Now what would I do with this?

[Vapor]

Something new coming from me....been hinting at it for awhile now, but now it's up and running













Specs:
3x San Ace H1011s
MCR320
MCR220 Res (fillport is now used as a drain port)
Koolance pump controller (7.65V to 12.5V range with a 12V input)
MCP350 modded to MCP355 under the XSPC Res top
MCP350 modded to MCP355 with the Apogee Drive bottom (MCP350, also modded to MCP355) on the EK Dual Turbo top.
Any CPU block I want
Bunch (six) of Dallas temp sensors on the air-in side.
FM17 flowmeter (good enough for approximations and comparing flow results to itself)

Liquid flow order:
CPU block of choice -> Koolance QD -> XSPC Res top + pump -> FM17 -> MCR220 -> EK Dual Turbo Top -> MCR320 -> Koolance QD -> CPU block of choice.

Air flow order:
Air in -> MCR320 -> San Ace H1011s -> MCR220 -> airflow over the pumps.

A self-contained unit entirely

[ILikeCosmosS]

yeah thats good but not as good as i would make it
changes: take the res/top and just put the top
ek res + male to male fitting with pump +top
directly mount the res male to male with a rad then entering into the loop forming rad+pump+res i

[SNiiPE_DoGG]

the point is here (I think) is that vapor has a nice little block tester going on here..... very portable and easy to set up

[Vapor]

the point is here (I think) is that vapor has a nice little block tester going on here..... very portable and easy to set up

Ding ding ding!

Fuzion V2 with all the various nozzles is up first (one block = 7 blocks of testing, great way to start and refine the method for only the price of 1 block )

Kentsfield is up first, then Bloomfield. Maybe Conroe as well.

5 mounts, each mount I'll test six different flow settings: all 3 pumps on, 2 pumps on (EK Dual Turbo top), 1 pump on (XSPC Res top), 10.5V on the XSPC, 9V on the XSPC, 7.65V on the XSPC Gets a pretty nice temperature vs. flow curve. CrystalFontz is of course used for logging of...like everything, cept CPU temps, OCCT will do that.

[Hondacity]

wooot wooot me can't

[NaeKuh]

so vapor if my math is correct, and you have 3 ddc's, assuming each gives off about 15feet of head... assuming head scales, so you'd have around 40-45 feet of head?

Or to make things easier roughly 15-17PSI?



but your holding the same flow rate.. oh man your DDC's are gonna get HOT.

[Vapor]

I'm not holding flowrates at all, I'm letting blocks dictate the flow rates

All flow modulation will be done by turning off pumps or lowering voltage

[NaeKuh]

I'm not holding flowrates at all, I'm letting blocks dictate the flow rates

All flow modulation will be done by turning off pumps or lower voltage

nah meaning your theoretical max flow rate is that of a single DDC without restriction.

Because you cant get a DDC to spin any faster.

:P

But your pressure scale is gonna be insane.

[Vapor]

I don't understand what you're getting at? Why would I want an unrestricted DDC's amount of flow?

I certainly don't need that much flow anyway

Here's some prelim data I took the past few days before I added in the QDs and finalized load conditions....this is on a 3.6GHz, 1.49V Kentsfield
(EDIT: image removed--no longer have that .xslm file so making a new one with a different y-axis label is a bit difficult )

Seems to be working just the way I wanted it to

[skinnee]

you took the cover off finally!

[brinox]

when you turn the other pumps off, you have added restriction from the disabled pumps yes?

[Vapor]

Of course I have no clue how much though

The flowrates are measured (via the FM17 being polled a few thousand times and then averaged, lol), not theoretical

[Jah]

Frankensteins monster, watch out for lightning strikes it may come alive

[Serpentarius]

Of course I have no clue how much though

The flowrates are measured (via the FM17 being polled a few thousand times and then averaged, lol), not theoretical

have you tried placing 3 parallel pumps? that'll be truly xtreme!!

[basserdan]

Add some backpack straps and you are close to an xtreme notebook lc system

[Ttocs]

Is that graph correct?
So you're looking at approximately 85 degree load temperatures on that Q6600 (Y axis says delta t).

[Webby]

Parallel pumps are only effective with low restriction components for increasing flow as head does not increase and it is head that is needed to force more water through a block, yes you will increase the flow through the block a bit with pumps in parallel but not by nearly as much across all situations as if you had them in series.

Vapor on that prelim figure I guess you have mislabelled your y axis hopefully that is Core temp not a deltaT value! Oh and don't forget to factor in heatdump from the pumps into the situations while not a huge issue if you move from 3 pumps running at full blast ~60W to a single pump at ~10W you will obviously see a difference in water temps and heat dissipation through the radiators. So perhaps a comparison value of core temp against water temp entering the block would be more appropriate as a measure of performance?

[Vapor]

Yeah, that's a mislabel...copied the DeltaT graph and then changed the Y values to ambient corrected values without changing anything else in the graph (oops!)

I wouldn't look too much into those graphs in any way whatsoever, that was with a setup that's different than what I plan on running full-time.

FWIW, I have no intention of measuring water temp. I've considered it, but I think it's fair to ignore it because of a few reasons:
1) I'm not really trying to isolate WB performance--instead, I'm trying to observe cooling performance as WBs and flow rates change.
2) Increasing flowrates requires increased pumping power, it makes sense to me that is it acceptable to embed that variable--finding a water/core delta will likely paint a 'sunnier' picture for highflow setups than what is reality. There's the distinct possibility that a block may perform worse with increased flow simply because of the increase in water temp--this will show that while an isolated WB performance analysis will not. If I'm not mistaken, pump heatdump will also vary from block to block :-/ (this relates to #4)
3) I have a lot more radiator/fannage than what's necessary in the real world, so I really am minimizing the effect of pump heatdump from that end of it anyway.
4) There are others out there that are going to do WB isolating tests I'm not trying to isolate components to put into an estimator, but rather provide some high accuracy real-world data

[Zehnsucht]

provide some high accuracy real-world data

Mm I like! I'm looking forward to your results!

[naokaji]

But your pressure scale is gonna be insane.

The pressure is almost screaming for a certain restrictive block

[gillbot]

Very nice, similar to what i've been shooting for though I'm using 1 D5 and 1 MCR220 but I plan to add another rad soon.

How are those Koolance QD's as far as restriction?

[warriorpoet]

Sleeve It!!!!

[lyl]

Cool test setup! should provide a consistent benchmark for relative test results