) share what voltage settings/clock settings you're running at so I could have an idea of what to shoot for reasonably? I'm reluctant to do the xt->xtx flash, I'm normally very comfortable with this, but on a 500 dollar card =) I'm a poor high school computer teacher... heh.

Thanks all.

 

[Yreka]

LMK how it works out. I just ordered the same block for my 1900XT. Should be installing it weekend after next.

 

[s_s256]

the Storm v2 doesn't even perform all that much better over a Storm v1, certainly not $53.

http://www.swiftnets.com/products/storm-r2.asp

 

[DFI Daishi]

thanks for directing me to that review.. I assume you mean http://www.systemcooling.com/swiftech_apogee-01.html it was very thorough but I'm not sure what you mean by "distinct results". I gathered that they pretty much perform on par with each other, and the updates at the end make it sound like they have addressed any quality control concerns.

Basically my interpretation thus far is that the apogee performs on par with the storm for about half the price.

And as gabe pointed out to me, the MCW60 utilizes the apogee base and has been receiving very positive reviews.

no, the apogee doesn't perform on par with the storm based on systemcooling's results. all other reviews that i have found of the block utilize the internal temperature diode of a live proc and are therefore useless for comparison of any two modern CPU blocks.

http://www.systemcooling.com/swiftech_apogee-10.html

comparisons using hydraulic pumping power take into account the apogee's lower flow resistance and therefore higher flow rates with a given pump.

using this measure, or the plot vs flow rate, the storm handily beats out the apogee by better than 3C on an overclocked proc running 100 watts load. the internal temperature diode can't reliably resolve that temperature difference.

reviews using live CPUs CAN give valid results........if they integrate a calibrated temperature sensor into the proc, as per thermal solution validation guidelines. reviews that don't even do that aren't worth the bandwidth that you use up while reading them.

so far as the MCW60 goes.....i have not read a review of it or any other GPU block that does a decent job of it.

 

[scavengerCO]

received mcw60 today.. removed HSF. seems like I should be using some sort of shim or pads.

the radeon kit with the mcw60 is just the single bar, two screw variety. The geforce kit has a cross bar with four holes and looks like it'd actually work better on the x1900.

Anyone who has used the mcw60 w/x1900.. did you use the geforce kit perhaps?

Anxious to get this bad boy under water and eliminate that annoying fan.. not more atitool clock control!

Is the card going to barf without having a fan attached to the header? If not that is kind of concerning.. if the fan fails, does the card just burn itself up?

cheers

 

[Budwise]

the X1900 uses the Nvidia mounting system...

 

[dnottis]

No the X1900 is fine without a fan connected. I have an X1900XTX with a Maze 4 on it and some swifty ram sinks. I'm maxed at CCC overdrive 690 / 800 (stock volts still) and my highest load is 41.06 Deg C (according to Riva Tuner in Q4 Nexus Hub after 10 mins). Most apps don't push it past 38 Deg C though.

 

[scavengerCO]

So i'm finally up and running (pics to come later)..

I used ceramique instead of the AS5 I had, I figured why risk it. Although I did use AS5 on my naked processor.. I'm typing this so it must not have oozed out and fried anything around it =) (knocks on wood floor of kitchen where this whole setup is at the moment =) The AS5 *might* help a degree or so, but running 40 at max overclock is hardly something to sneeze at versus the 60-70 degree crap I was seeing before at stock speeds with 90-100% fan.

Anyway... I used CCC to overclock and it easily jumped to 690/800 max in CCC. Idling around 30 and topping at 40. God bless water. So that DXStress.exe that CCC runs during auto-overclock is interesting..

Ok, so I'm totally happy with these settings, but I gotta ask.. does atitool really serve a benefit to me in this situation? I was previously using it to keep my stupid-arse whiny fan at 100%, but now.. I guess if I wanted to raise voltage. But then don't I have to turn off the 2d/3d switcher to overclock in atitool. I like that the CCC route allows me to preserve that functionality.

By the same token, any benefit to flashing to xtx? What does CCC top out at for an xtx? As I understand it (and part guessing here), flashing to xtx would raise the "stock" voltage, thus allowing me to clock higher, true?

And avoiding the flash would in theory keep my warranty intact (I know I can flash back). Nevermind the fact that I epoxied ramsinks onto my card But they wouldn't notice that surely.. laugh.

But I mean 690/800. I can't bitch. And in truth, running higher voltages and going higher on clocks than that would make me nervous about the life of my card. After spending $519 on this little biatch I want it to last atleast a few years =)

Thoughts/feedback? Anyone else using the MCW60 with the x1900xt?

What are those of you with boxed xtx's seeing on water?

 

[Steeeeve]

no, the apogee doesn't perform on par with the storm based on systemcooling's results. all other reviews that i have found of the block utilize the internal temperature diode of a live proc and are therefore useless for comparison of any two modern CPU blocks.

http://www.systemcooling.com/swiftech_apogee-10.html

comparisons using hydraulic pumping power take into account the apogee's lower flow resistance and therefore higher flow rates with a given pump.

using this measure, or the plot vs flow rate, the storm handily beats out the apogee by better than 3C on an overclocked proc running 100 watts load. the internal temperature diode can't reliably resolve that temperature difference.

reviews using live CPUs CAN give valid results........if they integrate a calibrated temperature sensor into the proc, as per thermal solution validation guidelines. reviews that don't even do that aren't worth the bandwidth that you use up while reading them.

so far as the MCW60 goes.....i have not read a review of it or any other GPU block that does a decent job of it.

I just disagree. The storm is just as good as the apogee but with the apogee not restricting flow as badly, it makes it a better choice IMO. In real life the apogee has preformed just as well as the storm.

 

[DFI Daishi]

Ok, so I'm totally happy with these settings, but I gotta ask.. does atitool really serve a benefit to me in this situation? I was previously using it to keep my stupid-arse whiny fan at 100%, but now.. I guess if I wanted to raise voltage. But then don't I have to turn off the 2d/3d switcher to overclock in atitool. I like that the CCC route allows me to preserve that functionality.

uninstall the CCC first and then load up ati tray tools or ati tool. i have things set up with ATT so my card is only OC'd when in 3D mode and i'm pretty sure that i used to do the same thing back when i was still using ati tool. i don't know about the voltage thing though......i didn't think you could tweak that using ati tool.....might give it a try again to have a look for that feature.

i never had good luck OCing in CCC, it was constantly locking up on me at clocks that were hardly above stock and well below what i hit with 3rd party apps. maybe that's just my bad luck, once again.

 

[DFI Daishi]

I just disagree. The storm is just as good as the apogee but with the apogee not restricting flow as badly, it makes it a better choice IMO. In real life the apogee has preformed just as well as the storm.

"in real life" you don't have the slightest clue what is performing better as a function of the uncertainty generated by the poorly calibrated internal diode.

the processor manufacturers aren't reccomending that you use an external thermocouple to validate cooling performance just for the hell of it.

the spread between the best performing blocks on the market and the worst is smaller than the temperature diode can resolve. the internal diode can't reliably tell the difference between a storm and a maze 3, let alone the difference between the storm and the apogee.

oh, but differences that small don't affect overclocking......right.......so why aren't you running a cheap older block yourself?

you don't know for a fact if your present OC is thermally limited or not, so you get a good performing modern block hoping to eliminate that as a possibile limiting factor and take things to the limits of what your hardware will tweak to. you choose to drow the "it cools well enough" line at the apogee. fine. just admit, if only to yourself, that it's an arbitrary choice as opposed to a reasoned one.

by the same token, you would not know for sure if temps were the limiting factor when reading a review using how far the proc will OC as it's measure of cooling performance.......meaning that it's not exactly a good measure of which block is doing the better job in a comparative review.

if the proc needs more/cleaner voltage to go farther than it is presently at, what block generates the best results has nothing to do with how the block is cooling and everything to do with room temps, airflow around the board, line voltage and power handling component conditions. how do you isloate for factors like that? repetition? just how many reps with how much hardware would it take? what happens if someone else trys it and gets different results? who's results are "better" results? just how much unaccounted for variation is there in "real world" tests?

synthetic testing allows you to have a definate answer to a definate question, which you can accept as true to the extent that it is repeatable with controlled and measured conditions. the question answered is not "how well will a given block let me push my hardware" unfortunately. the question answered is what block cools better under the testing conditions used. a lot of people seem to have trouble doing the math to translate those numbers into the conditions that they experience at home. that's fair, it's not easy. you can do it if you sit down and hammer on the charts.

testing using the internal diode of a real proc doesn't answer any question at all, since you can never know how much uncertainty in the results is resulting from what factor.

 

[texuspete00]

It seems there is a lane of power components that used to touch the stock sink. Next to the long secondary sink that is already on some other power components. What do you guys do about it?

 

[centvalny]

You need to use MCW 60's nv6800 back plate (so the card won't bent) and longer screws.....

Heres the run...single card...http://service.futuremark.com/compare?3dm05=1830033

 

[DFI Daishi]

It seems there is a lane of power components that used to touch the stock sink. Next to the long secondary sink that is already on some other power components. What do you guys do about it?

for myself, i have the rad venting out the top of the case and the rear fans and PSU fans bring in air around that area. others watercool the mosfets. still others leave the PSU and rear case fans as normal. it's my impression, based on my own experiences, that the mosfets do fine with normal case airflow.

 

[scavengerCO]

It seems there is a lane of power components that used to touch the stock sink. Next to the long secondary sink that is already on some other power components. What do you guys do about it?

I believe you're referring to the vregs, and I was seeing those get up to 85/90 according to atitool. I pointed a 120mm fan at them and it dropped them considerably, so I'm pursuing mounting a fan in my window to blow down on the card. I have an antec slot blower next to the card but that wasn't helping all that much, although that was with the case open.

Danger Den's x19 block is supposed to cover those, but I'm not waiting for that to come out and my mcw60 rocks.

 

[scavengerCO]

uninstall the CCC first and then load up ati tray tools or ati tool. i have things set up with ATT so my card is only OC'd when in 3D mode and i'm pretty sure that i used to do the same thing back when i was still using ati tool. i don't know about the voltage thing though......i didn't think you could tweak that using ati tool.....might give it a try again to have a look for that feature.

i never had good luck OCing in CCC, it was constantly locking up on me at clocks that were hardly above stock and well below what i hit with 3rd party apps. maybe that's just my bad luck, once again.

I haven't used tray tools yet, I'll have to try it. atitool does let you modify voltages under prefs.

Had a small leak on the outlet of my mcres.. had bought a dangerden fill port and some 1/2" ID at home depot. Tonight I'm pulling out the 7/16" and replacing it with the 1/2" and switching to the t-line/fillport instead of the res. 3 days of water cooling and i've already drained my system twice and i'm already replacing my tubing =) But practice definitely makes perfect.

I like the 1/2" ID better, more rigid, but the natural curve fits the flow better. Also with the thicker wall, theoretically it will be better with the t-line/fill versus thinwall because of slower evaporation.

Doing pump->cpu->gpu->rad->t-line/fill->pump

I'd prefer to go from rad to cpu but that makes for longer runs and less natural curves.

--j