Quote Originally Posted by bluehaze View Post
I think i've got it now seeing what Boxgod posted got me thinking, if there is a parallel loop setup where you Y out of the pump then Y back into the res being that water is incompressible you will see no gain. However if you Y out of the pump and your res has dual return ports the res can act as an expansion chamber of sorts. So in effect due to the restriction in the loop the water doesn't need to compress, if you put a Y on the outlet of the pump the water will by nature fill both sides of the Y due to the restriction of each loop and the water taking the least path of resistance.

The least path of restistance will always be the loop that is not full, the problem comes in when you add a Y going back into the res you no longer have any room for expansion therefore you will never see any gain as the volume of water flowing through the 2 loops going back through the Y remains the same because you cannot fit more water through that Y going back into the res.

However if you have two lines going back into the res then you can effectively have 2x the volume of water flowing through the two loops combined when the pressures equalizes and both loops are full.

I can see it in my head but it's really hard to explain what i'm thinking lol not sure if any of this makes sense
I get what you're saying...but that's not the case.

If you took a loop with a CPU block, a radiator, and other stuff, it would have a pressure drop curve of y= AX^3 + BX^2 + CX + D (well, D=0). If you doubled the components in serial, you would have a pressure drop curve of y= 2AX^3 + 2BX^2 + 2CX + 2D. Basically, you'd double the resistance....

But if you took those same components and split the (doubled up) serial loop into two parallel loops....the overall restriction is much lower than even AX^3 + BX^2 + CX + D (how much so depends on the Y split and all associated with that).

But because the restriction drop is so large from going from the doubled-up serial loop to the parallel loop, flowrate through the pump increases. But overall gains in flow are not necessarily imminent. Going in parallel will only increase the flowrate outside of the parallel'd region (by drastically reducing the restriction the parallel'd components create)....as for what happens within the parallel'd components when using a Y-split, the verdict is still out (there should be some cases where it's still beneficial, but not on the scale of what the T3 does).

The reason why the T3 succeeds at doing this is because while another pumptop would require you to add components (and therefore restriction) in order to go to parallel, the T3 is a native parallel design that requires you to add components (and therefore restriction) to go to serial.

EDIT: BoxGods' PEC explanation is the true way to look at it