Best Pump from Pet Store

[Skeen]

I used a Via in the past and liked it alot. It was cheap as hell, had 1.5 meters of head and never failed after 2 years of use. Hell, in 2 years you may not even care if it breaks. I feel water cooling is mainly hype anyway (unless you're using a water chiller). You can get a high end heatsink that'll cool just as well. Noise you say? My heater core used 2 12cm Deltas. Now I use one.

 

[greenman100]

I used a Via in the past and liked it alot. It was cheap as hell, had 1.5 meters of head and never failed after 2 years of use. Hell, in 2 years you may not even care if it breaks. I feel water cooling is mainly hype anyway (unless you're using a water chiller). You can get a high end heatsink that'll cool just as well. Noise you say? My heater core used 2 12cm Deltas. Now I use one.

that just doesn't hold up

if it were true, why would Intel and other big OEMs be prototyping self contained watercooling solutions?

 

[Skeen]

Well if you look at the numbers (as posted by overclocker.com), you're right it doesn't add up. The best water block has a c/w of .12 while the best heatsink has a c/w of .22. Of course I could use a a better fan then they did, and you, for instance, could use a greater diameter hose, a more free flowing rad, whatever. My point being is that everyone's mileage may vary. I should have said that I feel watercooling is inferior from a price/performance standpoint.

As to why OEM's are finally trying to get a piece of a new computer enthusiast market, I'm sure it has more to do with greater marketshare than an inability to cool a part in any other way.

/edit- The XP-90 has a c/w of .13. Would you rather pay $40 or $100 for similar results.

 

[greenman100]

/edit- The XP-90 has a c/w of .13. Would you rather pay $40 or $100 for similar results.

You can't compare that ot the bumbers of the die simulator. If you read the testing methodology, the P4 numebrs come from a montherboard. To compare, you'd have to go socket A to socket A, in which case the best tested socket A sink, the SLK900, has a C/W of .23. The best WB has a C/W of .12.

Assuming a radiator C/W of .04 (medium speed fans on a double heatercore) and a 120w CPU... There's a 8.4C difference.

8.4C cooler is not similar IMHO, and you'll have less noise with the medium speed fans than the 74 DBa 6000 RPM vantec

 

[Graystar]

that just doesn't hold up

I agree.

A major contributor to this perception seems to be bad temperature sensors, where the same sensor reads 3 degrees low at idle speeds and up to 11 or 12 degrees low at max power output. Couple that with sensors that read higher because the HSF airflow is gone, and you got lots of computers that don’t appear to run much cooler under water as they did under air.

But read the CPU’s internal sensor and story changes dramatically. Even a mediocre watercooling system can drop the CPU temp by 10C.

I have a dual processor system and my hard drive is the loudest part of my computer now. If the benefits of watercooling weren't real this forum wouldn’t have lasted this long. People would have dropped the effort long ago.

 

[Alacritan]

Well if you look at the numbers (as posted by overclocker.com), you're right it doesn't add up. The best water block has a c/w of .12 while the best heatsink has a c/w of .22. Of course I could use a a better fan then they did, and you, for instance, could use a greater diameter hose, a more free flowing rad, whatever. My point being is that everyone's mileage may vary. I should have said that I feel watercooling is inferior from a price/performance standpoint.

Watercooling, until recently, has always been an enthusiast market only being adopted by OEMs because they now NEED it. Price/performance is much less important to the enthusiast than the average person, which is why we buy expensive graphics cards and cooling equipment and the average person buys a Celeron based rig from Dell with stock cooling and isn't even aware that hardware temperatures affects PC performance. The cost of performance goes up significantly the cooler you want your PC to run. It's not a linear thing.

 

[Alacritan]

please stop posting what you don't know about

kinetic energy?

Considering watts fed into pump is always increasing while the pump is on

KE=1/2mv^2

if KE is increasing, mass of water is constant, then velocity keeps increasing?

so my water is going to eventually reach the speed of sound, then the speed of light?

hi, my name is friction, and I don't let that happen. Friction turns that extra KE back into heat, including fristion of the water hitting other water molecules and barriers such as metal and plastic.

As for the C/W rating: Speaking of BillA, his tests show a thermochill 120.1 at ~.045 C/W with a medium speed fan, and ~.025 C/W with a high speed fan. He also found a C/W of ~.04 with a medium speed fan and ~.02 with a high speed fan for the thermochill 120.2.

http://www.overclockers.com/articles778/

So, what happened to gravity in there? When I throw a ball up does it stop going up because of friction? You're ignoring some important **** here.

And a Thermochill 120.1 is NOT an average radiator. It's a crappy radiator. A heatercore is significantly better, and a heatercore that can fit 2 120mm fans is WAY the hell better than a Thermochill rad including the 120.2. The closest thing to an average heatecore would be a Bonneville but there are enough with a single 120mm heatercore to say there is no real average. Thermochill rads are a bad example in particular considering their poor performance.

 

[greenman100]

So, what happened to gravity in there? When I throw a ball up does it stop going up because of friction? You're ignoring some important **** here.

And a Thermochill 120.1 is NOT an average radiator. It's a crappy radiator. A heatercore is significantly better, and a heatercore that can fit 2 120mm fans is WAY the hell better than a Thermochill rad including the 120.2. The closest thing to an average heatecore would be a Bonneville but there are enough with a single 120mm heatercore to say there is no real average. Thermochill rads are a bad example in particular considering their poor performance.

ok, so you are implying kinetic energy is stored as potiental

but that potiental energy is released when the water comes back down the other side of the loop.

LOL care to back up the thermochill statements?

a 2" thick heatercore is not suited well for a 120mm axial fan of any decent noise. A blower is needed to optimize performance

 

[Alexwai]

And a Thermochill 120.1 is NOT an average radiator. It's a crappy radiator. A heatercore is significantly better, and a heatercore that can fit 2 120mm fans is WAY the hell better than a Thermochill rad including the 120.2. The closest thing to an average heatecore would be a Bonneville but there are enough with a single 120mm heatercore to say there is no real average. Thermochill rads are a bad example in particular considering their poor performance.

I don't think so. According to my casual testing to compare the performance between the 2-342 and Thermochill HE120.2, the delta between the water temp and the radiator air-in for both radiators is very close. 1-pass did better but a very little on my setup. Anyway, Thermochill HE120.2 is not that bad.
Click here

 

[ls7corvete]

So, what happened to gravity in there? When I throw a ball up does it stop going up because of friction? You're ignoring some important **** here.

And a Thermochill 120.1 is NOT an average radiator. It's a crappy radiator. A heatercore is significantly better, and a heatercore that can fit 2 120mm fans is WAY the hell better than a Thermochill rad including the 120.2. The closest thing to an average heatecore would be a Bonneville but there are enough with a single 120mm heatercore to say there is no real average. Thermochill rads are a bad example in particular considering their poor performance.

I had the same missconception until BillA set me straight. Until we find some single pass 1" or .75" thick HCs we are working with less-than the best for our application. Not that the price doesnt make up for it......

 

[Alacritan]

That's why I switched to a Derale transmission cooler. I get the same temps as with my 2-342, except my fans now only have to run at 5V instead of 7V to say at 40C or under.

 

[greenman100]

That's why I switched to a Derale transmission cooler. I get the same temps as with my 2-342, except my fans now only have to run at 5V instead of 7V to say at 40C or under.

you still haven't backed up your thermochill statements

 

[Alacritan]

Well, by your reasoning, a 2-342 should most definitely be better. It's 1 1/2" thick whereas the Thermochill rads are 2" thick (admittedly though I did have a BIX in my mind). There may not be as much difference in performance as I had thought now that I've checked out the Thermochill rads. They actually look like repackaged heatercores.

But I maintain that a 2-342 is better than a Thermochill. If an appropriate party would like to borrow my 2-342 for testing to officially test it, I wouldn't mind at all.

 

[Alexwai]

It's 1 1/2" thick whereas the Thermochill rads are 2" thick (admittedly though I did have a BIX in my mind).

Have you messed them up?

1-pass heatercore is: 2"
Thermochill 120.2 is : 1 1/2"

 

[Skeen]

You can't compare that ot the bumbers of the die simulator. If you read the testing methodology, the P4 numebrs come from a montherboard. To compare, you'd have to go socket A to socket A, in which case the best tested socket A sink, the SLK900, has a C/W of .23. The best WB has a C/W of .12.

OK fine but this still doesn't invalidate my main claim that watercooling isn't the best bet from a price/performance standpoint. The best on the list, the Hydrocool Hydro-Stream HS5, is around $50 just for the block.

As to whether the price/performance ratio should be a factor, I think yeah, since [KX3]rAge is posting asking if he can shave a few bucks off by buying some pump from the pet store. If that wasn't the case, why screw around with watercooling? Why not jump right into phase change?

 

[JerMe]

Hey, I'm trying to put together a cheap yet very effective water cooling system and I'm now working on the pump. I figure getting an aquarium pump from a pet store would be a lot cheaper then buying a pump specifically designed for water cooling yet it would/should serve the same purpose.
What pumps have you guys exerimented with that were bought from a pet store? Are they in general cheaper then a water cooling specific pump??
All input appreciated

Head over to Petsmart's outdoor pond section. You can probably pick up a Danner Mag3 for about $25-$30, YMMV. Don't trust their website price - my mag5 is listed as $57.99, but the shelf tag priced it at $27.50. I think the mag3 was $25 or so, but was OOS. Make sure you walk PAST the AQUARIUM section and go straight to the POND section, that's important if you want to find your way to the pump.

Remember if you have a choice, get the mag3 over the mag5, as they'll perform the same in our W/C environment (actually the mag5 does worse as higher flow restriction). If you want to be hardcore, then go for a mag7 if they have it (although I'm not sure what the rest of the guys here will think of that). Of course that'll mean more heat, (more noise?), and more $$$.

If you choose to buy from an online vender, you'll find the mag3 for $40 at Marine Depot (http://www.marinedepot.com/md_viewItem.asp?idproduct=DN1121).

That's your question, right?

 

[greenman100]

OK fine but this still doesn't invalidate my main claim that watercooling isn't the best bet from a price/performance standpoint. The best on the list, the Hydrocool Hydro-Stream HS5, is around $50 just for the block.

As to whether the price/performance ratio should be a factor, I think yeah, since [KX3]rAge is posting asking if he can shave a few bucks off by buying some pump from the pet store. If that wasn't the case, why screw around with watercooling? Why not jump right into phase change?

top of the line air: $50
top of the line watercooling: $170
top of the line phase change: $500+

by your logic, why screw around with a corvette? why not just buy a ferrari?

watercooling IS the best bet from a price/performance/NOISE standpoint, though

 

[Skeen]

top of the line air: $50
top of the line watercooling: $170
top of the line phase change: $500+
by your logic, why screw around with a corvette? why not just buy a ferrari?

This wasn't my logic. I made my comment about phase change in response to Alacritan's assertion that:

"Price/performance is much less important to the enthusiast than the average person, which is why we buy expensive graphics cards and cooling equipment and the average person buys a Celeron based rig from Dell with stock cooling and isn't even aware that hardware temperatures affects PC performance."

watercooling IS the best bet from a price/performance/NOISE standpoint, though

I might give you the noise factor though I personally used 2x12cm Deltas on my heater core and now I use just one on my SLK900.

 

[greenman100]

those deltas were probably overkill

I have two low noisefans on my heatercore, my pump is louder than the fans, and I can overclock to 2.55ghz just fine

 

[Skeen]

those deltas were probably overkill

Yeah could be.