Help me understand this setup

[Gnomer]

One of our propane water heaters is leaking. I'm not replacing it myself and have someone setup to come out. But, I'm trying to understand the set up. I drew a diagram of how they are installed below. I believe these are considered to be in parallel but I don't understand the circulation pump on the cold side with two cold supplies and cutoffs.

I was hoping that I could isolate the damaged heater and stop flow but there are no cutoffs except for the cold side. Is there any drawback to installing cutoffs so each can be isolated or was this not done on purpose?

Any thoughts? I'm just curious about it and annoyed there's no way to shut down the broken tank and still use the functioning one.

Thanks!!

 

[hj]

I asssume tank #1 is the one that failed, because I cannot see how #2 is doing anything

 

[Sponsor]

 

[wwhitney]

Given the circulation pump, presumably the right hand "cold supply" is really a hot water return, for a hot water circulation loop.

The tanks are plumbed in parallel, so they both should be contributing to the hot water demand. For a true 50/50 split of demand between the two tanks, the piping to and from the two tanks should be identical, from the common cold supply point to the common hot distribution point. I'm unclear on how sensitive the division of demand is to asymmetries there, and so how much less of the hot water Tank 2 will be supplying because it has somewhat longer cold and hot piping.

As to the question, I would think having separate shutoffs would be prudent and normal practice, and I'm not aware of any reason not to do that. But that answer is from first principles rather than experience with such installations.

Cheers, Wayne

 

[wwhitney]

I asssume tank #1 is the one that failed, because I cannot see how #2 is doing anything

How is that? Tank #2's output is in parallel with Tank #1's. Are you saying the extra length shown on the piping to Tank #2 (if to scale) is enough to skew the division far from 50/50?

Cheers, Wayne

 

[John Gayewski]

How is that? Tank #2's output is in parallel with Tank #1's. Are you saying the extra length shown on the piping to Tank #2 (if to scale) is enough to skew the division far from 50/50?

Cheers, Wayne

Yes that's what he's saying. Tank 2 isn't doing anything. The cold in and hot out are on the same tank why would tank 2 supply anything?

 

[John Gayewski]

Normally the cold only will have a valve so your not accidentally ruining the tank with a valve on the hot. This situation calls for a bypass setup to run on one tank at a time if need be.

 

[wwhitney]

Yes that's what he's saying. Tank 2 isn't doing anything. The cold in and hot out are on the same tank why would tank 2 supply anything?

I'm not following. Both Tank 1 and Tank 2 are pressurized by the city water pressure on their cold water inlet. Both tanks have hot water outlets, and the two outlets are combined in a tee to form a single pipe that supplies the house with hot water. Why wouldn't water from both tanks flow into that tee to supply the house?

Cheers, Wayne

 

[John Gayewski]

I'm not following. Both Tank 1 and Tank 2 are pressurized by the city water pressure on their cold water inlet. Both tanks have hot water outlets, and the two outlets are combined in a tee to form a single pipe that supplies the house with hot water. Why wouldn't water from both tanks flow into that tee to supply the house?

Cheers, Wayne

There's water in the second tank of course, but it's not leaving the second tank as the path of least resistance is in, and then back out of, the first tank. Why would water take a path backwards toward the second tank? High pressure goes to low, in a straight a line as possible.

 

[wwhitney]

There's water in the second tank of course, but it's not leaving the second tank as the path of least resistance is in, and then back out of, the first tank. Why would water take a path backwards toward the second tank? High pressure goes to low, in a straight a line as possible.

Yes, water moves from high pressure to low pressure, but it takes all available paths. Just like electricity. I redrew the OP's diagram below to better illustrate the logical arrangement of the piping.

There are two paths from the cold tee labeled A to the hot tee labeled B. Water flow will take both paths, as shown by the arrows. If Path 1 through Tank 1 and Path 2 through Tank 2 have identical resistances to flow, then the flow through the two paths will be identical. So best practice would be to make everything from tee A to the two tanks exactly symmetrical, and likewise from the two tanks to tee B, and that should give you approximately equal flow.

If, as the drawing appears to show, Path 2 is longer than Path 1, then Path 2 will have more resistance than Path 1, and you'll get more flow along Path 1. During flow, there will some pressure value at the inlet to tee A and a lower pressure at the outlet of tee B. That pressure difference is what is driving flow through both Paths, and both flows will have identical pressure drops.

So say Path 2 has a 50% greater effective length than Path 1, then its pressure drop for a given flow will be 50% greater. Pressure drop doesn't vary linearly with flow, it's approximately quadratic (double the flow gives about 4 times the pressure drop). So the flow through Path 2 would be approximately sqrt(2/3) = 80% of the flow through Path 1.

Cheers, Wayne

 

[John Gayewski]

I understand your what your saying but I think your missing something. I know in hydronics that when there's an imbalance there's no flow and therefore no heat given. This example there is a huge imbalance. The path of tank one has close to 0 resistance to flow since there's no no piping. It literally flows in the heater and back out of the heater. The other heater never sees the pressure drop from opening a faucet. It could be the differential being much greater in plumbing rather than that of a hydronic system your theory is true.

It would be interesting to see a live demonstration one tank filled with red dye and the other filled with blue. If purple comes from the faucet then there's at least some coming from tank 2.

 

[John Gayewski]

I actually lived in a house with water heaters piped like this once. We had 80 gallons of 125 degree water in the basement and yet when the upstairs neighbor took a shower I could not get through a shower without the water turning cold.

When I changed the piping that problem ceased.

 

[wwhitney]

I know in hydronics that when there's an imbalance there's no flow and therefore no heat given.

I'm not so familiar with hydronics (just did my first project), but I assume that where there's a small imbalance, there's a difference in flow and thus less heat given--it's not all or nothing.

This example there is a huge imbalance.

I'm not sure why you say that. First, I'm not convinced that the OP intended the diagram to be to scale; maybe the installation is entirely symmetric, but drawn with longer pipes to Tank 2 simply to make it fit easier in the diagram.

But even if the drawing is to scale, I marked it up below to show the imbalance. Below the black line is the section where there are two parallel paths. And the green piping represents the extra piping to Tank 2; if you deleted that piping, it would be balanced.

[Exactly so on the hot side; on the cold side, I assumed that the cold tee has Tank 2 on the straight path, and Tank 1 on the side outlet, and that the equivalent length of "straight path of tee plus 90" is less than the equivalent length of "side branch on tee", so the non-green piping after the tee represents that difference in equivalent lengths.]

So maybe worst case the piping to Tank 2 is twice as long? It depends on what equivalent length, if any, should be assigned to the tank itself. I already did the math for the case that the Tank 2 path is 50% longer; if it's twice as long, then the flow through Tank 2 should be sqrt(1/2) = 70% of the flow through Tank 1.

Cheers, Wayne

 

[Jeff H Young]

of course either tank can leak. Kind of hard to tell exactly without a more detailed drawing

 

[John Gayewski]

I'm not so familiar with hydronics (just did my first project), but I assume that where there's a small imbalance, there's a difference in flow and thus less heat given--it's not all or nothing.


I'm not sure why you say that. First, I'm not convinced that the OP intended the diagram to be to scale; maybe the installation is entirely symmetric, but drawn with longer pipes to Tank 2 simply to make it fit easier in the diagram.

But even if the drawing is to scale, I marked it up below to show the imbalance. Below the black line is the section where there are two parallel paths. And the green piping represents the extra piping to Tank 2; if you deleted that piping, it would be balanced.

[Exactly so on the hot side; on the cold side, I assumed that the cold tee has Tank 2 on the straight path, and Tank 1 on the side outlet, and that the equivalent length of "straight path of tee plus 90" is less than the equivalent length of "side branch on tee", so the non-green piping after the tee represents that difference in equivalent lengths.]

So maybe worst case the piping to Tank 2 is twice as long? It depends on what equivalent length, if any, should be assigned to the tank itself. I already did the math for the case that the Tank 2 path is 50% longer; if it's twice as long, then the flow through Tank 2 should be sqrt(1/2) = 70% of the flow through Tank 1.

Cheers, Wayne

View attachment 84015

Um no it's nothing in hydronics you can feel exactly where it stops.

I don't really know what to say other than water takes the path of least resistance and flow is not possible without differential pressure. If there is no differential there is no flow.

 

[wwhitney]

I don't really know what to say other than water takes the path of least resistance and flow is not possible without differential pressure. If there is no differential there is no flow.

Agreed on the differential pressure comments, but water takes all available paths between points of differential pressure.

So as soon as water flows on Path 1, there will be some pressure drop due to frictional losses from the flow on that path, and that pressure drop will provide the pressure difference to give you some flow on Path 2.

Cheers, Wayne

 

[Gnomer]

Wow, y'all are giving me some great info, thank you! #2 in the drawing is the one that failed, and the drawing is not at all to scale but the tanks are the same size and both propane.

 

[Gnomer]

Yes, water moves from high pressure to low pressure, but it takes all available paths. Just like electricity. I redrew the OP's diagram below to better illustrate the logical arrangement of the piping.

There are two paths from the cold tee labeled A to the hot tee labeled B. Water flow will take both paths, as shown by the arrows. If Path 1 through Tank 1 and Path 2 through Tank 2 have identical resistances to flow, then the flow through the two paths will be identical. So best practice would be to make everything from tee A to the two tanks exactly symmetrical, and likewise from the two tanks to tee B, and that should give you approximately equal flow.

If, as the drawing appears to show, Path 2 is longer than Path 1, then Path 2 will have more resistance than Path 1, and you'll get more flow along Path 1. During flow, there will some pressure value at the inlet to tee A and a lower pressure at the outlet of tee B. That pressure difference is what is driving flow through both Paths, and both flows will have identical pressure drops.

So say Path 2 has a 50% greater effective length than Path 1, then its pressure drop for a given flow will be 50% greater. Pressure drop doesn't vary linearly with flow, it's approximately quadratic (double the flow gives about 4 times the pressure drop). So the flow through Path 2 would be approximately sqrt(2/3) = 80% of the flow through Path 1.

Cheers, Wayne


View attachment 84010

Maybe that is a hot water return, I'll have to check when it's fixed. But, at the time, that copper was cold to the touch.

 

[wwhitney]

Yeah, so getting back to your question : - ) I don't think there's any reason not to add cold shutoffs to both tanks, and it would be a good idea.

Also worth bearing in mind that any changes you make to the piping it would be helpful to keep the parallel paths as symmetric as possible.

Cheers, Wayne

 

[Fitter30]

Think what the plumber was trying to do is reverse return ( first in last out) which what you have is a mess.

 

[Reach4]

There are advantages to the 2 heaters being in series. There is not a balancing problem that way.

You can add valves to let you take one tank out of action. With your parallel setup, you would need 4 valves. In series I think 6 valves.