[Mycelus]

Title says it all. I heard that sharp turns and whatnot affect cooling performance... doesn't water just push through? It makes no sense that bends would have a huge effect on the temps.

Someone enlighten me?

 

[100cotton]

They restrict the water flow more than a straight shot, and having too slow of water can result in worse temps.

 

[Trogdor]

90* fittings do cause flow restrictions, it's been awhile so I can't tell you the exact amount.

A few elbows (1-4) aren't going to kill flow. If you need a lot just get another pump.

 

[Mycelus]

Ok but has this been tested and shown to make a huge hit in performance? Water flows slower at some points, but I can't see that actually increasing temps by say 3-5+ C?

What about people who have bent pipes, theirs would also face the same problem no?

 

[Trogdor]

AHha! Found what I was looking for! Read this - http://martinsliquidlab.org/2011/01/30/fittings-and-elbow-impacts/

 

[Mycelus]

Ok, so basically minor, and negligible...

 

[Artikbot]

Quote:

Originally Posted by Mycelus

Ok, so basically minor, and negligible...

Apparently so, but my experience conflicts with Martin's.

If I put a 90deg straight out of the pump or straight into it, it becomes SUPER loud. Performance is similar, but the noise is unbearable (compared to straight output/nput).

 

[Mycelus]

Urrr.. how would a 90 degree bend cause noise..

 

[Artikbot]

Quote:

Originally Posted by Mycelus

Urrr.. how would a 90 degree bend cause noise..

I have no idea, but as I said, in my loop, if I put a 90 degree going into our out of the pump, it does.

 

[Martinm210]

My test does not include pump close proximity effects. I only tested restriction losses. I have done some pumpnelbow testing, but it likely depends on the pump. In industrial pump system design practice is to avoid elbows at the pump inlet side of the pump within 4x the pipe diameter. Reason for this is the elbow can cause enough negative pressure to induce cavitation as well as unbalancing the fimpeller flow. Outlet is generally less of a concern.

I have measure the benefits of removing elbows at the inlet side such as going from a factory DDC top to aftermarket top with inlet elbow removed etc.

Anyhow, I would generally avoid elbows in close proximity to the pump if possible, my test was purley based on pressure drop or restriction loss.

 

[Mycelus]

Yeah the ONLY restriction I could see is if the 90 degree is right near the pump, at other places maybe it'd slow the water down and reduce the pressure, but I would imagine not enough to actually drop temps or cause problems.

Im assuming, I dont actually know.

 

[nleksan]

I have a 90deg single-rotary adapter and rotary male-male coupler (all Bitspower) attaching my MCP35X2 to the bay res, and I havenever once had an issue... In ffact, I can push close to 3GPM through my loop at max, and I have a crapload of rotary angled fittings (not to mention a fairly high Restriction CPU block). Even with 2 of the most powerful common pumps @ full speed, I get zero cavitation...

 

[Mycelus]

Alright, in that case this is all bullocks

 

[DiGiCiDAL]

Yep... it's really pretty simple if you just look at flow being a characteristic limited by the single most restrictive part in your loop. I you have magical blocks that somehow managed to be less restrictive than a 90-degree fitting... then in that loop, adding 90-degree fittings might reduce your flow (albeit still by only a very tiny margin - and definitely not noticeable). In reality the CPU block and possibly the GPU block(s) are so much more restrictive that you can reasonably add a dozen elbows and not make any difference at all.

 

[ZytheEKS]

Or if you have one of the Watercool Mo-Ra radiators that have 27 180degree bends XD
Those cause a hell of a lot of restriction

 

[DiGiCiDAL]

Quote:

Originally Posted by ZytheEKS

Or if you have one of the Watercool Mo-Ra radiators that have 27 180degree bends XD
Those cause a hell of a lot of restriction

That's actually a good point... in many cases your radiator alone will have tighter turns (using possibly much smaller pipe/tubes) and yet it's not a concern for almost anyone. Naturally, since you have to have one or you're not going to be cooling much of anything!

I guess the best rule of thumb would simply be only use 90-degree fittings where you actually need them.

I think most people would do this instinctively however - it's not like I've seen anyone say "hmmm... I could run this tube perfectly straight from point A to B... but instead I think I'll make it zig-zag across my case a few times first"!

 

[Mycelus]

Yeah.. I mean Im used to bogus crap being thrown around, so for a while I was wondering if 90 degree bends have a huge effect.

I really hate wannabe noobs who spread around this crap...

It didnt click in my head at all, thought maybe I was missing some technical details about pump rates and whatnot

 

[DiGiCiDAL]

LOL... there's way less in watercooling circles than there is in audiophile circles. At least I've never heard anyone claim that they got a faster overclock by placing rocks on their case, or by using cryogenically treated SATA cables made of unobtanium.

Thermodynamics and hydrodynamics are a little hard to grasp for many (myself included to a large extent) - but in general, and largely thanks to people like Martin - you don't have to rely on speculation and theory... because someone has tried just about everything in pursuit of performance.

On topic... if the 90-degree-bends theory is bugging you... try some of the threads arguing about loop order - they can get really entertaining. (Coming from someone that has had their CPU block last in the loop to great success).

 

[SynchronicBoost]

This is probably the last thing in the world to worry about. But for your reference in fluid power terms, two 45's are better the a 90 for pressure drop. But honestly, do you guys even think there is much pressurization in there to impact pump efficiency? And this liquid is a relative incompressible, and not air which heats as it compresses.

 

[ZytheEKS]

Quote:

Originally Posted by DiGiCiDAL

That's actually a good point... in many cases your radiator alone will have tighter turns (using possibly much smaller pipe/tubes) and yet it's not a concern for almost anyone. Naturally, since you have to have one or you're not going to be cooling much of anything!

I guess the best rule of thumb would simply be only use 90-degree fittings where you actually need them.

I think most people would do this instinctively however - it's not like I've seen anyone say "hmmm... I could run this tube perfectly straight from point A to B... but instead I think I'll make it zig-zag across my case a few times first"!

Actually most radiators use a dual pass U flow system, so half of the tubes are all running in parallel. The combined area of the chambers running in parallel is usually quite a bit larger than the tubing you are using, and the chambers themselves have no bends or curves, but rather the chambers at the end create the U flow. This means most radiators will add little to no restriction to a loop

THEN there's the radiators that use tubes with 180 degree turns instead of chambers and charge 3x more money , because **** logic, that's why!