Debunking the 90* myth?

[SuperJew62]

Hello there,

Im more of a lurker than a poster and I wanted some input in regards to my current project.

My system has been running on cardboard with Watercooling on my floor for the past few months and I finally ponied up and bought a case. The case I purchased is a Giagabyte Auora 570 I will be mounting a MCR-220 rad on the back. Inside I will have a Laing G5 with a t-line, my question is better put in illustrated form. This pic from performance pc shows what i want to do add 90 degree elbows on the rad to decrease the tubing "loop" outside of the case. Will using 90 degree albows hurt my flow that badly?

Your thoughts and comments would be appreciated thank you.

[Digz]

I've never fully got my head around this Idea that a 90 degree
elbow kills the flow of a loop as it's not like your putting a kink
in thats flattening the tube.

In a waterblock the water has to change direction twice and you
never hear people having a flow tantrum when people add a GPU
block into a loop.

[RickH]

That actually sounds like a pretty good guideline. If your loop and pump have enough head and flow to handle adding another block, you can probably put in a couple elbows for a similar pressure drop and be OK. If your current flow rate is too low to handle another block, don't do it.

[STEvil]

the style of the bends can make large differeces as well. A nice smooth bore will help flow while a rough cut really hits on higher flow setups. Curving the inside of the bend and "elongating" it somewhat will help, too.

EDIT

For the case pictured it would have been better to pull the fittings out of the side of the tank and place a cover plate over the holes then drill in new holes to point vertically down from the tank.

[Marci]

Nothing to debunk - it's a matter of fluid dynamics and physics, fact, NOT a myth. We don't make this stuff up y'know.

Nothing you can do about it in a block. It's an accepted fact that adding another block adds further restriction due to the block's internal design and the fact that every block effectively adds a 180 deg bend into the loop, before considering any internal channels within that block. Why do you think we tell folks not to bother with chipset blocks and HDD blocks? Not only is it another heatsource, but it's another 180 degree bend... it all adds up to gradually reduce performance.

If elbows / 90 deg fittings can be avoided, then do so. If they can't, use em, but an exact identical rig WITHOUT 90 deg fittings would outperform one WITH 90 deg fittings - laws of physics dictate this...

Every 90 deg bend in a system removes an amount of head pressure due to friction. A 90 degree fluid bend has the flow resistance approximately equal to 3 times an identical length and diameter of straight tubing. Each 90 degree bend also adds turbulence...

Due to the nature of PC Construction and the angles of PCBs in relation to motherboards, and to achieve a good thermal transfer within a block, bends are necessary. When not confined within a block, bends are unnecessary. Tubing runs should be as short and straight as possible if one is seeking optimal efficient performance.

If you really want to know the whys and wherefores, go buy this book - http://www.tp410.com/tp410.htm

As a starting point, consider what happens at the molecular level when a pump moves water through a pipe. The pump exerts work on the water to generate flow, and some of this energy will be wasted by the generation of friction. This friction may come in the form of water molecules colliding with one another, or it may be friction of water with the walls of pipe. The consequence of these frictional forces is a drop in pressure in the same direction as the fluid flow. The terms head loss and pressure drop are used interchangeably to describe this loss in pressure across a length of pipe. The Darcy equation defines this mathematically as:

hL = f(L/D)v2/2g

where hL is the velocity head, f is the friction factor (which may be found in tables or determined experimentally), L/D is the ratio of pipe length to internal diameter, v is velocity of flow and g is the acceleration of gravity. This equation is fairly intuitive; for fluid velocity to increase, so to must the head loss. To maintain high water velocity in restrictive pipes, pressure (aka hydraulic head) must be increased.

For fittings and valves, one is no longer simply considering the length of a straight pipe. Instead, bends and restrictions also affect the head loss. The above equation now requires a modification:

hL =K v2/2g

where K is the resistance coefficient, which is defined as the number of velocity heads lost due to the valve or fitting. If desired, one can relate this resistance coefficient to equivalent pipe length by using the Darcy equation:

K = f(L/D)

where f is the friction factor and L and D are the length and diameter of the pipe, in feet.

Since head losses are additive, it is theoretically possible to estimate the total head loss of a cooling loop provided that estimates for friction factors are available for all tubing and fittings and that data exists for the water block and radiator of choice. For our purposes, however, examining the tabulated K coefficient values in Crane 410 should be sufficient.

Source: http://www.procooling.com/index.php?...s&disp=37&pg=1

The pressure drop for turbulent single-phase fluid flow around sharp 90° pipe bends has proven to be difficult to predict owing to the complexity of the flow arising from frictional and separation effects. Existing models accurately predict the frictional effects, but no precise models are available to predict the flow due to separation. It is the purpose of this work to propose a model capable of such prediction. The proposed model is presented and added to an existing model to predict pressure losses over the turbulent Reynolds number range up to 3 × 105. The predicted data is within a spread of + 3 to − 2 per cent of existing experimental data. Future work will validate this model experimentally and computationally

Another book to read on the subject (and there are HUNDREDS) - http://journals.pepublishing.com/con...r1l073415806v/

To put it in more general terms, the average joe will NEVER understand WHY 90 deg fittings are bad and why they impact flow. Just accept it as a fact that they do.

[clokker]

One hardly has to try to "debunk" the effect of 90 deg. fittings (why fight science?) in order to use them.
Simply accept the fact that your flow will be reduced and live with it.

Incorporating said fittings in my last few loops has greatly simplified tube routing and had a negligable effect on temps (admittedly I'm not a bleeding edge overclocker), so the trade off seems acceptable to moi.

I've never seen a HDD cooler that addressed the heat signature of a drive so I don't use them but I disagree about the efficacy of chipset coolers.
On my NF4 chipsets, both the Maze4 and the Swiftech blocks lowered temps dramatically compared to the Jing-ting, Thermalright and Evercool HSFs that I've tried.
Adding the chipset block seemed to have no real effect on the CPU/GPU temps and obviated the necessity for extra airflow over the motherboard, thus lowering noise (increasingly, my reason for watercooling at all).

I guess it depends on the reason one watercools- if wringing every last MHz out of the system is the goal, then 90deg, fittings would be a real no-no.
If the desired result includes a modicum of silence and esthetics are a consideration, then the fittings and the chipset block can be a real boon.

[berknip]

Great post Marci!

[thunderstruck!]

I guess it depends on the reason one watercools- if wringing every last MHz out of the system is the goal, then 90deg, fittings would be a real no-no.
If the desired result includes a modicum of silence and esthetics are a consideration, then the fittings and the chipset block can be a real boon.

How does using elbows reduce noise? :

[clokker]

How does using elbows reduce noise? :

Obviously, they don't.
Using a chipset waterblock can(one less fan in the system).

[MaxxxRacer]

Marci = ownage.... thanks Marc!

I will have to stick that in the guides somewhere if you dont mind.

[Bad213Boy]

great post marci. physics and hydraulics, gotta love those classes in college

[Okda]

Marci = ownage.... thanks Marc!

I will have to stick that in the guides somewhere if you dont mind.

QFT

yes please do Max

[SuperJew62]

Thank you Marci, Clokker and other posters for the info, I wasn't saying it wouldn't decrease flow I was just curious how severe it would be. Reason I want the radiator on the back instead of internally is for the resale value of the case, the last case I put a dual rad hole in I had to give away almost. Basically the loop would be D5-TDX-External Rad-Gpu. Thank you for the recommendations and sage counsel

[voigts]

There are some situations where 90 elbows just make sense when routing tubing. If you have to use them, do, otherwise dont. If you do, the best thing to use are copper long sweep elbows as they really add very little restriction at all due to the long sweep not having an abrupt change in direction. I have one in my setup now going into the GPU block and there really is very little difference between it and using a very sharp tubing bend, except that using the long sweep eliminates in my situation the stress of a tight bend on the block that tubing would make.

[eXa]

if u have to use elbows or bends those ^ would be best!

[kemist]

those custom made voigts, or can you find them at a hardware store?

[voigts]

Most hardware stores aren't going to carry them. I got mine from good old mcmaster (1/2" #5520K603, or 1/2" street elbow #5520K184) and just soldered in a short piece of 1/2" copper into the end to make it a lot easier to get the 1/2" tubing onto. A little black paint and voila!

You can find them elsewhere online also. They are called long sweep or large radius copper elbows.

[Digz]

slightly less flow and killing the flow are two different things.

When a person mentions using a 90 degree elbow you would think
that it will stop the flow completely the way people reply on forums.

[SiGfever]

Those are refrigeration 90's which are "long radius" instead of plumbing which are "short radius". In refrigeration a "long radius" 90 degree elbow is equivalent to approx. five feet of additional piping in restriction.

A "street elbow" as stated is male solder by female solder (M x F).

[Vapor]

What about something like this?

               |       |
               |       |
               |       |
              /       /
            /       /
          /       /
        /       /
______/       /
            /
          /
        /
______/

2 45* angles would be a little better....and still be able to make the turn, would that be any good? Or would one of those refrigeration 90s still work better?

I've run into a few times where having a 90* would have been easier, even with the added restriction it would have been acceptable, but really never wanted to touch a right angle barb...

[STEvil]

part of the problem is the inlet and outlet of the bend having a "step" for the water to hit.

[virtualrain]

Marci = ownage.... thanks Marc!

I will have to stick that in the guides somewhere if you dont mind.

Perhaps we can add Marci's comments on this to this thread and get this thread stickied...

http://www.xtremesystems.org/forums/...d.php?t=119190

[Vapor]

part of the problem is the inlet and outlet of the bend having a "step" for the water to hit.

Yeah, that's why I'm asking about a dual-45*, would that be better than a single 90*?

Okay, here's a thought experiment that hopefully someone can answer....what's the best way to get a 90* change of direction with a given radius? A variable diameter arc? An decreasing/increasing radius arc? Dual rounded 45* bends (giving a straight shot for an inch or two to reduce turbulence)?

[STEvil]

golf-ball textured wide curve

Failing that, an angled "jet" (taper) might do well. Which side of the taper to be shorter is a good question..

[voigts]

part of the problem is the inlet and outlet of the bend having a "step" for the water to hit.

You are going to have that little "step" every place that tubing goes onto a barb no matter what you do, and that little bit isn't going to make a measurable difference at all.

Two 45 degree work fine, although I have grown to use a long sweep elbow in that it provides a smooth yet compact 90 degree transition. Even two 45 degree elbows are not smooth and take up more space.