Reservoirs vs T-lines in Flow Rate = Insignificant

[ranker]

Just one elbow won't amount to much, just don't go putting a dozen of them in to straighten out your tubing and you'll be fine.

I'm really curious on the effect of using a T-joint as an elbow. What I'm implying is that I'd like to connect a T directly to the radiator. The other end of the T will be capped off with a G1/4 temp probe. The branching T itself will be connected with tubing to continue on to the blocks.

I'm wondering if that sort of implementation of an "elbow" will result in a larger pressure drop.

[Martinm210]

I'm really curious on the effect of using a T-joint as an elbow. What I'm implying is that I'd like to connect a T directly to the radiator. The other end of the T will be capped off with a G1/4 temp probe. The branching T itself will be connected with tubing to continue on to the blocks.

I'm wondering if that sort of implementation of an "elbow" will result in a larger pressure drop.

The only thing I could relate that to without having done the test myself is some generalized engineering equivalencies to fitting losses. A more typical fitting loss pipe length equivalent is something like this:

Fitting vs Equivalent Pipe length
Elbow = 1.5'
T flow through = 1.0'
T branch = 3.8'

Unfortunately because most fittings have a smaller ID than hose fittings we actually experience higher pressure drops than those typical published PVC pipe type length equivalencies. For example I found one 1/2" Elbow is worth about 4' of 1/2" tubing....that's a 1/2" elbow meant for 1/2" hose that actually has an ID of about 3/8". If an elbow is worth 4' of tubing a T branch "could" be worth twice that much.

Now that sounds pretty bad....BUT tubing itself is not very restrictive so that doesn't tell the whole story.

If I have a more typical system that's getting 1.53 GPM for example and add 8' of tubing to it, I'd only reduce the flow rate down to about 1.45 which is some, but in terms of temperture not enough to worry about.

If possible though for absolute best efficiency, putting that probe on the 90 side of that T so flow passes straight through would be even better.

[ranker]

If possible though for absolute best efficiency, putting that probe on the 90 side of that T so flow passes straight through would be even better.

I would if I could, but I can't find any other way to connect my RD-30 to my radiator AND include a G1/4 water temp probe at the same time. My case is so cramped and tight it's the only solution I can come up with. It's why I'm asking for the theoretical impact now to see if its even worth using the temp probe.

[safan80]

personally I'd go with a res. the EK res doesn't need a funnel which is nice.

[Maviryk]

Try hard mounting the Fuzion, no springs at all.

Just be sure to keep track of the turns and use a criss cross pattern for tightening.

[bobo5195]

2 points immediately spring to mind (would right more but have to catch a bus)
1. when modelling anything in watercooling the restriction tends to be of the form a*x^2. This is the case with your coke bottle rad above. This applies to everything and if your modelling something always compare your complicated formula to an x^2 graph unless there is a huge discrepancy. The x^2 result will be more “right” and any difference is measuring error. This might not apply for really complicated flow scenarios like a jet impingement block or a pump but should give you a very close answer. T

2. A coke bottle is not a good radiator as the water just rushes in and out, without settling, which is bad reservoir design. This has two effects
a. Air does not filter out as easily as the water never settles (moves really slowly) so air can’t escape.
b. There is lower restriction as the inlet and outlet are coupled together in a flow pattern. This reduces restriction but means the flow is not settling.

Restriction of a good reservoir with the same sized pipes going in and out will always be equal to (for SI units)

P = V^2 *rho

Where:
P= the pressure drop
V = velocity of water in m/s in the pipes going in and out.
Rho = the density of water 1 kg/ litre (don’t start putting stupid numbers of significant figures in)

[Martinm210]

I would if I could, but I can't find any other way to connect my RD-30 to my radiator AND include a G1/4 water temp probe at the same time. My case is so cramped and tight it's the only solution I can come up with. It's why I'm asking for the theoretical impact now to see if its even worth using the temp probe.

Ahhh, I see. Yeah it won't amount to any real temperature change. Having a water temp probe is a really good radiator performance indicator. I have one down my T-line in both of my loops. It's not hard mounted, but it gives me an after by blocks measurement and shows me if I get any heat buildup.

[Martinm210]

2 points immediately spring to mind (would right more but have to catch a bus)
1. when modelling anything in watercooling the restriction tends to be of the form a*x^2. This is the case with your coke bottle rad above. This applies to everything and if your modelling something always compare your complicated formula to an x^2 graph unless there is a huge discrepancy. The x^2 result will be more “right” and any difference is measuring error. This might not apply for really complicated flow scenarios like a jet impingement block or a pump but should give you a very close answer. T

2. A coke bottle is not a good radiator as the water just rushes in and out, without settling, which is bad reservoir design. This has two effects
a. Air does not filter out as easily as the water never settles (moves really slowly) so air can’t escape.
b. There is lower restriction as the inlet and outlet are coupled together in a flow pattern. This reduces restriction but means the flow is not settling.

Restriction of a good reservoir with the same sized pipes going in and out will always be equal to (for SI units)

P = V^2 *rho

Where:
P= the pressure drop
V = velocity of water in m/s in the pipes going in and out.
Rho = the density of water 1 kg/ litre (don’t start putting stupid numbers of significant figures in)

But how do you know when you've gone from a simple to a complex restriction?
Even a T line is going to have multiple losses going on, one at the step down in ID, one at the opening, and one at the step-up. The reservoir is similar, step down loss, entrance to the reservoir velocity loss, entrance loss into the sharp barb fitting, and another loss as it steps up into the hose ID. I found I needed to go clear to the 4th to plot my pump curve of the D5. It threw me off when I collected the data and saw a dip and then this big buldge, but I guess the hydraulics and friction losses combined with changing impeller efficiency is a whole different ball of wax. Found the curves at Laing's site and they match what I recorded. I've also seen some strange flatten of flow rate in the fuzion and some other blocks so by default I've been going to a third order poly though I have seen what you noted going too high is often just plotting error. But there are a number of losses that aren't well captured by the second power, and figuring out when you need to make that change isn't easy.

One thing I've learned through all my testing is alot of engineering equations, while the theory is good, can be off slightly because of water coolings specifics. I applied Darcy's equation to the friction loss of tubing only to find out it was slightly less restrictive than my actual test confirmed both by curve calculations and real world pump tests. Also fitting losses that are commonly published are off and I think that is because hose fitting IDs don't don't match the hose. It's all still something that can be captured mathematically, but sometimes it seems to be just slightly different in environment.

The reservoirs would all vary slightly and you would have been suprised how well this coke bottle worked in terms of still water, I was impressed

The inlet and outlet of my test case were protuding with inlet K value losses around .9. I bet I could reduce alot of that loss by simply cutting the outlet barb flush and rounding the edge which should bring that K value down to .2. A bay reservoir where the threads are nearly hidden and flush with the walls probably flows better as well..?

[virtualrain]

I would if I could, but I can't find any other way to connect my RD-30 to my radiator AND include a G1/4 water temp probe at the same time. My case is so cramped and tight it's the only solution I can come up with. It's why I'm asking for the theoretical impact now to see if its even worth using the temp probe.

Ranker, if you use a DD Delrin T (which is very compact) then I would think that with the temp probe plugged into one of the pass-through ports, it would effectively turn the T into an elbow... the temp plug would protrude enough into the T to block that branch completely. I would think this would be better than a more extended T where the coolant can travel down the dead-end path.



What do you think?

[Eddie3dfx]

Well I figured I'd post since today I took out some components and installed a T line.
After the 8800 the temps went up..
35-38c idle (e6300 to 3.34gz)
8800gts 42c
Swiftech resevoir>D5 pump>apogee gt>MC60>mcw30>mcr220>thermaltake alum resevoir>swiftech resevoir>

After getting yelled at on here I decided to go the route of the T line and dump the alum in the process.

Now it goes
D5>mcr220>Tline>apogeegt/Mc60>mcw30>D5
The T line installed 2" after the mcr220 hole.
So i took out 2 resevoirs and some tubing.. numbers are now
Cpu 31-33c idle
Gpu 40-42c idle

I'll add in redline later and see if it makes a difference.
So definite thumbs up for the T line, though I did change a few other things
up as well

[sick_g4m3r]

Well I figured I'd post since today I took out some components and installed a T line.
After the 8800 the temps went up..
35-38c idle (e6300 to 3.34gz)
8800gts 42c
Swiftech resevoir>D5 pump>apogee gt>MC60>mcw30>mcr220>thermaltake alum resevoir>swiftech resevoir>

After getting yelled at on here I decided to go the route of the T line and dump the alum in the process.

Now it goes
D5>mcr220>Tline>apogeegt/Mc60>mcw30>D5
The T line installed 2" after the mcr220 hole.
So i took out 2 resevoirs and some tubing.. numbers are now
Cpu 31-33c idle
Gpu 40-42c idle

I'll add in redline later and see if it makes a difference.
So definite thumbs up for the T line, though I did change a few other things
up as well

how hard was bleeding?

@virtualtrain, can you try to rephrase that. Are you saying turn that T into an elbow with a temp sensor?

[Eddie3dfx]

it still is lol..
It's been about 45mins so far, I'm not sure how long it's going to take to get every bubble out.

[trans am]

Well I figured I'd post since today I took out some components and installed a T line.
After the 8800 the temps went up..
35-38c idle (e6300 to 3.34gz)
8800gts 42c
Swiftech resevoir>D5 pump>apogee gt>MC60>mcw30>mcr220>thermaltake alum resevoir>swiftech resevoir>

After getting yelled at on here I decided to go the route of the T line and dump the alum in the process.

Now it goes
D5>mcr220>Tline>apogeegt/Mc60>mcw30>D5
The T line installed 2" after the mcr220 hole.
So i took out 2 resevoirs and some tubing.. numbers are now
Cpu 31-33c idle
Gpu 40-42c idle

I'll add in redline later and see if it makes a difference.
So definite thumbs up for the T line, though I did change a few other things
up as well

we have a similar setup. I am adding this mcw60 (8800gts) to my loop this weekend and I was concerned with my temps on a dual radiator like yours. At the same time I am going to remove my 3.5 res and going with a t-line and DD fillport. thanks for posting your temps.

[virtualrain]

how hard was bleeding?

@virtualtrain, can you try to rephrase that. Are you saying turn that T into an elbow with a temp sensor?

Sort of... Ranker needs an elbow connection out of his rad... he also wants to use a temp probe in the loop and he was concerned that using a T with a dead-end on the pass-through route would hurt performance worse than just using an elbow. I also tend to think a dead-end on a T would be worse than a normal elbow.

However, if you use a compact T like the one I pictured above and plug the opposite pass through opening with a temp probe, you effectively have an elbow and it should perform like any other elbow... not like a T with a dead-end path.

[Eddie3dfx]

Transam, not a problem.
Let us know what your temps are when you install it.
I'm thinking of adding in an other mcr220 or maybe a single radiator (black ice x-flow 120 at microcenter for $29.99 with barbs)
If I get a Q6600 then I definitely need a mcr220.

On a side note, how do you guys plug your tubing on the T line?
home depot plug or fillport?

[sick_g4m3r]

Sort of... Ranker needs an elbow connection out of his rad... he also wants to use a temp probe in the loop and he was concerned that using a T with a dead-end on the pass-through route would hurt performance worse than just using an elbow. I also tend to think a dead-end on a T would be worse than a normal elbow.

However, if you use a compact T like the one I pictured above and plug the opposite pass through opening with a temp probe, you effectively have an elbow and it should perform like any other elbow... not like a T with a dead-end path.

yeah, an elbow and temp probe in one. that sounds like a great way to do it.

@ eddie, fillport FTW. safer and easier

[ranker]

Sort of... Ranker needs an elbow connection out of his rad... he also wants to use a temp probe in the loop and he was concerned that using a T with a dead-end on the pass-through route would hurt performance worse than just using an elbow. I also tend to think a dead-end on a T would be worse than a normal elbow.

However, if you use a compact T like the one I pictured above and plug the opposite pass through opening with a temp probe, you effectively have an elbow and it should perform like any other elbow... not like a T with a dead-end path.

Thanks for the suggestion and I do have those DD blocks. Unfortunately, the temp probe seems to push pretty deep inside. I guess the dilemma is which is worse, a partially blocked elbow or a dead end T/elbow. I'd test but then again, I'm really getting sick and tired of building/rebuilding my Horizon. o.0 I just want to find a way to be happy about it and then get started on my Bob Slay project. After 3 personal UFO builds, I'm anxious to get away from those cases.

But anyhow, enough of me whining, I'll just sit back and see if I can find another place to throw the temp probe in and go with this route as a last resort. I can't believe I talked myself into doubling my water temp probes to 6 to measure water in and water out of the radiator.

[sick_g4m3r]

Thanks for the suggestion and I do have those DD blocks. Unfortunately, the temp probe seems to push pretty deep inside. I guess the dilemma is which is worse, a partially blocked elbow or a dead end T/elbow. I'd test but then again, I'm really getting sick and tired of building/rebuilding my Horizon. o.0 I just want to find a way to be happy about it and then get started on my Bob Slay project. After 3 personal UFO builds, I'm anxious to get away from those cases.

But anyhow, enough of me whining, I'll just sit back and see if I can find another place to throw the temp probe in and go with this route as a last resort. I can't believe I talked myself into doubling my water temp probes to 6 to measure water in and water out of the radiator.

just stick one in the res and be done with it. lol

[trans am]

Transam, not a problem.
Let us know what your temps are when you install it.
I'm thinking of adding in an other mcr220 or maybe a single radiator (black ice x-flow 120 at microcenter for $29.99 with barbs)
If I get a Q6600 then I definitely need a mcr220.

On a side note, how do you guys plug your tubing on the T line?
home depot plug or fillport?

Eddie3dfx, can you run a 3dm2006 bench or game and hit alt tab while you have some temp monitors running? I am curious what your load temps are while overclocked. Also what fans do you have on your rad and how many? I'm going to use 4x120mm medium speed yate 79cfm
(push/pull) on this heatercore.

[virtualrain]

Thanks for the suggestion and I do have those DD blocks. Unfortunately, the temp probe seems to push pretty deep inside. I guess the dilemma is which is worse, a partially blocked elbow or a dead end T/elbow. I'd test but then again, I'm really getting sick and tired of building/rebuilding my Horizon. o.0 I just want to find a way to be happy about it and then get started on my Bob Slay project. After 3 personal UFO builds, I'm anxious to get away from those cases.

But anyhow, enough of me whining, I'll just sit back and see if I can find another place to throw the temp probe in and go with this route as a last resort. I can't believe I talked myself into doubling my water temp probes to 6 to measure water in and water out of the radiator.

LOL! I hear ya! The last niggly details seem to take forever to resolve and just drag the build on and on and on... I know!

If you can put the T along the route somewhere else, it might be better... or tap a G1/4 opening in something where the probe would be benign. It would be nice if RADS came with an extra G1/4 opening in one of the tanks to put a probe in (or a plug if you didn't want to use it for that).

[Eddie3dfx]

8800gts xfx xtreme
mc60 block
oc
670/981
shader 1568 (i believe)

idle 41c
3dmark 06 - alt-tab 45c
ati tools scan for artifacts 44-45c

4 fans (sandwhich 2 front - 2 back)

front
1 silvestone 110cfm
1 aerocool xtreme turbine 89cfm

Rear
2 thermaltake cfm

fronts are pulling, rears are pushing

[trans am]

8800gts xfx xtreme
mc60 block
oc
670/981
shader 1568 (i believe)

idle 41c
3dmark 06 - alt-tab 45c
ati tools scan for artifacts 44-45c

4 fans (sandwhich 2 front - 2 back)

front
1 silvestone 110cfm
1 aerocool xtreme turbine 89cfm

Rear
2 thermaltake cfm

fronts are pulling, rears are pushing

wow! looks pretty good also considering you have 3 blocks and your gpu is clocked pretty high. very good temps.