Object lesson in flow rate

[hafa]

While working the other day, I noticed that my CD-RW was not showing as a CD-RW. In fact, Windows showed the drive as a "late-g@n" rather than a "lite on." Upon inspection of the drive, I found the back of it to be wet. NOT a good sign in a watercooling system. Shut down right away and found that the 90-degree elbow on the waterpump output had developed a stress crack. This gave me the incentive to finally make some overdue changes to my system. I moved the pump to the bottom of the case, got rid of the 90-degree elbow and re-routed the tubing.

All I have to say is that this should have been done long ago. Temps are now 6C lower and the system is quieter, since the pump is more securely mounted to the bottom of the case.

As for the CD-RW, only the connector got wet. A happy ending for what could have been an unmitigated disaster.

[GOD Macktruck]

nice improvement.

those 90 degree elbows will really kill a system

: ) :

[Mikeonatrike]

So it took a kick in the cd rw drive to figure out that 90s are crap eh ;)

[Shadowcat]

Nice to hear that things didn't totally mess up for you... for once it seems a leak proved helpful!

[Thornbrand]

Indeed, glad to hear you caught the problem...that could have turned out much worse.

A good lesson learned, and hey...you got a nice reduction in noise and a generous improvement in temps :thup:

[Korndog]

wow... 90s are that bad eh?
well.. what if u have a tube splitting into parallel then have a 90 on each of those parallel tubes, then back to one tube? is there less loss in flow?

[ILikeMy240sx]

I doubt it was the 90 degree elbow that caused the 6C difference in temperature... but cant hurt without it.

[hafa]

Originally posted by ILikeMy240sx
I doubt it was the 90 degree elbow that caused the 6C difference in temperature... but cant hurt without it.

The only things that changed were the removal of the 90, re-positioning the pump to the bottom of the case, and the requisite increase in tubing length to accomodate the new configuration.

Given that the increase in tubing length and lower elevation of the pump are potentially negative changes, that leaves only one thing...the 90.

I'm pretty confident of my temps, as they're dervied from a themocouple on the cpu die side connected to the digital readout on my fan bus.

[ILikeMy240sx]

Originally posted by hafa


The only things that changed were the removal of the 90, re-positioning the pump to the bottom of the case, and the requisite increase in tubing length to accomodate the new configuration.

Given that the increase in tubing length and lower elevation of the pump are potentially negative changes, that leaves only one thing...the 90.

I'm pretty confident of my temps, as they're dervied from a themocouple on the cpu die side connected to the digital readout on my fan bus.

I thought you removed your block and installed it again... that could have a significant affect but I guess not in this case....

The position of the pump doesn't really matter - pump position has nothing to do with performance. I doubt you increased your tube length significantly to have a measurable difference... That only leaves 90 def bend but seriously I highly doubt that one single 90 degree bend will drop your CPU temp by 6C... there has to be something else. Besides, Maze4 is not a direct-center cooling block which means that even with increase in flow, temperature difference between before and after will not be much.
ie - Swifty 5002 - since its not a center cooling block, even with increase in flow there is not a significant difference in c/w to cause a significant difference in change.

[I.M.O.G.]

Originally posted by Korndog
wow... 90s are that bad eh?
well.. what if u have a tube splitting into parallel then have a 90 on each of those parallel tubes, then back to one tube? is there less loss in flow?

Much worse. Simpler is better.

[Korndog]

Originally posted by IMOG


Much worse. Simpler is better.

eep! good thing i'm getting a res soon, won't have to deal with elbows anymore.
i don't get how it can be worse though, but i've barly started fluid dynamics in physics so thats probably why ;)

[hafa]

Originally posted by ILikeMy240sx


I thought you removed your block and installed it again... that could have a significant affect but I guess not in this case....

The position of the pump doesn't really matter - pump position has nothing to do with performance. I doubt you increased your tube length significantly to have a measurable difference... That only leaves 90 def bend but seriously I highly doubt that one single 90 degree bend will drop your CPU temp by 6C... there has to be something else. Besides, Maze4 is not a direct-center cooling block which means that even with increase in flow, temperature difference between before and after will not be much.
ie - Swifty 5002 - since its not a center cooling block, even with increase in flow there is not a significant difference in c/w to cause a significant difference in change.

All this notwithstanding, the only changes were the ones stated. As I stated, the pump position change and tubing length increase could not have an appreciable effect on the flowrates and therefore the temps. Unless the thermocouple suddenly developed a glaring inaccuracy, (a bit too coincidental) the fact remains that the temps dropped significantly. I'd call that emperical evidence with all variables accounted for. One mitigating circumstance here may be the restrictive nature of the type of elbow (90-degree, straight, not curved) and its position (screwed directly into the pump output). Even a Maze 4 would see better efficiency with a significant increase in flow rate.