Hello, I wan't to watercool my Bitfenix Prodigy, and I have room for either a dual 120MM rad up front such as an Thermochill TA 120.2 or I could get a Phobya Xtreme 200mm Radiator, what would give me the most cooling? Thanks
When it comes to figuring out how much radiator you need for your specific loop, you have to start doing some math. I know that we all have been building a loop and thought, ‘how many, what size and what kind of radiators do I need for this loop to stay cool like I want?’
First tip: Google is your best friend to help find TDP (Thermal Design Power}
Finding out what the TDP or your CPU or GPU is can be as simple as doing some searches by searching for ‘i7 2600k TDP', ‘GTX 580 TDP’, or ‘AMD 6970 TDP’. Remember to account for all components…if you run a multi-card graphics setup, you need to include the TDP values for all cards in the total. For example, our i7 2600k has a stock TDP of about 95 watts at 100% load (estimated). If we have a 2x SLI setup of GTX 580’s, we are looking at about 244 watts at 100% load, per card. Total? About 583 watts in heat that these three components can potentially produce when at 100% load, simultaneously; it's also safe to consider that heat dissipation can never be 100% efficient of power consumption, so even calcuating 85-90% of your TDP total is pretty safe. (This also translates very closely to wattage when you need to consider a power supply for your system, but you need to account for the remaining components: motherboard, fans, hard drives, DVD drives, etc. To help calculate a full system TDP, you can use a tool like the Extreme PSU Calculator (link). In short, when calculating loop TDP, simply add up the total values for components being cooled in the loop...if you have more than one video card, make sure you add in TDP for each one. If you want to simply calculate the overclocked TDP wattage of your CPU, just adjust the CPU section of the calculator or utilize the calculation listed a bit later.
Once you have calculated your total loop TDP potential, you need to consider radiators that dissipate heat in watts depending on flow rate of your loop and fans being used and their speeds/power. For this task, I almost always refer to Skinneelabs.com/radiators (link) for all of this crucial information, graphs and comparisons.
For example, I am going to reference the XSPC RX360 radiator for this loop. Given the total TDP of 583 watts, I want to know if this single radiator is enough for my loop, or if I should consider another radiator.
Looking at this chart, we can see that the maximum amount of heat this radiator can dissipate is around 555 watts using 2800 rpm fans (very fast, very loud). You could get better results in a push/pull scenario, but that’s even louder; you may be able to live with a 15-20° delta and loud fans if you went this route.
In short, Delta-T is the load temperature of the water in your loop when compared to ambient air temps; if your room ambient is 27°C, and load water temp is 34°C, this gives you an approximate Delta of 7°C if you are running 100% load on all components being cooled by the loop. Basically, delta-T is a mathematical derivative of your ambient room temperature, flow rate, heat to be dissipated (in watts) and the ability of your radiator to dissipate heat (in watts) depending on fans used to produce the cooling impact by the loop as a whole. You’ll notice the chart above has a listing of different fans in the upper-left corner: this determines the angle of the graph and the temperature delta on the left side of the graph. Lower fan speeds correlate to a higher delta-T as you add more heat in watts to the loop. The more heat you produce, the more important it is to remove it from the loop; and fans help accomplish this goal. If you notice the actual temperatures on the lines of the graph at the determined points (around 300 watts of load and around 555 watts), you’ll see that the fan speed allows the heat dissipation to be rather normalized. However, the further to the right (and up the graph you go), you’ll also notice that your delta-T rises. Below a 5° is incredibly good, 10° is still very good and even 15° deltas are very much the norm. If we wanted to run this loop at a 10° delta, we would need to run two of these RX360 radiators to keep the heat load in watts below 300 watts dissipated per radiator with fans of 600-2800 rpm (in a single-fan setup; push/pull would allow some leniency here…perhaps a RX360 and an RX240, instead).
Granted, TDP and determining our delta-T isn’t an exact science, but it gets us pretty close. It’s a bit more tedius to calculate CPU overclocked wattage; however, here is a great calculation to help CPU overclocking and estimated TDP:
For this example I will use a relatively average overclock voltage of 1.35v to reach 4.5ghz (4500mhz)
OC Wattage = TDP x ( OC MHz / Stock MHz) x ( OC Vcore / Stock Vcore )^2
OC Wattage = 95 x (4500/3400) x (1.35/1.25)^2
OC Wattage = 95 x (1.3235) x (1.08)^2
OC Wattage = 95 x 1.3235 x 1.1664
OC Wattage = 147 (which is exactly what was calculated by the PSU calculator for overclocked CPU watts on this chip)
Radiators & Fans
The radiator is the heat exchanger for your water loop; water passes into its thin channels which run parallel down and back with small fins to help dissipate the heat. They are typically rectangular and match fan sizes commonly for 120mm and 140mm fans, but there are others to match 200mm fans sizes as well. Most radiators used are the 2x, 3x or 4x of these 120-140mm versions, but there are large radiators that also use 4, 6 or 9x 120mm fan-size in a grid pattern for a very large rad. There are also 180mm and 200mm rad sizes out there for several different fan placements and mounts for newer cases with larger footprint fans.
Radiators are typically listed and classified with FPI or 'Fins Per Inch'; this means that for every 1", there are 'X amount' of heat dissipating fins. Common low FPI rads are 7-11 FPI, while high FPI models are 20-30 FPI. This is important to understand as it directly relates to the radiator's performance (more FPI = higher cooling potential), but take note: this also means higher CFM fans with very good static pressure to move air over the densely packed fins. Higher CFM and static pressure fans are often more costly than lower speed fans that can be used for lower FPI rads. While the FPI-to-expensive-loud-fast-fan concept is a good rule of thumb to maximize performance of a 30 FPI rad, there isn't anything that says you have to run these kinds of fans on them, as normal, mid-range fans also perform quite well despite the extra FPI restriction. Expand the image below for an example of low vs. high FPI.
(Images from SkinneeLabs.com)
Again, Skinneelabs.com has a very good radiator comparison and benchmark; triples are the most commonly implemented radiators, but they are also mainly used to keep an apples-to-apples comparison among the tested radiators. Once you start talking radiators, you start talking Deltas; or the difference in temperatures in comparison to ambient room temperatures and the water inside your loop. Most folks run a modest 10-12°C delta or even slightly more. Once you get below 10-8°C of Delta, you are getting your water temps closer to ambient. While I'm not a thermodynamic fluid engineer, there are some fairly easy-to-read explanations on Deltas around the web and you likely have already read my short description earlier in the sticky.
I've put together a chart that defines some different cooling properties of common radiators and their cooling potential based on total volume in cubic millimeters (mm^3). The list below is ranked based on thermal coefficient; essentially a product of a radiator's heat in watts for a 10°C delta-T with 2000rpm fans divided by total radiator volume to achieve the average cooling potential of all 15 radiators listed and reviewed by skinneelabs.com/water-cooling-radiators.
This can also be used for a very quick cooling performance estimate for total volume of a radiator based on the average thermal coefficient:
[volume LxWxH in mm] (LxWxH) x 0.00023129193 = Watts dissipated for 10°C delta-T (estimated)
When it comes to deciding on fans to use on your radiators, there are several sources around the web with comprehensive testing; including from one of our own, 4ryan6, who has a great fan guide. Be sure to read through the links below for questions and comparisons on many fan manufacturers.
Okay, I am still stuck on the calculations. I don't understand where I would use the FPI, I am comparing 2 rads, the 200MM has an FPI of 8, but the 240MM has an FPI of 20, I don't understand if I should or where I should use this in the calculation
The FPI in the calculation is an average, but that's across 15+ different radiator models used for the source data in the equation. First I would determine the total radiator volume - In theory, if both rads are the exact same LxWxH, higher FPI typically has a higher potential to cool better with medium speed or faster fans due to the increased fin/tube surface area. Determine what volume each radiator has first, and go from there.