Analysis of Plumbing Loops and Other Configurations

[mas985]

I, like most people, have long believed that plumbing loops can help balance return flow through multiple returns in a pool or spa but I was curious as to how much they really help and if there are other ways to balance return flow. So I performed a detailed analysis on 6 different plumbing configurations and I thought I would share the results.

Configuration Diagrams



Configuration Descriptions

- Multi-Line - Separate lines with valves to each of the returns going from the pad to the pool.

- Full Binary - This configuration is completely symmetrical and the plumbing to each return is identical thereby guaranteeing that each return will have the same flow rate.

- Semi-Binary - This configuration uses a single split combined with a series fed configuration and is partially symmetrical. Returns 1 & 4 will have the same flow rates and returns 2 & 3 will have the same flow rates.

- Symmetrical Loop - This configuration is similar to the Semi-Binary in that the same return pairs will have the same flow rates although the values are swapped. However, because there is extra pipe to travel through, this configuration will have slightly higher head loss.

- Offset Loop - This is probably a more common loop configuration and will tend to have more imbalanced returns than the symmetrical loop.

- 2"/2.5" Series - This is probably the most common configuration for a pool and of course has the most imbalance between the returns. But as I will show, this can be improved by increasing the feed pipe.

Pool Analysis Configuration

40' Pool Length
10' Distance between returns
2' 1.5" return pipe into pool
1" Eyeballs
2" Feed pipe unless otherwise specified
80 GPM Total flow rate split between the returns
Head loss calculated to edge of pool for all scenarios

Analysis Results

Multi-Line (Valves 100%) - Return 1-4 Flow Rates = 20.6, 20.2, 19.8, 19.5 - Head Loss = 1.7'

Multi-Line (Valves Balanced) - Return 1-4 Flow Rates = 20, 20, 20, 20 - Head Loss = 1.8'

Full Binary - Return 1-4 Flow Rates = 20, 20, 20, 20 - Head Loss = 5.6'

Semi Binary - Return 1-4 Flow Rates = 19.7, 20.3, 20.3, 19.7 - Head Loss = 4.7'

Symmetrical Loop - Return 1-4 Flow Rates = 20.3, 19.7, 19.7, 20.3 - Head Loss = 5.8'

Offset Loop - Return 1-4 Flow Rates = 21.4, 20.8, 18.7, 19.1 - Head Loss = 2.5'

Series (2" pipe) - Return 1-4 Flow Rates = 24, 20.1, 18.2, 17.7 - Head Loss = 3.5'

Series (2.5" pipe) - Return 1-4 Flow Rates = 21.9, 20, 19.1, 18.9 - Head Loss = 2.4'

Series (3" pipe) - Return 1-4 Flow Rates = 20.8, 20, 19.6, 19.5 - Head Loss = 1.9'

Observations

- The Multi-Line configuration gives the most flexibility and performance but requires a lot of piping. But keep in mind that 4x2" pipes is the equivalent of 1x3.7" pipe so not exactly apples to apples.

- The Full Binary configuration has ideal return balance but it also has the highest head loss of all the configurations.

- Semi Binary is fairly balanced but still has fairly high head loss.

- Symmetrical loop has the same balance as Semi Binary but at the cost of more head loss.

The first three scenarios all suffer from the long feed pipe to the center of the pool which adds a significant amount of head loss. However, if the equipment is located closer to the middle of the pool, these configurations might be a better choice.

- Offset loop has worse balance than the symmetrical but better than the 2" Series.

- The 2" Series has the worst balance between returns.

- The 2.5" series has similar balance to the offset loop and slightly lower head loss.

But in reality, none of the configurations are that imbalanced and probably would never be an issue. So in most cases, this is probably nothing to really be concerned about unless you are feeding a water feature that requires exactly the same look at all the returns.

In addition, I looked at a 8 jet spa using the Semi Binary configuration with 4 jets in each series after the split. Assuming 2.5" pipe is used, there is so little head loss in the 2.5" pipe between the jets compared to the head loss in the jet itself, that the flow rates out of the jets are nearly identical even without a loop. So this shows that loops are not really required for spas either.

Comments/questions from the forum are always welcomed.

 

[Leak Specialist]

The more plumbing underground the more chance of a plumbing leak underground.

The absolute best solution in my mind is every return has its own individual line and valve.

The end result is infinite adjustability and the option of abandoning the leaking line without any concrete cutting.

 

[Drewskii]

Thanks for the analysis. I think this kind of study is extremely helpful in understanding plumbing designs.

In the series design, or others that are imbalanced, can the difference in flow rates be optimized by changing the size of the branches inversely to the distance they are away from the feed pipe?

For instance:

Return 1 branch - 1.5" pipe
Return 1 branch - 2" pipe
Return 1 branch - 2.5" pipe
Return 1 branch - 3" pipe

 

[Leak Specialist]

Changing the size of the hole in eyeball fitting would be cheaper, starting small then getting larger as you get further from the pump. Also they would be adjustable after the pool was plumbed and decked.

 

[Drewskii]

That makes sense for returns.

Suppose we were referring to a static exit like a bubbler?

 

[Leak Specialist]

I have not seen different size bubbler nozzles, but true bubbler systems generally have their own booster pump.

If your bubblers are standard wall returns with venturi's you could still change the size of the hole and get similar flow.

 

[mas985]

The more plumbing underground the more chance of a plumbing leak underground.

The absolute best solution in my mind is every return has its own individual line and valve.

The end result is infinite adjustability and the option of abandoning the leaking line without any concrete cutting.

If the leak is between the valve and the pool, it would still leak, just not as fast so the eyeball would need to be plugged as well.

I would agree that running multiple lines is a good solution although it may not always be possible or practical.

But not to leave out the mutli-line method. If 1.5" pipe where used for each line and all valves are 100% open, the flow rate distributions would be the following:

Return 1-4 Flow Rates = 21.6,20.4,19.4,18.6 - Head Loss = 2.0'

However, if you adjust the valves for all the lines to be 20 GPM, then it becomes:

Return 1-4 Flow Rates = 20 GPM - Head Loss = 2.3'

Not much head loss needs to be added to make all the lines equal flow rate.

 

[mas985]

In the series design, or others that are imbalanced, can the difference in flow rates be optimized by changing the size of the branches inversely to the distance they are away from the feed pipe?

For instance:

Return 1 branch - 1.5" pipe
Return 1 branch - 2" pipe
Return 1 branch - 2.5" pipe
Return 1 branch - 3" pipe

Changing the pipe size does not work in the series design because you want all of the lines to be at the same pressure and the only way to do that is with the largest pipe possible to minimize the head loss between the return ports.

However, you could use this method for parallel runs but the difference in head loss is quite significant. So for example, a 1.5" pipe has 3.5x the head loss as a 2" pipe so the 2" pipe would need to be 3.5x longer to have the same head loss. Even a 2" pipe has about 2.4x the head loss as a 2.5" pipe so the difference in lengths is quite significant. It would be easier to just pick a pipe size and make the equivalent lengths the same.

 

[Leak Specialist]

Yes the line would have to be plugged in the pool also.

There are a few designs in the op that are just as impractical as running a separate line to each return, they just don't leave an option of abandonment.

I'm not an engineer nor a math wizz but simplicity is largely overlooked and vastly underrated in the pool industry.

 

[mas985]

Which is why many PBs opt for the series fed returns because it is the easiest to install and requires the least amount of pipe. Many won't even entertain the idea of multiple returns or will charge a fortune to do it.

 

[Leak Specialist]

Which is why many PBs opt for the series fed returns because it is the easiest to install and requires the least amount of pipe. Many won't even entertain the idea of multiple returns or will charge a fortune to do it.

That's the truth. Plumbing in series leaves a lot to be desired when it comes to adjusting the output of the jets and pump efficientcy, it also puts several of the weakest plumbing fitting there is under your pool deck, the Tee.

It must be noted that I look at this much differently than most, I'm not trying to get my pump as efficient as possible nor equal flow between returns, I'm trying to make it as painless as possible when your pool starts leaking.

 

[Drewskii]

I wanted to create a visual way to interpret Mark's study of piping systems detailed earlier in this thread. There is a natural conflict between flow and head loss (friction) and depending on your goal, a perfectly balanced flow, minimum head loss, or somewhere in between, it is mind numbing if you aren't an engineer. I hope the chart helps simplify a complex set of choices. Perhaps a PB could help by quantifying the cost to build each design.

Mark had nothing to do with this chart and if it is inaccurate the blame is solely mine.

If a piping system has four ports then a perfect distribution of the flow would be 25% to each port.

This chart shows the % variance from 25%, the optimum flow, at each port depending on which system is used.

HL-# is the Head Loss (friction) for each piping system and is also indicated to the right of each curve or line.

Multi-Line valve balanced and Full Binary designs both have perfectly balanced flow rates of 25% at each port so their lines are at 0% and represent the Benchmark for other systems.

I'm hoping a similar study is coming that studies a 3 and 5 port system and maybe even round systems. :-D

 

[mas985]

Technically, the flow rates out of each port are discrete points and not a continuous curve so it would be more accurate to show them as points or on a bar chart.

As for other # of ports, for odd numbers, some of the configurations would produce uneven flows such as the full binary. That really only works for 2^n number of ports. Even the semi-binary/symmetrical loop would have a larger unbalance because there is not an even split.

But in general, the order of performance is still the same as for the 4 port. There is just less of a difference between the best and worst ports for a 3 port configuration and more of a difference for a 5 port configuration. The same effect is still there so nothing new would be learned from a separate analysis.

But if you are asking about your specific scenario which is slightly different than a pool, then here are the results for that. Assuming:

2.5" feed pipe
3 returns
3/4" Bubbler
Series fed

Ports 1-3 Flow Rates = 29.84, 30.06, 30.12 - Head Loss = 11.5'

The error in the calculation due to the installer is much higher than the difference in flow rate. Installers are not all that precise and generally don't have perfect glue lines.

 

[Duffatola]

Multi-Line (Valves 100%) - Return 1-4 Flow Rates = 20.6, 20.2, 19.8, 19.5 - Head Loss = 1.7'

Multi-Line (Valves Balanced) - Return 1-4 Flow Rates = 20, 20, 20, 20 - Head Loss = 1.8'

Assuming valves 100% means 100% open can you please advise what valves balanced means?

 

[mas985]

Valves balanced is where the flow rate through each of the line is set to the same value (e.g. all 20 GPM). So shorter lines will have the valves restricting the flow more than the longer lines. The longest line would be fully open.

But I wouldn't bother equalizing flow to the pool outlets. There really is no benefit to it unless there is a significant problem with the plumbing setup.