#3 lol this is XS
#3 lol this is XS
[SIGPIC][/SIGPIC]
PENT E8400 batch #814A014 ...4.3 at 1.34v~4.7 at 1.45v
FOXCONN MARS
COOLIT Eliminator 7*c idle~27~38*c load $95bucks !
BUFFALO FireStix's ddr2-800 do 1200 eazy at 2.1v
OCZ 2x2 kit pc2 8500 - 1066 @1069 atm
Quattro 1000W
Radeon 2-4850's in crossfire
OCZ Vertex SSD thanks Tony!
ALL PIPED INTO HOUSE AIRCOND ;}
*QUANTUM FORCE* saaya & sham rocks !
*REAL TEMP*
At least you've got some Xtreme software now for working in Xtreme situations! "Unclewebb" rocks !
*MEMSET* Felix rocks !
*SUPER TEC MAN* UncleJimbo rocks !
OVERCLOCKERS MAG..http://www.xtremesystems.org/forums/...d.php?t=197660
LOL, this is getting rather interesting....
I just bought another Freezone cause I couldn't stand it any longer, this time the Eliminator model (NIB, $199 retail, got it for $99 on Ebay, I think the seller didn't know what they had), going to add it to the present loop (Oringinal Freezone + DangerGen 120mm Xflo), so it goes like this, warm cpu block water hits rad, gets precooled first, then on to the new Eliminator for second cooling phase and then immediately to the Freezone for the last cooling 'rinse cycle' before heading back to the block.....by the time the water hits the block it should be rather cold, hoping for no condensation.....I am interested in full load temps, want to see if I can knock off 20*c, getting below 30*c full load....I opted for the Eliminator because I will be using a single PS to run the whole rig, hoping a 700watt OCZ unit will suffice for everything...
I reported my present temps many posts back with some flack from bigwater enthusiasts, will report the improvement, or not, when the parts arrive....should have them in about 5/6 days....
I am doing this to further show that bigwater thinking doesn't apply to pelt chilling....
laterzzzz..............
Last edited by bldegle2; 02-08-2009 at 10:11 PM.
Asrock 970 Extreme4, Vishera 8320 @4.6Ghz, 1.39v, 16 gig Gskill RipJaws X DDR3 2133 @2284, OCZ 700w, OCZ Vetex 4 256gb boot, ATI 6850, all on big air..
smoke and mirrors
I plan to insulate my mobo.Originally Posted by bldegle2
I will have to look closely at the waterblock and figure the best way to position the frosty king insulation tape (1/8 inch thick sheet) both top and bottom of the mobo.
Under that I will dielectric the socket, seal around the socket with rubber adhesive and conformal coat all around the socket area.
Can you get sub ambient with your setup, I think you probably could.
I hope it doesn't condensate, dielectric grease and closed cell foam is a pain.....I stopped using my phase unit, mostly because of the constant hassle changing CPU's, and also because it is rather expensive on the electrical costs to run 24/7....with my 6400+ I was able to get to single digits c* idle with just the Freezone only on cold days, didn't have the radiator in the mix then....
It will be a couple of days, proly the weekend....
laterzzz.............
Asrock 970 Extreme4, Vishera 8320 @4.6Ghz, 1.39v, 16 gig Gskill RipJaws X DDR3 2133 @2284, OCZ 700w, OCZ Vetex 4 256gb boot, ATI 6850, all on big air..
smoke and mirrors
This will take me a while to properly integrate
Water temps are likely to be condensation risky.
Need to insulate the mobo and cpu, planning on going subambient, at least for testing.
Flat out and sweating I am expecting the boreas to get cpu temps down close to single digit C at idle and ??? at load.
I will probably start a new thread, first pics ...
That Boreas should hold the full load temps to under 30*c, which is what I am shooting for with the double chiller + rad loop....I would have gotten a Boreas unit, but, jessh, the costs approached what I paid for my Promie Phase change unit back in the day (still have it), and, when you get that sucker up and running you will find out the 92's make hella noise at full load.....and you will prolly have to change out the water block, high flow blocks don't workie so well with low flow chillers.....
I changed mine out from the stock 92's to 120mm with a funnel (92 to 120) adaptor, same performance, mucho, mucho quieter.....
The Boreas packs some power.....just remember, when adding a pelt chiller, one has to slow the water flow down to around 210/230 L/HR (not the 500+ that most bigwater setups use), use the pump that came on the Boreas, it will suffice.....if you add this to your setup as it stands now, you will be disappointed in the performance as the fluids will be moving too fast to allow the pelts to chill the water.....slow is the way to go with pelt chillin'.
Post your results, I am adding a second chiller to my loop this weekend and still using the original Coolit pump (single pump for the whole enchilada), will post back with some numba's....
pelt chillin' is fun....on cold mornings I was able to get my 6400+ down to 4/5*c at idle with just the original Freezone and some powerful fannage, and without the radiator, bummer, I added the rad after I got the 940's.....
Laterzzzz..............
Last edited by bldegle2; 02-12-2009 at 08:19 AM.
Asrock 970 Extreme4, Vishera 8320 @4.6Ghz, 1.39v, 16 gig Gskill RipJaws X DDR3 2133 @2284, OCZ 700w, OCZ Vetex 4 256gb boot, ATI 6850, all on big air..
smoke and mirrors
We're back to this dis information again
Slower flow rates does not mean better temps. It just leads to greater temp variance along the loop.
If you slow the flow rate down to reduce the temp of the water exiting the chiller . You are therefore also slowing the water down going though the CPU. this results in more time for the water to heat back up in the CPU block. So any reduction in temps leaving the chiller is lost because the same effect that is applied to the cold side is applied to the hot side but in reverse. So the net result is no gain
Last edited by Ultrasonic2; 02-12-2009 at 11:42 AM.
Ultrasonic,
No, YOU are spreading misinformations, pelt chilling, especially the Coolit way, is designed for low flow...you put a high flow pump in the mix and your temps will soar, or in the case of SteveRo, maybe no change in temps as he is running a bigwater loop already.....the only way one can cool the warm fluids with their design is to pass warms fluids rather slowly past the pelts....especially given there is no real radiator in the mix....thus the reason I added a rad, and then saw some good reduction in full load temps, actually HUGE...all with low flow 210/230 l/hr, and the Boreas is no exception.....why do you think they used 1/4" tubing???? think about it logically....
I already tried a higher flow pump with disasterous results temps wise. but, hey, until you actually try it.........and maybe a question regarding this to Coolit's Tech support will get you a more informative answer, I didn't design the unit, just tinkered with it....it is you spreading FUD, not me.............
whatever...
laterzzzzz..............
Last edited by bldegle2; 02-12-2009 at 01:27 PM.
Asrock 970 Extreme4, Vishera 8320 @4.6Ghz, 1.39v, 16 gig Gskill RipJaws X DDR3 2133 @2284, OCZ 700w, OCZ Vetex 4 256gb boot, ATI 6850, all on big air..
smoke and mirrors
This is total bull
Pelts remove given wattages with certain temperature differentials, and so keeping the pelt with not-colder-than-loop-average water is best (by giving lots of flow). Waterblocks also like flow. With more flow, more wattage is removed by the pelts, and the waterblock's differential between coolant temps and core temps decreases. Don't bother arguing back unless you tell me which part of this is incorrect.
e8600: 6261 mhz (LN2)
e8500: 5830 mhz (dice)
It's really unfeasible, unpractical, and for all intensive purposes... SHOULD NONE THE LESS BE ATTEMPTED!
I plan on testing with the coolit pump - should be lower speed and with a MCP 355 pump - so should be higher flow because more powerful pump but still through the Boreas 1/4 barbs (looks like this can't be changed, at least not by me). So it should be an interesting comparison. I am conformal coating my board now.
eligray,
Mixing bigwater hiflow parts with a Coolit product is not going to work....
There are many ways to cool with a pelt, with this I totally agree, again, with the Coolit design and 1/4" tubing it is built for LOW FLOW....this is a senseless argument, you obviously have not worked with the Coolit product, especially if mixed with bigwater tech and hi flow thinking....
And, the reason it works is low volts to the pelts (4v to 5v) and a preponderance of pelts, it is a balance between heat generated by the pelts themselves and cooling efficiency that can be obtained because of the low volts....you just can not mix the two techs (bigwater and Coolit tech), they will fight each other....
If you are so adament on your thoughts, I suggest you ACTUALLY make a setup and report back your HONEST results...
SteveRo, this rant was not directed towards you, but I will caution you, mixing of Bigwater 1/2" lines, hi flow pumps, hi flow blocks and coolit low flow 1/4" lines will not work and you will be disappointed, and the converse will disappoint you too, low flow pump with the rest high flow parts......since you have the parts you can at least do the testing to verify what I have already confirmed with my testing.....just make sure you have the tubing buttoned real tight, the head pressure created by a high flow pump and low flow 1/4" barbs could be a recipe for disaster.....
laterzzzzzzzzzzzzzzzz...................
Last edited by bldegle2; 02-12-2009 at 02:57 PM.
Asrock 970 Extreme4, Vishera 8320 @4.6Ghz, 1.39v, 16 gig Gskill RipJaws X DDR3 2133 @2284, OCZ 700w, OCZ Vetex 4 256gb boot, ATI 6850, all on big air..
smoke and mirrors
yep - I plan to clamp everything down good with hose clamps.
They use 1/4 inch piping to reduce the cost.
they use a low flow pump for 2 reasons
1 to reduce the cost
2 a less powerful uses less electricity it then generates less heat to be cooled by the tecs and heat sinks
I may plan to try this also - would this be good? -
block - feser 480 radiator - coolit pump - boreas - block
If the water is below ambient then this is counter productive.
But i guess the water temp will be above ambient because you'll be using a low flow rate . Give it a go . you can never do to much testing
"They use 1/4 inch piping to reduce the cost.
they use a low flow pump for 2 reasons
1 to reduce the cost
2 a less powerful uses less electricity it then generates less heat to be cooled by the tecs and heat sinks"
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Ultrasonic2,
Lets see, the real reason they use 1/4" is because the tubing is in close quarters (because it is a self contained unit) and has to bend very sharply, in fact, the lines are wrapped with spring coils to facilitate this, not to reduce the costs, as you stated.....3/8" or 1/2" would not work....
and the reason the lines can be 1/4" and the pump low flow, you werent' even close, again, the way pelt chillin COOLit style works is specifically BECAUSE of low flow, it allows the heated CPU and chipset block water enough time to actually 'supercool' to temps lower than what would come out of a bigwater unit only.......so when it hits the blocks, it can absorb enormous amounts of heat and still keep temps in control....
why do you keep knocking this tech? i strongly suspect you do not have one of these units, methinks it is argueing for arguements sake only....
if you want to continue the harranging, PM me, I would be more than happy to hear your SPECIFIC explainations and see your testing data....
laterzzzzz.................
Asrock 970 Extreme4, Vishera 8320 @4.6Ghz, 1.39v, 16 gig Gskill RipJaws X DDR3 2133 @2284, OCZ 700w, OCZ Vetex 4 256gb boot, ATI 6850, all on big air..
smoke and mirrors
Lets see, the real reason they use 1/4" is because the tubing is in close quarters (because it is a self contained unit) and has to bend very sharply, in fact, the lines are wrapped with spring coils to facilitate this, not to reduce the costs, as you stated.....3/8" or 1/2" would not work....
That is your opinion and your entitled to have one
im going on about it because what you said is incorrect and always will be.
Better performance due to greater flow rates has been tested time and time again (excluding increased heat dump due to the increased power of the pump)
it's easy as for anyone to test you just squeeze the tubing to reduce the flow
i have a multi TEC water chiller that is keeping my pc cool right now. Whether or not someone has one is irrelevant. The Coolit is controlled by the same laws as my chiller and everything else in the world.
As stated before i have no doubt that reducing the flow rates will produce lower water temps entering the cpu block but this does not result in lower CPU temps because the same effect is applied to the cpu block bringing the temps back up again.
It's logical the effect that your applying to the TEC cold side is also being applied to the CPU hot side canceling any possible gains out.
Last edited by Ultrasonic2; 02-12-2009 at 05:57 PM.
Way to spell it out Ultrasonic. I think we've explained this as well as possible, and he just doesn't seem to get it.
Yes, by lowering flow, the coolant will be colder directly after the pelts. However, it will also heat up more in the waterblock.
e8600: 6261 mhz (LN2)
e8500: 5830 mhz (dice)
It's really unfeasible, unpractical, and for all intensive purposes... SHOULD NONE THE LESS BE ATTEMPTED!
IMHO, I doubt their credibility and testing accuracy. E.g. take a look at TR Ultima90 at:
http://www.hardwarecanucks.com/forum...review-11.html
And TRUE black at here:
http://www.hardwarecanucks.com/forum...-review-8.html
It shows that Ultima90 is 2.1C cooler than TRUE black for full load test (Q6600@3.4GHz).
My LOBO rig:
AMD X2 Windsor F3 3GHz 1.20Vcore
2x1GB Crucial Rendition DDR2-750 4-3-3-4 @ 2.1V
Abit NF2M mATX
Voodoo 3 2000 PCI graphics card
Gigabyte Odin PRO 550W PSU
40GB Seagate Baracuda SATA II
15" monitor
The truth is lets say you have a CPU with a TDP of 125 watts and you have a TEC with a Qmax of 125 watts and the hotside is able to displace all of that heat you would end up with no net gain or loss. No matter what flow rate, if the flow rate was lower the temp of the water going to the CPU would be lower but would also be neutralized by the energy the CPU outputs. In terms of thermal dissipation higher flow is better because of minimal temperature gains and losses. That does not change with TECs however the maximum thermal dissipation does. If you have a higher Qmax it would in theory absorb more energy from the water than the CPU would put out. However if your Qmax is lower than the CPUs energy output then you would end up with a constant net gain of energy until it reaches energy equilibrium. The argument about colder and hotter is just irrelevant and stupid. So lets just go back to the original topic.
I am a senior electrical engineering student at Cal State Northridge and in the case of closed loops, Ohm's law applies to fluids as well.
Ohm's Law: V=IxR
Voltage (V) is equal to the difference in water pressure between the inlet and outlet, current (I) is equal to the flow rate of water, and resistance (R) is any obstruction in the path of the water. A pump in our case would be a voltage and current source at the same time. Anything not a pump, such as water blocks or radiators, is a resistor.
A single loop containing a pump, boreas, water blocks, radiator, and other parts is a series loop. In a series loop the current (flow rate of water) is the same for each resistor in the loop. But the voltage (pressure) across a resistor varies with each resistor. This is the common setup for water systems.
Next case, take all the resistors and put "T's" on each inlet and outlet, then string the inlet "T's" together and attach it to the outlet of the pump. String all the outlet "T's" together and attach it to the inlet of the pump, cap both ends. This is the parallel case, the current from the source (pump outlet) now splits to pass though each resistor and rejoins back at the source (pump inlet). The voltage (pressure) across each resistor is the same.
So how does this apply? Their seems to be an issue with high and low flow rate (current) and what is best for heat transfer given certian parts. I dont have the parts and i can't test this high or low flow issue, though other people have or are going to test this.
I do have a solution to have both high and low flow in the same loop at the same time. Build the loop in series, but when the loop gets to the Boreas unit, split the series into two parallel sections. One parallel section contains two of the four Boreas TEC blocks and the other parallel section contains the other two TEC blocks. Then recombine the parallel sections into series after the Boreas unit. I assume the Boreas blocks are all the same so they have equal resistance and hence the current (flow rate) in each parallel section of the boreas is half that of the series current (flow rate). When the parallel sections rejoin into series the current (flow rate) adds, returning to normal. The pressure drop due to the parallel section will be minimal compared to the series. One step farther is to split the Boreas section into four parallel sections, one for each Boreas block, quartering the original series flow rate for each parallel section.
Since the barbs on the Boreas are smaller then 1/2ID, each Boreas water block will incur a very large resistance. But when put in parallel, assuming all the blocks have the same or close resistance, the halving or quartering of the current (flow rate) due to being in parallel will make up for the large resistance and the voltage (pressure) drop will be minimal. The minimum pressure drop and flow rate would be the four parallel Boreas sections. The two Boreas parallel sections has the resistance of two blocks added together and only half of series current (flow rate), making the pressure drop more then the four section, but less then if it was in series.
Last edited by pby5cat; 02-13-2009 at 01:45 AM. Reason: spelling
Ideally they would put things like TECs in parallel however that increases the expense because then you need more blocks instead of one larger piece but yeah I do agree with the way you put it there. If it was in parallel the more energy each TEC group could absorb in series the first set absorbs the most energy and by the last set it has little work to do.
Actually, in a 1 gpm system, its something like 300W is required to heat up the water by 1 degree c on its way by.
e8600: 6261 mhz (LN2)
e8500: 5830 mhz (dice)
It's really unfeasible, unpractical, and for all intensive purposes... SHOULD NONE THE LESS BE ATTEMPTED!
FYI - Just in case you didn't see it yet - I have started a new Boreas assessment thread at -
http://www.xtremesystems.org/forums/...d.php?t=217758
gpm is a unit of time how much heat it takes is not related to time itself but mass or quantity. Note why you always calculate thermal units in J/gram or some such. water is roughly 1g per ml so (2 000mL)(10C) = 20 000 Cal, 4.184 J/Cal so 20 000 Cal*4.184 J/1 Cal = 83 680J, 1w = 1J/sec so 83 680J * (1w/h)/(3600j/s) = 23.26W/h. So 300 watts in one hour is 300w * 60min = 18 000J/min. 18 000J/min * 60sec/1min = 1 080 000J. 1 080 000J =! 83 680J. Water is one of the few elements that has such a high ability to store heat. It requires an immense amount of energy to heat up or cool down water. Someone could check my math but I'm pretty sure it's right not positive since I'm a bit rusty at this.
Posting Permissions
Reply With Quote
Bookmarks