Thanks, will do. I will switch back to stock nozzle tonight after getting off from work.
Printable View
Great work Martin! My Quantitative Methods professor would love you. ;)
I can't wait to see some Nehalem results. Larger, non-square IHS and a single, large monolithic core with about the same surface area as a dual-die Penryn. Think it would make a difference any?
At least from the tests I ran, the GTZ was doing a little better at holding the core temperatures between cores closer to each other. this is probably due to the larger slot like nozzle of the GTZ vs a smaller circle of the fuzion.
Yeah, I'm not so sure I would be too worried about flow rate though. They were close enough to me not to get really excited about it. Especially when you start considering pump heat. While the fuzion ran higher flow rates it also ran with water temperatures that were almost .1C higher even with the larger quad radiator and 1550RPM fans. If you had a smaller radiator that could mean .2C or hotter water temps. I'm not sure the hydraulic gain would be there to make up for the heat. Regardless fractions of a degree are pretty hard to measure, even after 5 mounts and all these sensors, I'm still only good to about a half a degree or so.
My nozzle plate fits very accurately, make sure you don't have it backwards or upside down. Mine will come out and slip back in without much effort at all. I need to get some internal pictures done then I'll be able to show you a better picture of the nozzle orientation.
They might be the same threading, but the standoff are different. You could probably get either one to work, but I would recommend using the backplate that comes with it otherwise you might not install the block with the intended pressure. They both bottom out very specifically and you could overtighten it with the V2 backplate.
Andy seemed to confirm with his quad that was the case. My first round that I had issues with my ambient and had to throw out was showing the quad midplate doing about a degree worse than the stock midplate. But you do need the midplate for sure.
Just watch the o-ring seating on the compressions. I don't have any bitspower compressions, but the TFC compression will impact the swiftech logo that's raised above the block top because the barb seats are recessed. I think you could get them to work, but you might have to remove the logo first or use thicker o-rings.
No, I consider anything over 1GPM good, and 1.5GPM excellent. I want to do some pumping power vs heat dump experiements later, but I suspect the DDC3.2 with top as I had tested is too much pumping power for optimal performance on some of the blocks. Hard to say without testing, but a CPU only loop I would never worry about flow rate, just bottom line performance. Only in a multiple block loop would I worry at all about flow rate and both of these blocks are pretty good with that using a D5 or DDC, you'll still have plenty of reserve pumping power for most setups.:up:
Martin thank you for hard work, totally appreciated. Looks like my next block with be GTZ :up: compression fittings and mounting mechanism sounds so welcoming..
Thanks!
Update the op with some additional pictures of the interior and one with compression fittings on there...
ths is a great block..i cant wait to see your extended testing results!
although i have a fuzion v2 im actually considering switching to a GTZ mainly because of it wider barb spacing (having trouble on the v2 with 1/2" ID tubing and zip ties..) and because of its sexy looks :D )
:) Thats good to know they are the same.
Would the quad insert help for a quad or would it matter much?
Hey martin.
How about doing the exact same test.
Only with say a swiftech 220 radiator?
I'm going to try it again when I go into round two, but this was intended to test straight out of the box as shipped without buying any accessories. My first round of testing I had problems with because of ambients was however showing the quad midplate performing worse than stock. Swiftech's published results show performance of the 4.5mm nozzle which was the best for them. D-Tek noted that some of the nozzles are finicky about installation and some are better for different types of quads, etc. But they also suggested trying the red quad nozzle, it may be helpful for the Kentsfield.
Sorry, it's incredibly time consuming and I've already tested the MCR220. If you want an estimate you can use my MCR220 radiator estimator and estimate your water delta. My testing here with the TFC480 resulted in about a 2.5degree delta, so for a rough number just add the difference from 2.5 degrees. That's not going to be a perfect number though because as the processor gets hotter, it produces more heat. I purposely chose the TFC480 for two reasons, one is because it's a quad rad and I can then use all 8 of my air in sensors and 4 air out sensor. In addition the larger radiator reaches equillibrium much faster than the higer delta of a small radiator, so 10 minutes is enough time for warmup.
Ok cool.
I have never gotten around to taking my Tc320 off and testing my ST220.
I did do it in the reverse order with the same sys only difference was the two radiators.
I wouldnt figure there would be much difference anyway.
Pretty much anyone running a cpu and gpu on a 220 radiator or larger, one wouldnt see a very big drop in temps.
Now of course if one ran an SLI/CrossFire set-up plus a cpu then i could see a 220 style or smaller radiator posing problems.
Anyway thanks always for the test and feedback.
Yeah, probably not going to be more than a couple of degrees differeence depending on the fans use.
The interesting thing I found with my CPU block testing is the .2 to .3C error I was fighting before I fixed my ambient problem. The hotter the processor, the more heat it produces, so a couple of degrees hotter water temperature might translate into a little more than that after you figure in the extra processor heat that's coming out. It's very complex in how it all works, and interesting.
One thing to think about is as the cpu/waterblock/NB/memory, all watercooled or some watercooled.
Will raise the avg/overall temps on the mobo as a whole, and heat will spread over a givin area.
Even more so with the sys under load.
And give a good 1/4 or more of a degree variance in testing.
Ambient heat so too speak.
Does that make any sense?
IMO something like a thermal camera with time recorded video of said sys
from idle to full load would best show when/where and how the heat spreads.
Plus case airflow and fan orientation would also be a big factor for that type of test.
But then again that is a whole nother hair pulling, long, time consuming, lots of drinking and expensive test in it-self.:)
Thanks for all your effort martin!
I have a question, as can be seen there's a 0.4 GPM gap between the FuZion and the GTZ. Now, according to your flow estimator (:clap:), the following system:
FuZion V2
MCW60
2x MCR320 QP
D5
1x Res
6ft. 7/16" tubing
would yield 1.53 GPM. Let's replace the FuZion with a GTZ. 1.53-0.4 = 1.13 GPM.
Isn't it a big difference? I'm afraid 1.13 is not that good, esp. for future upgrades. What would be best for the above setup?
Thanks ^^
While having a target flow rate is somewhere to start, the more I've been playing with this CPU block setup where I get the full picture, the more complex I'm seeing this pump heat issue becoming.
For each CPU block tested I'm recording flow rates, but also water and air temperatures. What I'm noticing is that even with the same pump, the more restrictive blocks have a lower water temperature. Even with the TFC 480, I'm seeing around .1C hotter water with the more free flowing blocks.
So in the end you might have higher flow rates, but also hotter water temperatures. The question then becomes, which is better or worse or does the end resulting performance on your other blocks better or worse.
In the end I think there is probably some sort of "Optimal" flow rate to shoot for and it's going to be a function of pump heat dump.
It's all splitting hairs at this point to the point it really doesn't make a difference in the real world, but I'm not entirely sure running a system at 2GPM with a known pump would be better than say undervolting the pump to 1 or 1.5GPM. I'm not really sure what to shoot for at this point but I know the differences are very small above 1GPM.
wow umm this is very interesting... id like to find out what the optimal gpm is for water cooling?
.1°C hotter water for a free flowing block?......I think I can live with that. :rofl: ;)
:yepp:
It's all moot...
I feel like I'm with my buddy's racing bikes on the dunes. We get all giddy about a carbon fiber clutch lever that cost $50 for a 2 ounce weight reduction. Never mind the 1/4 pounder with fries we ate for lunch, that 2 ounces makes all the difference.:rofl:
It's not as simple as that. The more restriction already in your loop, the less of an impact replacing the FuZion with the GTZ will have in terms of flow reduction. Look at the PQ-curve of the pump, it's not linear. If you could do what you did, adding enough blocks would make the water flow the other way. :p: