right angle before a reservoir - OK?

[babysnake]

no reason i can't have a right angle connector at the input of a pump is there ?? :confused:

i can't think of one, but i thought i'd just ask those in the know ! ^_^

[fivecheebs]

Idealy you should probably avoid it. Most pumps like a fairly unrestricted inlet. Having said that it probably wont make much difference if your using a small bore loop. (i cant remember what your getting)

[dutchcedar]

If there's any place in a liquid cooling loop you want to keep as restriction-free as possible, its at the pump's inlet. You must feed the pump with an unimpeded source of coolant...

So yes, there is a reason not to use a 90 degree connector at the pump inlet. It might work just fine, but I would suggest avoiding it if its at all possible. <_<

[Knipex]

What pump are you using ???

[Starbuck3733T]

I've got a 90 on the inlet of my eheim 1250, and the temps didn't change at all from when I didn't have the 90 there.

[babysnake]

the pump is an alphacool one - the ap700 12v and i'm using small bore (8mm/6mm) tubing mostly, althought the inlet tube is slightly larger (10mm/8mm)

i can possibly avoid the inelt right angle but i'll end up with a very small radius curve there instead, as it's all pretty squished in !! :o)

[Knipex]

If those are your options then I would preffer the 90 degree bend over a really tight bend on the hose.

[babysnake]

If there's any place in a liquid cooling loop you want to keep as restriction-free as possible, its at the pump's inlet. You must feed the pump with an unimpeded source of coolant...
my understanding of the physics is that through a circular pipe, flow is inversely related to the 4th power of the radius of the pipe (ie half the radius equates to 1/16th the flow)

i assume a manufactured, metal right angle fitting is a constant radius, hence flow should be unimpeded.....even though at the gut feeling level that doesn't seem right somehow :confused:

[Pug]

my understanding of the physics is that through a circular pipe, flow is inversely related to the 4th power of the radius of the pipe (ie half the radius equates to 1/16th the flow)

i assume a manufactured, metal right angle fitting is a constant radius, hence flow should be unimpeded.....even though at the gut feeling level that doesn't seem right somehow :confused:

With a low flow from the pump, the flowrate isn't hit quite as hard as with a bigger pump (like the AP900 or 1500).
I've not checked to see if Alphacool recommend against this procedure (as I think Swiftech do with some of their pumps, not sure).

I use an Eheim 1046 with a right-angle on the inlet due to the fact that I have it mounted in a 5.25" bay.
Upgrading this configuration to a heftier pump like a 1048 or 1250 won't give me the benefit that it might if I had straight connectors, less blocks and upped the tubing from 8/6mm to 10/8mm...
...but for what I'm cooling/using, it definitely works.


What sort of bend radius are you seeing if you do go with a straight connector? I have Polyurethane Polyether which can go pretty tight at a good range of temps.


Tip If you do/did try a piece of hose to see if it would kink, make sure you run warm water through it with the pump running, as the shore hardness of the pipe walls can vary with temperature change and allow collapse and subsequent starvation of the pump feed.
Just a heads-up. :)

[dutchcedar]

my understanding of the physics is that through a circular pipe, flow is inversely related to the 4th power of the radius of the pipe (ie half the radius equates to 1/16th the flow)

i assume a manufactured, metal right angle fitting is a constant radius, hence flow should be unimpeded.....even though at the gut feeling level that doesn't seem right somehowI'll leave the physics to those better versed in it. Even from the few posts here, you can see that some people have done this successfully, but always with a clear and evident need to do so.

The bottom line, to me at least, is simple... don't do it unless you have to. If you have to, be prepared for some possible complications. Personally, I would never do it at all. <_<

[babysnake]

thanks for all the input ^_^

the reason i was thinking of doing this is so i can then squeeze the reservoir right next to the pump itself, so saving some space and making mounting the reservoir easier (it's one of those straightforward cylinder ones with a screw fitting on top, and the 1/4" threads at 12 and 6 o'clock, as it were, with no obvious way of fixing it all to the case - next project gets a BullsEye or Aquatube for sure B) )

one alternative would be to take a big loop of 8/6 tube - which i worry might spoil some of the 'look' of the case - or do a '3 sides of a rectangle' turn (straight/90/90/straight) to get round the corner

i'm into cooling, of course :o), but mostly i'm after a quiet, good looking rig, as the stock CPU/GPU speeds play my games fast enought as it is

clearly some people have done this sort of thing OK and there's no specific Alphacool info against it that i can find, so i might mitigate the problem a bit by having a short run of 1/2" tube from the reservoir to the pump (as suggested here) and see how i get on

thanks again for the incredibly useful input. as a noob it's very reassuring to be able to discuss all these things without fear of being made to feel st00pid !! :)

[fivecheebs]

The only time you get made to feel stupid here is when you do it to yourself!

Remember there are no such things as stupid questions...

Let us know how it pans out babysnake ;)

[babysnake]

Let us know how it pans out babysnake ;)
will do........


[ BTW: wife & kids dispatched to a cousins this weekend, so 48+hrs of unrestricted modding beckons :D ]

HURRAH!!!!

[Fibbles]

What would be more ideal; a right angle on the pump's outlet leading to a straight fitting on a block or a straight fitting on the pump leading to a right angle on the water block (2 of them actually).

The pump is an Eheim 1046 with the head of a 1048 (Inno's HPPS) and the blocks are either an angled Graph-O-Matic with straight fittings or a straight Graph-O with bent/angled fittings.

Seeing as the loop will be on the low flow side, the bent/angled fittings won't really hurt the flow too much. The CPU block is Cape's X2 Pro and it doesn't require (or even improve with) lots of flow.

[BigBen2k]

It doesn't matter, as long as you avoid putting an elbow on the pump inlet.

[Da_Rude_Baboon]

Thats good as there will be a right angle connection on the return feed to my res. Theres no way i can avoid it.

[Fibbles]

http://wizdforums.co.uk/images/smilies/thumb.gif The pump has a filltank with an angle before the tank, so it wont matter and I will have an angle in front of the tank.

Doesn't Star have an angle before his pump inlet? With or w/out did nnothing to temps if I'm right, but thats temps http://wizdforums.co.uk/images/smilies/huh.gif

[Starbuck3733T]

I used to... now I've got that nice fat 3/4" ID tube feeding the pump. IIRC it did not much to temps but hurt the flow... this was when I was running the maze4 block (not the storm g5) that was not as dependent on high flow rate.

[jaguarking11]

ill be running 3 elbos on my pump. Ill let you guys know how it will run. But tell you this though. A machined elbow will slow down the flow but is less likely to impede it. My buddys machine has a 9degree elbow on the inlet of his mag drive pump. It does not make a single degree difference. Personaly just use a larger id elbow on the inlet and you should overcome the problem with starving the pump.