WaterChill CPU/VGA/Chipset Water-Cooling System

Flow Rate Testing

One of the first things I like to do with any new system or water-cooling component is to measure the basic flow rates.  The initial pump flow tests indicated that this system will be operating at a relatively low flow rate (<76 GPH) since that’s the best the pump could do without being connected to any components. 

I spent considerable time flow testing as many of the various component configurations as possible.  I checked the system flow with components connected in series, then in parallel, with a reservoir and without, with the Asetek pump and with a stock Hydor L30 II, with 3/8′ ID tubing and then with ½’, and finally just the Antarctica waterblock without the Chipset and VGA blocks. 

To perform these flow tests I measure how long it takes to fill a 5-gallon bottle (gravimetrically calibrated).  Each test is repeated three times.  Results are typically very repeatable with a margin of error of ~1%.  The precision flow meter I used during the pump PQ testing cannot be used for system flow measurements because the flow meter itself introduces flow resistance, which would make the results invalid.

Hydor L30 II pump free flow

Both the Asetek Hydor and a stock Hydor pump were connected to a Tupperware reservoir with a 3′ long piece of tubing.  The discharge of each pump was routed to the bottle with a 12′ length of tubing.  The height of the overflowing reservoir was maintained at the same height of the discharge to eliminate any gravitational head affects.

(click to enlarge)

I measured the free flow of the Asetek modified Hydor L30 II with push-on fittings and a stock Hydor L30 II with ½’ OD hose barbs (for use with ½’ ID tubing).  As you can see, this basically confirms and coincides with the PQ curve results of these two pumps.

Wow — that’s quite a big difference!

System Flow Testing

To begin our system flow tests I will use the flow path recommended by Asetek in the WaterChill Owner’s Manual.  This is a standard series circuit.  I mounted the waterblocks to a dead motherboard and video card to simulate the actual layout inside a working computer.  An overflowing (constant water level) Tupperware reservoir was used to supply water to the system during testing.  The height of the outlet tube flowing into the graduated bottle was maintained at the same height as the water level in the Tupperware reservoir to cancel out any head affects.

(click to enlarge)

All in all, I tested 10 different flow path configurations.

1) Asetek Series System Flow:  

Reservoir > Pump > Radiator > CPU Waterblock >> Y-connector > Chipset Waterblock > VGA Waterblock > into bottle

2) Asetek Parallel System Flow:

Reservoir > Pump > Radiator > CPU Waterblock >> Chipset Waterblock AND VGA Waterblock >> Y-connector > into bottle

3) Asetek Series System Flow (without Reservoir):

Pump > Radiator > CPU >> Y-connector > Chipset Waterblock > VGA Waterblock > into bottle

The next four system flow tests use a standard Hydor L30 II pump with ½’ OD hose barbs instead of the Asetek modified pump.  Adapters were used to connect the 3/8′ ID Asetek tubing to the stock Hydor L30 II pump.  The Asetek Reservoir was not used during these tests.

4) Series System Flow with Stock Hydor L30 II Pump (without Reservoir):

Stock Pump > Radiator > CPU Waterblock >> Y-connector > Chipset Waterblock > VGA Waterblock > into bottle

5) Parallel System Flow with Stock Hydor L30 II Pump (without Reservoir)

Stock Pump > Radiator > CPU Waterblock >> Chipset Waterblock AND VGA Waterblock >> Y-connector > into bottle