#4 or #5 for S-TDX / MCP600

[cuddles]

What nozzle? GPU / CPU setup with dual 120mm heatercore and an MCP600. I've heard stories of the #5 clogging and being too much for some pumps. I guess I should simply be trying them both, but I'm just curious what you all think would work best, the 4 or 5.

[Ancient_1]

Yes the #5 can clog and it does have more restriction but also performs better. As close as I can tell I get about 1°c better temps with the #5 compared to the #4. As long as you clean your stuff (mainly the rad) befor you assemble it and be carefull filling you should not have to worry about it. I take my RBX apart every couple months and clean it, I usually have 2~3 of the holes plugged in that time. I think it is from paint that I got in my rad through the barb holes when I painted them.

When I put the G4 or G5 in I will have to check more often since the jet holes in those blocks are much smaller. There are getting to be many blocks that have the Jet holes and will require clean cooling fluid and to me isnt a big deal, since it was my stupidity in the 1st place that is causing my problems.

As for being too much for the pump that makes no sense, all they can do is lower the flow rate a bit the way these pumps work it adds no more actual load to them it just lowers the flow. They pertty much draw less current at lower flow than higher, more restriction means less flow (I really dont understand myself why tho but it is true).

The MCP600 will have no problem with that load.

As for which plate, the way I look at it using the #5 plate is the easiest way I know of to gain 1°c which is quite a bit in water cooling. If the cooling fluid is clean there are no downsides. To check all it akes are a couple random spot checks to see, if its clean then all should be fine, if you are getting a clogging problem then you can decide to use the #4, but on my RBX with 3 holes plugged temps were still as good or better than with the #4.

I have switched between the #4 and #5 using each 4 times and checked temps over a week each time to see if the difference really was there and as far as I can say with just using my MB cpu temps readout it was about exactly 1°c dofference.

[MaxxxRacer]

if you are only running the cpu in the loop go with the #5. it will give better performance as ancient said.

if you are running with a gpu use #4 as it will provide the gpu block with some more flow.

[cuddles]

#5 it is :)

[Ancient_1]

Max it depends on which one you want the best performance on. The CPU will lose more than the gpu will gain by going to #4 over the #5. I would guess we are talking of about .15~.2gpm which would just be a couple 10ths of a degree on a gpu but will make about a full degree on the CPU

[MaxxxRacer]

look at ph's testing.. the #5 brings the flow down a little over .4gpm with a high flow pump like an iwaki. that is over hte #4 plate. With the mcp600 you get 1.1gpm with the #5 and 1.3 gpm with the #4. The differnce isnt as pronouced on the mcp600 as it is with the iwaki class pumps. pH's tests also show only a .2C change between the #5 and #4 with the mcp600.

So the #5 is only really what you want if you are running a high pressure pump with gpu or with a medium pump without the gpu..

Ancient, go check out the raw numbers on the nozzle tests.. its interesting. the best part is that pH used the 50z which is the mcp600

[Ancient_1]

I have and part of that is he is using 70watts as a heat source and all blocks are not exactly the same. That is the reason I have tested the plates so many times in my system. Also he is testing in a loop that is more restrictive than one we would have with a gpu block in (his flow meter is quite a restriction) more than a gpu. That test looks good but isnt quite the same as our loops. And I think also the smaller die of the XP he uses causes some of the differences, Cathar has said the same that PHs temp differences are lower by 40~60% than his own. I am just reporting what I have experienced with my P4 and am sure the A64 will act more like it than a XP as far as temps goes because of die sizes and wattage.

I didnt use the figures I did as wishfull thinking or just pull them out of my A$$. :) If I were not 90% or more confident in those figures I would not of stated them.

[MaxxxRacer]

hmm. yah i remeber about phs loop having some extra resistance from the flow meeter...

I mainly went to phs data as its some actual numbers.. I dont like giving reccomendation with some hard numbers. As i have none of my own, and my temp gauge on my mobo is fuxed up... but its good to have some more data ancient.

[Ancient_1]

I was not saying PHs data was wrong, I believe it is 100% correct but I also believe that my 1 whole degree is if anything a little lower than what I actually see and I am very confident that I got 1°+ better temps with the #5. I also dont see them as contradicting each other either, it is just my heatload is much higher and my die is bigger that I see as the difference (only thing that makes sense to me)

I also have a MSI 875p neo that when I updated the bios my cpu temps got lost lol.
It started to read my prescott temps in the teens under load (boy did that bios cool the prescott)

One of the ways I judge my over all temps is using MBM5s hi lo to check averages over hours and days, that way I can usually see a trend and a factor of how much difference, by doing that a few times with each plate and have each test period showing the same difference is the only way I became confident of the difference of the 2 plates.

[MaxxxRacer]

no no, im not saying that yours or phs data is wrong, its just interesting to see how the different setups chagne the temps.. I agree that 1c is worth the #5, but as long as it doesnt hurt the gpu temps too much. I might try the #5 as my iwaki is a monster and I get good flow no matter what I do.

[Ancient_1]

If you follow the way I just stated you should see a similar gain and be able to somewhat document it.

[MaxxxRacer]

yah the problem is that my board cant read the temps. It think my load temps are 58C.. so measuring and kind of temps on them is useless.