Is this water loop done the right way?

[For Science!]

Just now, Techea said:

Yes, the guy in the video explains clearly what's happening with the cpu temp when you decrease or increase pump speed. Exactly what I saying myself about the gpu temp that is directly related to the velocity of the fluid in the system. If you also increase the number of loops and add more radiators to the system maintaining the same velocity then the temperatures automatically go down. 

Yes, but did you not also see that going from "very fast" to "normal" speeds only increased the temperatures by 0.4 degrees? that's nothing, margin of error for most people's measurement devices. It is not directly proportional at all, you can really drop the flow rate down to very slow speeds before you start seeing the effects of bad flow. 

 

It is even entirely possible that because there is lower resistance with parallel flow, you might even see better overall flow than serial. And if you took such a bad flow rate and applied it to a serial flow, it will also be bad since the flow rate will be limiting at those very low speeds.

 

This is the simple concept of bottlenecking (as much as it usually used incorrectly). For the purposes of cooling, system flow rate is not a bottleneck, and thus you can reduce the flow rate by half, and there is very-little-to-no penalty on the cooling performance, simply because other things (such as transfer efficiency between the die to the IHS, or IHS to the waterblock) are more limiting. You can artificially impose a bottleneck in flow rate by reducing the pump rpm, but as you see, it will negatively affect both parallel and serial loops. If you reduce the pump flow rate where it impacts the performance of the cooling, it will do so under all configurations.

 

I am getting a bit tired of talking to you about this, so you should just go ahead and do your testing. I don't know where your confidence comes from, perhaps Dunning-Kruger effect.

[Techea]

man, it could be very true.