Overclock.net › Forums › Cooling › Water Cooling › duel pumps
New Posts  All Forums:Forum Nav:

# duel pumps - Page 3

Quote:
 Originally Posted by ira-k I haven't flow tested it with the 2 pumps...1 was 1.5gpm when I did it in a bucket at 5psi...If you look at Petra's testing the pumps in series actually have more flow then the pumps in parallel not less...If I'm getting 1.5gpm at 5psi thats 0.3gpm for 1psi...So I would think 10psi would translate into 3gpm...I don't see how it wouldn't.... I would break it down and do it in a bucket but its a pain with the Pelt in the loop... Take a look for some of nicksub1's testing there also...I think he came to the same double conclusion also... EDIT: I forgot you like visuals.... 5psi with single DDC.... Close up Side 10psi...I was still bleeding it in this pic, it went on up to 10psi...I'm back down to 9psi since I put a 480 GTX in the loop...10psi was with a 240 BIX... .... Dual Pumps
I wouldn't take it apart to test it, but this is what I'm seeing. The only thing I'm trying to show is that pressure does not equal flow rate, it's not a linear relationship for most systems. We have a pump curve and a system curve and where the two intersect determines the system flow rate.

Petra's results on parallel vs. series makes sense to me. Here is how the two parallel vs series curves would look like plotted on top of each other. For most water cooling setups we have quite of bit of system restistance where running the pumps in series makes the most sense. But that's not alwasy true, in extremely low resistance, you could get more flow rate with a parallel setup.

And if this is correct the only way you could possibly get double the flow rate using two pumps is if you ran them in parallel and you had absolutely zero resistance, otherwise it's only going to be a fraction of that even though your pressure is double when in series.
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
I think your wrong...How are you getting more flow with 2 in parallel...When Petra's test shows the DDC's out flowing in series in an actual loop?

My psi readings are in an actual loop in series...1psi in my loop is 0.3gpm as tested in a actual loop with block rad and res....
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
Because the actual loop has enough resistance that series is the better option, that's the system curve I'm showing.

In the real world you gain 35% in series and only 17% in parallel with that particular setup.

Parallel without any resistance would be like having two pumps filling up a bucket from separate sources of water not connected at all, you would get twice the flow rate of one pump. That's the only way you would get double the flow rate.

Pressure is a good measure of energy within a loop but it's not a linear relationship to flow rate.

I'm going to take my system apart next weekend and I can show you my results with one or two pumps. If I get 1.6gpm with one, I should get 3.1gpm if it was directly related, instead I bet I get something more like 2.1.
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
Quote:
 Originally Posted by Martinm210 Because the actual loop has enough resistance that series is the better option, that's the system curve I'm showing. In the real world you gain 35% in series and only 17% in parallel with that particular setup. Parallel without any resistance would be like having two pumps filling up a bucket from separate sources of water not connected at all, you would get twice the flow rate of one pump. That's the only way you would get double the flow rate. Pressure is a good measure of energy within a loop but it's not a linear relationship to flow rate. I'm going to take my system apart next weekend and I can show you my results with one or two pumps. If I get 1.6gpm with one, I should get 3.1gpm if it was directly related, instead I bet I get something more like 2.1.

Yeah try it and see...Leave your block and rad in it....Having lower head pumps shouldn't affect it any I wouldn't think....

Pressure is the force in a loop head is the energy..
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
Well, what is your setup going to be? Like, the order of things getting hit.
I'll have these components to work with in this test. I'd like to do a real good flushing this time, so everything will be apart.

MCP655 x 2 (both have been ported)
D-tek Fuzion with or without nozzles
MCW60
MCR320 X2
Misc T fittings, elbows, etc

I can run a couple of difference scenarios, I'll even try some parallel vs. series test for fun, but i should get some Y fittings to do that first.

What I always understood is: pressure head is the more common term in hydraulics as the pressure in feet of water and you can easily convert that over to psi or any unit, here is a table:
http://www.csgnetwork.com/pressconvtable.html

The thought on energy behind a system is Bernoullis equation which is a conservation equation that says the total system energy at one point equals the pressure engergy + kinetic engergy + potential energy if you had no friction loss. That equation is a better tool for figuring out energy on open systems where you migh have a inlet and outlet, but its a good starting point. The only energy we can really measure around a water cooling loop is the pressure energy. There is energy behind the flow rate, but because it is constant through the loop, you can't measure any change. Pressure energy should be at it's highest immediately behind a pump or pumps and pressure should drop around a loop as that pressure energy is used up due to friction in the blocks, fittings, and tubing. The hard part about how this all works is the frictional losses are a curve as well as the pump pressure vs flow rate. The less pressure the pump has to resist the more flow it can produce. At the same time the lower flow the system encounters the less pressure drop.

The only way you can determine flow rate is by plotting a pump P/Q graph and a system pressure/drop graph and where the two intersect is the flow rate. You can create an equation for both of these curves and solve it that way, but graphing is the easiest way and I'd have to graph it before I could create an equation anyway.

Of coarse this is all just theory and has it's own share of assumptions, but I'll try to take this and perform some flow rate tests next week with various configurations to see if I can overlay some real world figures in there to confirm.

Let me know if you have any particular scenarios you'd like to see in the mean time.

Thanks!
Martin
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
Yeah your right in a WC'ing loops psi is highest at the outlet and is consumed in the loop and is 0 at the inlet or suction side..At free flow its 0 at the outlet and highest at the inlet or suction side...And PSI converted to head will show flow on a PQ graph....So 10psi in my loop equals 23.066 ft of head...The D-5 will be easy to chart with their PQ chart....You can figure the friction head losses pretty easily also...I was wondering what you were talking about when you said you cant convert psi to flow...

Do you have a pressure gauge you can put in-line right after the pump? I'm sure their would be a lot of guys that would like to see that so they can convert psi to flow...EDIT: It needs to be as close as possible after the second pump to measure psi correctly...And the needle will bounce till its bleed completely...They are great for telling if your loop is getting clogged the pressure will rise...

I might be wrong about mine I don't know...But 3gpm at 10psi with the 2x18W DDC and Petra top would be close...Actually I think its 1.4gpm for 5psi on mine...Its been a while since I've thought about it...

This is the calc I use...PSI to Head
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
I don't have a pressure guage yet, but I think I know how I can create a pressure drop guage out of some T fittings and some extra line to creat a loop. If you create a loop between two points in a loop and connect that with tubing that has water in it and hang it above the loop with an air pocket you the pressure difference will be indicated by the two water levels in feet of water, you can then convert this to psi easily.

I'll have to give it a try but that would only give me pressure difference. Either way I can test a single pump and two in series and compare their flow rates. I'll also throw a parallel setup in there if I can find a Y adapter.
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
I have another thing that will increase pressure for any system. Take your T line and extend it vertically, like strap it to a wall and fill up the line so you have a standing column of water on the the loop. If you could get another 6' of T-line extended and filled up, that should translate to pressurizing your entire system another 2.6psi...yet that won't mean more flow rate.
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
 Daily Test/Play Rig (13 items)
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518
Desk
Yeah you can probably rig up something like that, its kind of like a water level....

I don't think I have any 1/2" Y's or I would send you one..If you ever do want to get a gauge they have the 0-15psi and McMaster-Carr...They have a 1/8" male on the bottom so what I did was get a 1/2-1/2 tube x 1/4" male npt on top and use a 1/4-18" reducing coupling...You can get them with 1/4"npt male fittings but they are 2" dial and the 1/8" is a 1.5" dial and I wanted the smaller one....I think they are +- 1% at mid gauge read-out...
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
 My System (13 items)
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620
Case
Open Bench
New Posts  All Forums:Forum Nav:
Return Home
Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › duel pumps