Waterblock Testing

A progress report – Joe

SUMMARY: The last 10% always takes the longest time.

I’ve been in the process of assembling a waterblock testing suite (with the kind assistance of Bill Adams) and thought you’d might be interested in my progress to date.

The goal in setting up a waterblock test suite is to control all the variables – water temp, heat load, water flow, etc – so that the only variable uncontrolled is the waterblock. So far, I’ve assembled the following components:

• Forma Scientific Heater/Chiller
• Omega PX2300 Low Differential Pressure Transducer (0-5 psid)
• King Flowmeter
• Eheim 1250 waterpump
• Black Ice Xtreme Rev.2 with Hydor L30 waterpump
• CPU Die Simulator
• Omega Digital Thermometer
• Fluke 2180A Digital Thermometer
• Omega RTD and Thermocouple temp probes
• Two Laboratory Power Supplies
• Insulated Cooler (reservoir)
• LOTS of clamps, valves and hoses

All this comes together as shown in the following diagram:

The trick is getting all this stuff to work together! There were some speed bumps along the way:

• Forma Scientific Heater/Chiller: I checked this unit out and it seemed to work OK. However, it turns out that the chiller is NG – so I had to find another way to cool the water down. Hence, the
• Black Ice Xtreme Rev.2 with Hydor L30 Waterpump. What I ended up doing was to drill a second pair of 1/2″ intakes just for the cooling circuit (the first for the waterblock circuit).

The way I chose to maintain a constant input temp to the waterblock is to raise the water temp in the reservoir to 28.5 C by varying the fan speed on the radiator. It takes about 1/2 hour to stabilize, but once it does, the whole system is at equilibrium; once there, readings are taken.

As an example, I used a BeCooling (thanks Brian!) Slit Edge waterblock as a test subject, with the following results:

What’s nice is that these results are repeatable – multiple runs yielded essentially the same data. There’s a couple of points that are worth noting:

• “Hot Water”: The waterblock’s temp differential is a small number – even at low flow rates, less than 0.5 C.
• Flow Rates: The more flow, the better the performance – no surprise. Note that the Slit Edge is relatively flat at the rates tested.

I’ve elected to keep testing around one gpm and less, as most watercooling systems are running at these rates.

I’m still fine-tuning things, so don’t view this as a definitive test. However, I’m getting close to finalizing the test rig – maybe I’ll pick up a good chiller (don’t really need a heater).

Email Joe