HWLabs GTX 360 Radiator Review – Data Supplement

Hardware Labs - HWLabs Logo

Testing Methodology/Specification

If you are regular Skinnee Labs readers then you already know we do things a little differently. For those of you who found your way here for the first time, strap yourself in because we go all out on our test methodology and specification. We are not here for publishing fluff pieces; we want the hard data on every product we test even if that means more work. After all, we want to know how well each product performs and where to spend our money as well… we are enthusiasts just like you are. With that train of thought in mind, we have completely revamped our radiator testing to cover more ground, and below you will find all the details of how we go about our hunt for radiator performance.

Pressure Drop/Restriction

When building your loop there should be a list of things that come to mind, flow and pressure should be near the top of that list. Pressure drop is the measurement of inlet pressure minus outlet pressure, or the pressure loss of flow through the radiator.

Test Method

I have the line from my washbasin in the mudroom hooked up to the gate valve controlling flow, which then runs into the King flow meter. The bottom port on the flow meter is the inlet, top port is the outlet. The outlet runs down to the Delrin T which I have Bitspower 1/2″ barbs on for the normal flow, and the negative pressure line connects via an EnzoTech 1/4″ fitting. After the negative pressure T, the component in testing is attached. I always use Bitspower 1/2″ fittings, which keeps everything on a common test platform…well from a fitting perspective anyhow. At the outlet of the component is the positive pressure T fitting, again with Bitspower 1/2″ fittings and an EnzoTech 1/4″ fitting for the pressure line. The tubing the runs back into the washbasin and down the drain.



From our previous radiator test spec, we ran the tests with 1.5GPM flow through the radiator, but this a wide range of data left unknown. Radiator performance will change with flow rate, C/W will decrease with less flow and increase with more flow. However, flow rate will not radically change performance, the C/W or performance curves will be the same… but we still want to know those curves in order to have a better understanding of what we will see in our loops and we do not all have the same flow rate. Our specification changes, but the overall methodology is still aimed at finding the capabilities of the radiator, not what it will do in a given system.

For radiator testing, the best way to conduct the tests is to apply a heat load, just like your CPU, GPU(s) and other components you can put a block on and add to your loop. In order to supply the heat load for testing, I use modified aquarium heaters. Aquarium heaters are available in a variety of different wattages, lucky for us wattage is exactly what we are looking for in heat dissipation results. However, the tricky part with aquarium heaters is circumventing the safety mechanism that shuts off the heater when the set temperature is reached. Modifying the aquarium heaters allows for a constant heat load to be applied to the loop rather than heating the water to a given temperature and shutting off.

Measurements and calculations, what exactly are we going for here? C/W or degrees per watt is the best measurement to determine the capabilities of a radiator. Ultimately, C/W is a calculation of water temperature minus air temperature being pulled or pushed through the radiator divided by the heat load (wattage). This calculation gives you the delta in degrees of the radiator leaving water for each watt of heat applied to the radiator. Confused yet? An easier way to think of C/W is the temperature of the water over ambient or air in temperature for each watt of heat in your loop. Now that we have this equation and results, we can specify a set delta and figure out what heat load the radiator can dissipate at that specified delta. In short, this helps answer the question, can this radiator handle a CPU and GPU and get me decent temperatures. A bit more on C/W, this time in relation to Fan RPM. On several charts later in the test report, you will see Fan RPM charted with C/W. This is to show the effect on radiator performance in using different speed fans, some radiators perform very well using low speed/CFM fans and get better the more CFM you push/pull through the radiator. Where some radiators perform sub-par with low speed/CFM fans and require high speed/CFM fans to effectively dissipate heat from your loop. Yes this has a lot to do with how the radiator was designed, but I feel it is an important piece of data to show as a misconception I had early on in the start of my liquid cooling addiction was low speed/CFM fans could not be used on radiators optimized for high speed/CFM fans. I was quickly shown how wrong my thinking was.

Besides the typical Fan RPM, air temperatures and water temp measurements you will see Air Capacity Used listed in the data table.


Testing is now split into four rounds, a set flow rate with four fan speeds/configurations makes up a ’round’, same heat load applied for each test as well. Four flow rates are used giving us sixteen tests. Each test duration is four hours, with the first two hours being a warm-up period; data from the first two hours is not used in any of the data used. The remaining two hours are the data used in the final calculations and data you see presented. The heat load continues to come from modified aquarium heaters residing in a custom-built one-gallon reservoir. Radiators are tested in a very limited loop where we can reduce other variables and focus the testing on the radiator itself, this loop consists of the custom reservoir, Swiftech MCP655 with EK V2 Top (Front Outlet), two delrin T’s with four temperature probes each, the radiator, Koolance FM17, King 7520 and a brass gate valve. The big changes to the loop from our old test spec are the Koolance FM17 for logging flow rate and doubling the number of water sensors. However, before I detour the test specification, let us get to the tools/parts list…

  • Temperature Monitoring and Logging: CrystalFontz CFA-635 with SCAB attachment – Used to log 20 digital thermal sensors (12 air, 8 water), three Fan RPM signals and flow rate from a Koolance INS-FM17 at 1 second intervals.
  • Thermal Sensors: Dallas DS18B20 Digital one-wire sensors – .5C absolute accuracy overall with a .2C mean error between 20-30C.
  • Test Bench Sensors Deployed:
    • Two (2) air in sensor per fan
    • Two (2) air out sensors per fan
    • Four (4) water in sensors
    • Four (4) water out sensors
  • Flow Rate: King Instruments 7520 0-5GPM, 10″ Scale – Accuracy 2% of scale. Flow Rate controlled by a brass gate valve with 1/2″ NPT 5/8″ Barbs
  • Flow Rate Monitoring: Koolance INS-FM17 – Calibrated to King Instruments 7520
  • Water: Minnesota Tap
  • Fans:
    • Scythe Gentle Typhoon 1850 (AP-15) 120mm x 25mm – 47CFM/1350RPM/28dB
      • Low Speed Fans – Undervolted to 800 RPM
      • Low-Medium Speed Fans – Undervolted to 1200 RPM
      • Medium Speed Fans – Undervolted to 1700 RPM
    • Scythe Gentle Typhoon 4250 (AP30) 120mm x 25mm – 116.5CFM/4250RPM/44dBA
      • Medium-High Speed Fans – Undervolted to 2300 RPM
      • High Speed Fans – Undervolted to 3000 RPM
      • Ultra High Speed Fans – Undervolted to 3800 RPM
  • Fan Voltage: Mastech HY3005D (0-30V, 0-5A) Variable DC Power Supply
  • Data Logging: Each temperature sensor and fan RPM channel is logged for 240 minutes, first 120 being the warm-up, remaining 120 is the data we’re after. We log 14400 samples for each sensor, fan channel and flow rate through the CFA-635.
    • Air Temperature Data: Two sensors are deployed over each fan location, on the intake and exhaust side of the radiator, representing Air In and Air Out accordingly. Per the standard, each sensor is logged at 1 second intervals.
    • Water Temperature Data: Four sensors are crammed into a Bitspower 1/4" fitting and then into a DangerDen/Thermochill Delrin T, the T’s are placed at radiator inlet and outlet. Again, per the standard, each sensor is logged at 1 second intervals.
    • Fan RPM: Fan RPM is monitored via the Crystal Fontz CFA-635 and logged at the same interval as all temperature sensors. Power is controlled via a seperate DC power supply (noted above).
    • Flow Rate: Flow rate through the loop is controlled by a brass gate valve and King 7520 flow meter, the Koolance FM17 is in the loop for logging purporses and provide a hands off test cycle (no manual logging).
    • Heat Load: Wattage for the aquarium heaters are still plugged into a KILL-A-WATT P4320, but the wattage is calculated using radiator delta and logged flow rate. This eliminates the minor fluctuation of wattage from my old manual method and also includes pump heat dump from the MCP655.
  • Test Lab Environment: Unfortunately, I do not have an environment test chamber. All tests are performed in 10×13 room in my basement which is temperature controlled via a wall thermostat and on a separate zone from the rest of the house. I am able to maintain a consistent room temperature this way.

If you were familiar with the V1 radiator testing, you have probably surmised the differences for V2. Overall, the major changes are fan choice and subsequently RPM, but we have some actual airflow figures to introduce in the collective comparison. The biggest change overall is flow rates, with the addition of the Koolance FM17 and running four flow rates versus one previously we get a greater spectrum of data to use for estimating system performance. I do not know if we found the holy grail of radiator testing, but this is definitely a step forward from the old method.

Enough jabbering, time to get to the results…

Pages: 1 2 3

Leave a Reply

You must be logged in to post a comment.