Frequently asked questions (FAQ)

Frequently asked questions

A lot of people ask various questions about Darwin PC builds, about custom PCs or about watercooling in general. Some questions seem to be more common, so here is a collection of answers to the most frequently asked questions.


How much did this build cost?

It can be difficult to give an answer to this question. Often components have been provided by project sponsors. However, if you’d like an idea of how much a Darwin PC project would cost to build, check out the completed builds at


How did you make the custom IO shield for Project Eris?

The IO shield on Project Eris is a custom piece that covers the IO ports on the motherboard. It is constructed from matt black plexi, which was custom laser cut. Solvent weld was used to join the pieces into a box shape. Finally a vinyl cutter to make the red detailing, and sticky pads used to attach to the motherboard. See this post for more details.


What is the purple coolant used in Project Erebus?

The coolant used in Project Erebus is a custom purple pastel from Mayhems. Samples of the paint and the cable sleeving were given to Mayhems so they could match the colour. Mayhems kindly custom made the coolant, as the project was sponsored by them. The coolant was made using the purple pastel coolant, with a little red and blue dyes added. If you try it yourself, be sure to add the dye slowly (drop-wise)!


Why not put components in this order within a watercooling loop?

This is something that is oft-repeated, and on the face of it may seem logical. However, order of components makes very little difference to temperature of the coolant at any point in the loop, and therefore to individual component temperatures. The rate of flow in a watercooling loop is so high, that each part of the coolant has very little time to pick up little heat at the heat sources (CPU and GPU) and likewise to lose heat at the radiators. Overall this means there is at most a few °C difference at any point in the loop. In terms of numbers, a 300W GPU would raise the temperature of a litre of water by about 4.5ºC. If a typical loop has a flow rate of 4 litres/min, the coolant would increase by about 1.125ºC after the GPU.

The only important part of loop layout is that the reservoir is immediately before the pump. This is to remove air before coolant reaches the pump, preventing damage from the pump running dry.


Why construct a serial watercooling loop and not parallel loops?

It is a common misconception that parallel loops perform better than a single, serial loop. refer to flow rate. Compared to a parallel loop, a serial loop allows the total radiator space to dissipate heat more efficiently, rather than for example having one radiator dissipating 300 watts from a GPU, while another dissipates only 75 watts from a CPU. A serial loop also removes the need for multiple pumps and reservoirs, and allows for a cleaner look, with less tubing.


Won’t a larger reservoir allow better cooling?

It may seem logical to assume a larger reservoir can ‘hold more heat’ and reduce temperatures of the coolant and the components. However, reservoir size (and total coolant volume) only affects how quickly the coolant responds to fluctuations and reaches a stable working temperature – not the actual temperatures.

Consider a watercooling loop as a closed system. If the heat conducted by the coolant increases (e.g. when the system goes from idle to full load), the coolant will heat up until it reaches a stable temperature, when the thermal energy it conducts is equal to the heat dissipation at the radiators. The system is now in equilibrium. Once at a stable temperature, there are only two factors affecting the coolant temperature: the rate of heat conduction away from heat-producing components and the rate of heat dissipation at the radiators. The volume of coolant is now irrelevant, as its total thermal energy remains constant.

The only way the reservoir – and therefore total coolant – volume could affect coolant temperature would be increased heat dissipation at the reservoir itself. However, any heat lost through the reservoir surface is minimal compared to the rate of heat dissipation at the radiators. The real purpose of the reservoir is to collect air bubbles before they reach the pump, as if the pump runs dry it could be damaged. It also aids in the initial filling of and bleeding air from the loop.


Why have two pumps in serial?

Having two pumps in serial gives a higher pressure in the loop, which is good for moving coolant around a large loop. Also, with two pumps in series, if one pump did fail, it would have redundancy so the loop would continue to flow. Having pumps in parallel would increase the flow rate, rather than the pressure, but with any decent pump, the flow rate is more than adequate already.


Why keep one reservoir sealed while you fill a dual (serial) reservoir loop?

In a series loop, for filling it is simpler to seal the first reservoir by tightening the fill cap. This means only one reservoir needs to be filled as the coolant is moved into the loop by power cycling the pump(s). Once the loop is full of coolant, both fill caps are tightened, and coolant moves through the whole loop including both reservoirs.