How do T's work

Snoopybiggs

If you have a 3/4 T on a boiler, how does the water split? I thought you'd need a pressure differential to cause the water to flow. 

Robert O'Brien

Everything that enters, leaves. Which branch it takes is dependent on differential between the branches

JohnNY

What Robert said. Is there more to this question?

Snoopybiggs

what does that mean "differential between the branches". Im still learning so Im trying to use Dan's advice of becoming a marble in the pipe.

EBEBRATT-Ed

Water goes to the path of least resistance. In the case of a tee you have two connections on the "run" of the tee and one connection to the "branch" of the tee. If water comes in one of the run connections it is easier for it to flow out the other run rather than turning the corner and going out the branch.

If the water comes into the branch connection then the flow could go either way...the have the same resistance. This is called "bull heading a tee" which is sometimes bad practice in some systems.

Lets say you have water pressure connected to one connection on the run of the tee. On the other two connections you have valves installed. If you open the branch valve wire and open the run valve wied water will flow from both but the branch will get a little less flow (90 degree turn).

If you closed the run valve part way with the branch valve open you could balance the flow because the valve is adding resistance.

Snoopybiggs

So water going into one side of the run, will split between the branch and other side of the run equally with no pressure change assuming we're pumping away from the compression tank. So me being the marble would go straight or turn 90% exactly 50% of the time?

Snoopybiggs

The reason this is so hard for me is that diverter T's make so much sense to me but the easy stuff I find challenging

delcrossv

Say the branch has 1/3 more resistance than the run. So two-thirds of the flow will go down the run 1/3 will go down the branch. Making the turn itself adds resistance.

Snoopybiggs

got it thanks

EBEBRATT-Ed

@delcrossv

U said it better than I tried to do.

Alan (California Radiant) Forbes

Say the branch has 1/3 more resistance than the run. So two-thirds of the flow will go down the run 1/3 will go down the branch. Making the turn itself adds resistance.

I disagree as I think all the water would continue through the run and nothing would go through the branch. Any resistance in the branch greater than the run would deter any water from being sidetracked. That’s why they came up with diverter tees, no?  And even with diverter tees, we’ve seen here how difficult it can be to bleed diverter tee systems. 

JUGHNE

A tee with a single reducing run size such as 3/4 inlet X 1/2 outlet on the run and 3/4 being the side branch would act just like a diverter tee.
Sme flow thru the branch because of resistance on the 1/2 run.

neilc

is this all hypothetical with either loop being equal length?
wouldn't overall loop(s) length and head matter ?

Snoopybiggs

Can someone ask Dan? He has a good way of explaining things. i think i get it but without seeing it in a 3D model i wont fully get it. im too visual

Jamie Hall

Well, I'm not Dan, but... consider. The resistance to flow in a circuit of pipe is related to the amount of flow. The more rapid the flow, the greater the resistance (more else varies as the square of velocity). So... Think about two loops of pipe, starting with a single pipe, then splitting at a T, the coming back together again after a journey at another T, and proceeding on from there. As it might be two loops in a hot water system.

Now at the first T, the pressure in both branches (whether leg or run) must be the same. Further, at the second T, the pressure in both branches, whether leg or run, must be the same. Therefore the resistance of the two loops must be exactly the same at whatever flow is in the two loops. The flow will divide between the two loops so that this condition is met. If one loop has less resistance than the other, it will have more flow. Both loops will always have some flow.

Monoflo Ts work this way -- but they aren't just a plain vanilla piece of pipe. Instead, the inside is cast in such a way that the resistance to flow on the branch is significantly less than the resistance on the run, forcing more flow through the branch to balance the overall resistance.

Snoopybiggs

Does the temperature of the return water make a difference on the return side?

Jamie Hall

Snoopybiggs said:

Does the temperature of the return water make a difference on the return side?

To a first approximation, no. If you're going to get really precise about it -- out to the 4th significant digit fussy, then yes, but even there the difference is slight. But for any purpose we are going to hit, no.

delcrossv

Alan (California Radiant) Forbes said:

Say the branch has 1/3 more resistance than the run. So two-thirds of the flow will go down the run 1/3 will go down the branch. Making the turn itself adds resistance.

I disagree as I think all the water would continue through the run and nothing would go through the branch. Any resistance in the branch greater than the run would deter any water from being sidetracked. That’s why they came up with diverter tees, no?  And even with diverter tees, we’ve seen here how difficult it can be to bleed diverter tee systems. 

Just stick a hose to the end of a tee, it'll come out the branch and the run. The difference in a system is that you have pressure at the outlet pushing back on the branch as well. Diverters throttle the run to give a differential to the branch. So Yes and No depending on back pressure on the run and the branch.

Snoopybiggs

I made what engineer's call a gross error. pumping away page 23. "In a closed hydronic system, there's no need to lift the water because the system;s already pressurized". The whole dammed thing is filled with water all the time. To make things completely fool proof is to underestimate a complete fool.

EdTheHeaterMan

Snoopybiggs said:

Can someone ask Dan? He has a good way of explaining things. i think i get it but without seeing it in a 3D model i wont fully get it. im too visual

Maybe this will help



The top piping system A shows a section of piping with a tee branch that is later returning to a tee branch. The path of least resistance is straight thru. over 90% of the flow will go straight. Some will take the detour, but most will not.

In B the path is equal resistance because both paths have the same numbers of branch tees and elbows to make the resistance to flow equal. 50% will turn and 50% will go straight.

In the bottom C illustration I have added 60 feet of piping in a loose coil as to not add any significant resistance. The length of pipe is as if it were straight (use your imagination) so the equivalent length of the piping is equal to the equivalent length of the branches and elbows. So both branches will have about the same flow.

Does this help at all?

Disclaimer:To all you expert pipefitters out there. these fittings were custom made by me for this illustration. I measured the flow and there were the exact numbers I found after vigorous testing by an independent laboratory that specializes in design drawings on my computer. Therefor I did not need to consult an equivalent length chart for my calculations.

Jamie Hall

I might add to all this... where things get really interesting is when you get to networks of pipes, rather than simple loops. Such as might be found in a municipal water system. Trying to figure out how much water goes through which pipe is a good exercise...

yesimon

Say the branch has 1/3 more resistance than the run. So two-thirds of the flow will go down the run 1/3 will go down the branch. Making the turn itself adds resistance.

I disagree as I think all the water would continue through the run and nothing would go through the branch. Any resistance in the branch greater than the run would deter any water from being sidetracked.

Neither is true, for a 3/4" tee the straight run resistance is approx equivalent to 2.4 ft of straight pipe while the branch is approx 5.3 ft of straight pipe. The difference is not substantial. Also, flow is not linearly inversely proportional to the length/resistance of each branch. A branch with double the resistance will actually have more than 1/3 of the total flow, more than 50% of the flow of the branch with less resistance.

Regardless, most heating circuits in NA are very overpumped and there will be substantial flow even in a much longer loop. For fin-tube, 1gpm of flow only decreases heat emitter output by 5% compared to 4gpm of flow.

Rich_49

yesimon said:

Say the branch has 1/3 more resistance than the run. So two-thirds of the flow will go down the run 1/3 will go down the branch. Making the turn itself adds resistance.

I disagree as I think all the water would continue through the run and nothing would go through the branch. Any resistance in the branch greater than the run would deter any water from being sidetracked.

Neither is true, for a 3/4" tee the straight run resistance is approx equivalent to 2.4 ft of straight pipe while the branch is approx 5.3 ft of straight pipe. The difference is not substantial. Also, flow is not linearly inversely proportional to the length/resistance of each branch. A branch with double the resistance will actually have more than 1/3 of the total flow, more than 50% of the flow of the branch with less resistance.

Regardless, most heating circuits in NA are very overpumped and there will be substantial flow even in a much longer loop. For fin-tube, 1gpm of flow only decreases heat emitter output by 5% compared to 4gpm of flow.

Please attempt to give folks who want to learn good information . Problem with what you wrote above is that the question is not about parallel circuits but a tee . Parallel circuits would be same as all circuits being branches and it is ENTIRELY different .

In this case your highlighted phrase / caption is not accurate .

https://www.pmengineer.com/articles/84197-determining-flow-rates-in-parallel-piping-systems-constructed-of-smooth-tubing

Jamie Hall

As I mentioned above -- sort of by implication -- things can get very complex very quickly when dealing with multipole flow paths for a liquid. Or, conversely, one can use a "for all practical purposes" approach. The practical purpose being defined by the problem at hand.

The pressure drop in a given, specific piping arrangement is proportional to the flow raised to the 1.83 power and directly proportional to the hydraulic length of the piping arrangement, all else being equal. The "all else" being holding the geometry of the piping constant. The hydraulic length of the piping arrangement is the physical length, of course, plus appropriate individual additions for the various fittings. The physical length is self explanatory. The individual additions are not strictly constant with velocity and hence flow, but for Reynolds numbers likely to be encountered in common applications they are close enough. Many handbooks tabulate them for various fitting geometries and sizes.

It is also useful to remember that for a flow configuration involving a branch, such as a T, the pressure in all legs of the branch at the centre of it must be the same. Thus if you have two paths for flow -- an almost trivial case in hydraulic engineering, by the way, the flow between the two branches must divide in such a way that the total head loss in each branch is exactly the same.

The "resistance" to which @Rich_49 refers is what is termed above the equivalent hydraulic length. If we suppose that we have two pipes in parallel flow, one of which has double the hydraulic length of the other, the flow will divide so that the pressure loss through each branch will be equal. Since one length is twice the other, the flow in shorter one must be just slightly less than 1.46 times the flow in the longer one or 59% of the total flow, and the longer one 41%.

I note that whether the arrangements which allow the flow to divide and rejoin are quite irrelevant -- Ts, Ys, whatever, connected any way you like -- provided that the correct additions are used for the geomety of the flow. There, are, for instance, several for flow dividing at a T...

I will leave it to the interested student, as I used to say so many times, to figure out how those numbers come to be.