Troubleshooting a mystery... leak/flow reported with no physical leak found. Expansion Tank/Indirect

rossn

​Apologies in advance for this being lengthy...

I have an odd situation going on. I have a moen flo smart water valve that has shut off the water after detecting 0.1-0.3 gpm flow for 10 minutes 6 times over the last 3 nights... only late at night. It then turns off the water. I've been up for some to check and turn on the water for half of the events, and then it happens again after I go to bed.

It appears to actually be flowing that amount of water, but to where... it appears to be disappearing? Almost all of my plumbing is exposed, with the downstairs unfinished... no notable leaks found anywhere (a gallon of water would be obvious), no sounds of toilets running, and the only drain connected appliances are a dishwasher and clothes washer + toilets.

I am on a 30 year old natural gas cast-iron boiler with a 20 year old indirect hot water heater. Supply side thermal expansion tank is maybe 6 years old.

I need more active events to test, though it did appear that the flow went to 0 when I turned off the CW supply to the indirect last night, then resumed when I turned the valve back on. On a prior event, I turned off the cutoff valve downstream of the smart meter... when I turned it back on, flow went to 0.

Maybe a year ago, Flo reported a thermal expansion tank issue, but that has not reared its head again until today when I ran maybe 30 Flo leak check cycles, and twice in a row I got that message (when I had turned off the DHW manifold valves off).

Since Sep 28, I have been getting drip reports from the meter. Based on today's testing... think it's somewhere on the hot water side of things.

The other notable issue is that the boiler system has been getting notable air in it for the last year and a half... I have not been able to track that one down.

Is there any possibility of a failure on the indirect that is feeding the water over to the boiler (or a bladder failure)? I would think I would see abnormal pressures there, if that was the case.

Sorry for the ramble... it's complex. Thanks for your thoughts on this one!

Moen Alert Times (shutoff 5 minutes after):

Oct 12, 12:43am 6 mins at 0.3 gpm

Oct 12, 11:25pm 8 mins at 0.6 gpm

Oct 14, 1:51am 5 mins at 0.2 gpm

Oct 14, 10:44pm 5 mins at 0.2 gpm

Oct 15, 12:16am 6 mins at 0.2 gpm

Oct 15 1:37am 5 mins at 0.2 gpm

Robertw

Try food coloring in your toilet tanks to make sure they aren't running? Sometimes its so minor you cant hear it.

rossn

Thanks - I will try that tonight. I assume no staining with that. I can usually hear water downstairs through the abs, but am game for trying anything.

Jamie Hall

My bet would be a toilet tank float valve. It won't be the main valve (likely), but either the level in the tank set just a tad too high, and overflowing through the overflow, or more likely the float valve not quite shutting off and discharging through the tube which goes to the overflow. You won't hear that, and you won't see it in the bowl, either.

If it's the bowl recharge line from the float valve, food colouring in the tank won't show you that.

rossn

Thanks, Jamie. I will note and check that. Odd thing is it only happens late at night and wee hours of the morning. Maybe tonight I will just turn off the toilet supply lines. One of those toilets has a toto washlet on it, as well.

Larry Weingarten

Hi, Another way to check this is to shut off water to one toilet and see in the morning if the leak went away. Try it with a different toilet the following night. With some luck you'll now know where the problem is, and from there can focus on just one appliance.

Yours, Larry

rossn

Thanks, Larry.

Last night I did shut off both toilets. I got the 10:30 alert with those off, as well as the hot branch for the showers and tub. I then turned off the west end of the house, the boiler supply, and the other pex manifold DHW supply lines, and didn't get a second alert, which had happened the 3 prior nights. If tonight runs clear, I'll enable those one at a time until we get there :)

JimBerreth

Is the boiler relief valve or the backflow preventer piped to a drain? One of these could be discharging water during a boiler run cycle and the evidence (water) is going down the drain so you would not see it.

Mark Eatherton

I love a mystery… Who dunnit? As one who has been chasing leaks for over 50 years, I can tell you one thing all leaks have in common. They waste water.

You are on the right track by eliminating potential sources, but I have to caution you that you might have more than one leak, but you're on the right track.

As others have noted, it could be a relief valve (T&P or P only) whose termination is in or near the floor drain. Is there standing water in the floor drain that you didn't put in there? The trap seals on most of the floor drains I've seen evaporate, but I'm on the edge of the desrt (Colorado).Easy way to determine that is to place a cup or vessel of some sort at the end of the relief waste. It's caled a "witness cup". If you find water in the cup in the morning, you've found your leak. Now figuring out WHY the relief valve is doing it's job is your next task.

If the leak is in the heating system, like buried lines between basement baseboard sections, then it may require equipment that you don't have, like infrared cameras, hydrophones or helium injection/detection equipment. Generally speaking tho, a leak beneath the floor would be constant and consistent, and your provided charts don't indicate that.

Another possible issue may be that it's not actually a leak at all, but instead thermal expasion creating NEW water volume that is backing through the flow sensor. If there is a backflow preventer in the incoming cold water line, then the chances of that are slim to none. Go back to the witness cups and see if there is discharge occuring when you're not looking. And it might be the T&P relief valve. During the day, there may be enough water use to keep the pressure low in the DHW tank so the relief valve doesn't actuate when you are looking, but at night, when no use is occurring, the tank has to recover energy loss, and may fire and cause the relief valve to do its thang,

Exess air in the heating system is an indicator of fresh oxygenated water being introduced, but the air can be coming from a WHOLE bunch of other reasons (read Pumping Away).

And yes, potable wate expansion tanks do fail. Biggest problem is incorrect adjustment on new installations. They come precharged, but you must adjust the air charge to the normal operating pressure of your system. And even at that, the diaphragm can fail and cause issues that I've already addressed. I've seen the diaphragms stick to the tank and not allow for any expansion, and tear and fail and on and on.

Keep looking. You'll find it. Good luck and let us know your findings.

ME

Robertw

As Jaimie hall poined out:

If it's the bowl recharge line from the float valve, food colouring in the tank won't show you that.

The refill tube being too far down the over flow will siphon the tank water out and yes likely not be heard or possibly not seen, but maybe a ripple in the bowl. This can be fixed simply by attaching the refill tube to the top of the over flow as indicated in fill valve instructions.

But food coloring in the tank would definitely show if the tank is leaking thru the actuator seal, or if the fill valve itself isnt shutting off completely as they fill the tank from inside the tank, thereby mixing with the tank water

hot_rod

food coloring in the bowl would be an indication of one or both failures. The flapper valve, or the fill tube overflowing or siphoning. It would be easy enough to eliminate one of the variables, to pinpoint the problem.

At some point from a $$ standpoint replace both the fill and flush valve and eliminate both concerns.

GrafHeating

I for one would like to know the resolution on this one. Please report back when it is found.

rossn

Wow… was surprised to come back here and find all the responses from October! For some reason I didn't get the notifications, back in October, but then again, I was also having issues with my e-mail. Thanks, all - for the thoughtful responses! The issue disappeared after my last post, but tonight it reared its ugly head again, and I'm determined to get to the bottom of this.

@Mark Eatherton, it's nice to hear from you. You actually know me and my project from 5ish years ago, though you never saw the installed system. You were a great help to me. I think you might be on to part of this with the expansion tank. G put that tank in maybe 6 years ago and I haven't had reason to check the tank or it's pressure before now (perhaps I should have). It's set at 60psi, and the pressure coming into the house cycles from about 52 psi@7am to about 56psi@2pm, then declines until 7am the next day. Tonight I adjusted it to 56psi. What should I be setting it to? Should it be below 52psi to work effectively?

All,

Tonight's troubleshooting was meaningful (and long… it's now 3am), though there are still some mysteries.

I think Mark may be right that there are likely 2 things going on. I suspect there is a slow drip, possibly a toilet, but I think there is a different and more significant issue causing the 0.1-0.4GPM flow rates, and I am confident those flow rates are not related to a leak. I am thinking the root of that issue either has to do with the thermal expansion tank or city water having a cyclic pressure change between 50-54 psi with a period of about 10 seconds. I haven't looked at this previously, so I don't know if this is normal, or out of the ordinary. However, it does strike me as odd. I now believe water is literally flowing in and out of the house, and that impeller in the Moen Flo is spinning both directions on a cycle.

Here's how I came to that thinking… When I re-plumbed the house, I put a few strategic ball valves at the major branches. Only 3 fixtures can I not shutoff through those valves… cold water supplies for the master shower, steam generator, and cold water supply. When the issue arose again today (err… yesterday), I shut off all branches, including the west end branch, the east end branch, the boiler fill valve, everything downstream of the DHW thermo mixing valve, the recirc pump, even the boiler DHW supply and return.

After this, the sporadic flow between 0-0.4 GPM continued just as before (maybe slightly less pronounced). I then shut off the CW supply to the indirect, and the 'flow' stopped, despite having no outlets downstream of the DHW. Turn back on the DHW supply… issue resumes.

You can see it in this pressure graph, where I have everything downstream of the DHW off, then I shut off the CW supply to the indirect, then turn it back on.

And, you can see what happens to the mystery flow when I turn off the CW supply… it goes away.

When I installed the Moen Flo, I installed a ball valve upstream and downstream of Flo. So, I took turns independently shutting the ball valves, while using Flo to measure pressure vs time…

First, I checked the house pressure. I do think there may be a small, consistent leak at a toilet, and need to investigate further. A couple tenths of a psi in 4 minutes. And we all know if it was seeing 0.1-0.4 gpm for 4 minutes with the main shutoff off, pressure would have dropped like a rock - so we know these are independent things.

Then, I checked the street pressure (valve upstream of the Flo shut):

Varies about 3-4psi, 6 times a minute.

So, what do you all think is the issue? City water pressure issue? Expansion tank issue? Something else?

Based on what I see above, I'm starting to think this may be abnormal city pressure while the side arm is serving as a poor man's thermal expansion tank (bladder compliance), meanwhile the thermal expansion tank either needs a lower pressure, or it's not functioning correctly?

Keep in mind… this has happened a handful of days in October, and now again in January. Always starting 10-11:30 at night, and similar flow. Sometimes it goes away after a short while or is intermittent in the same night. Tonight it was consistent for hours for the first time.

Alright, let's hear it!

PS - Mark, after all these years, I just (now) got a check from the Kitec settlement… thanks for making me aware of its fitting issues when you did. I guess I can now recycle all those fittings.

Mark Eatherton

Simple fix. Install a single seated spring check on your incoming cold water line. Then adjust the expansion tank bladder to the highest anticipated pressure. Anything above that will be absorbed by the tank and not back into the main. The flow sensor is looking for pulses from water movement and doesn't discern forward flow from back flow. Check valve will stop that. Ask your water department to check the presure regulator serving this area. Looks like it is "hunting", which it shouldn't.

Glad you got resolution on Kitek.

Thanks for getting back to the group.

ME

rossn

Mark, I'll reach out to the water department to check and will leave the pressure where I set it to the expected peak.

I'm several hundred feet from the street on 1" copper, so dramatically undersized, and as a result see some notable pressure drop as a result of moderate flow. It is an issue, and I'm sensitive to further impacting flow (i.e. no pressure regulator). Any idea what type of pressure drop those check valves typically see, and if there is a style that is less impactful than others?

Mark Eatherton

In a static situation, with no flow, length of main would not create ANY pressure drop or fluctuations. Any droops you are seeing during non draw are from the city. All devices have a CV factor for them, and it varies by manufacturer. Once you know the CV factor (feet of head OR PSI at a given GPM flow rate) you will be able to determine how much pressure drop you will realize during flows. A swing check usually has the least pressure drop, but it brings baggage of noise (rattling during flow) with it to the game.

mattmia2

You could also parallel a small and a large check valve. You would want a higher cracking pressure on the large check valve because you don't really want a small flow through a large valve, it will damage the seat.

If there is no demand at night and the city has regulators, they may have trouble regulating the pressure.

I think you get this now but the pressure wouldn't necessarily drop with a small leak with the inside isolated depending on how much water is in the expansion tank. It will keep the pressure close to its precharge pressure until it is empty.

hot_rod

I would not use a swing check in an application like that. A few things about swing checks.

They need a pressure differential to close and seal tightly. A small backflow condition really.

There long travel distance makes them prone to water hammer.

The Cv, typically high, is in the full open position.

A 1" swing with an 11 Cv will have the flapper floating between full open and full close at low flows.

A soft seat, conical spring check travels less than 1/4" as flow stops. The spring immediately starts to close the check as flow stops, no backflow needed.

Pretty much all backflow devices, PRVs hydronic circulators, etc all use spring checks for this reason.

Also an example of water hammer pressure spike on an 80 psi line closed quickly to flow stoppage, i.e. a swing check.

Plenty of good videos showing the operating difference between swing checks and spring checks online.

A spec sheet, with Cv numbers on a low lead spring check.

Swing checks are ideal for dirty or debris water, good for sewer and sump pumps with delta P to close them

Screenshot 2025-01-17 at 2.06.00 PM.png

Screenshot 2025-01-17 at 2.05.39 PM.png

Screenshot 2025-01-17 at 2.06.23 PM.png

Screenshot 2025-01-17 at 2.16.06 PM.png

Screenshot 2025-01-17 at 2.19.09 PM.png

rossn

Thanks, everyone!

Mark - definitely more concerned about the dynamic pressure drop. As an example, I drop 10-12 psi (to 40-42 psi) when drawing 10 GPM. And the house is very long… means some pretty serious pressure drop by the time I get to the rental above the garage, so someone taking a shower when a sprinker zone goes off is a real issue.. Hence, the concern about pressure drop. However, I think my concerns about pressure drop with the check valve are unfounded… more below.

Mattmia - here's what the water manager told me:

"The water storage tank maintains pressure for your side of the service area, and there is a lag between a valve opening and the water “refilling” from the tank.  It starts to get into physics as the speed the pressure recovers is based on a couple variables: head pressure, head loss, the speed of sound of water, the speed of sound of all the pipe materials, the size of the mains and temperature… The little wiggles of water back and forth are also fairly normal.  Water valves in a home typically are not airtight so you can get some backflow when pressure dips outside in the watermain."

I haven't checked again to see if those pressure oscillations are ongoing or more significant than normal, but will. Assuming that tank is not pressurized, and is gravity feed, I'd be surprised that filling it would translate miles downstream to a 3-4 psi pressure differential.

Bob - thanks for the education on and depictions of the swing valve. Before seeing your screenshot, I was looking at the NA51376, and it sounds like this is the series you suggest (I assume 'low lead' for Caleffi is equivalent to 'lead free' of other mfgs, and is ok for potable use).

From the submittal: "Operating pressure differential 0.25 psi (sizes ½" to 1¼" ), 0.5 psi (sizes 1½" to 2"). Flow capacity: Cv 17 (½" & ¾"), Cv 30 (1" & 1¼"), Cv 75 (1½" & 2")."

And looking into the flow coefficient equation, it seems Cv represents flow rate for pressure drop at 1psi. My trunk is 1-1/4" with a calculated 40 WSFUs (26.4 GPM), though I anticipate a more typical max flow of 15-20 GPM. If I'm using the Cv formula correctly, pressure drop looks like:

@5GPM = 0.03 psi
@10GPM = 0.1 psi
@15GPM = 0.25 psi
@20GPM = 0.44 psi
@25GPM = 0.7 psi

Is the quoted differential pressure of 0.5psi on top of this, or just a generalization of the expected total pressure drop.

While I'm stingy on any pressure drop due to my complications, this is a lot less than my initial assumption (mistakenly based on a typical PRV), and so I think I'm fine with it. I'd rather not have it for simplicity/maintenance, but seems it is needed if the water authority says no issue on their end.

Now, the question is where is the right location for the check valve, upstream or downstream of the smart valve (which measures pressure, flow, identifies leaks and (supposedly) other events such as a malfunctioning expansion tank.

Depiction of current system:

CityWaterSupply - - - (Enters home) - - - 1st Ball Valve - - - Smart Valve - - - 2nd Ball Valve - - - Trunk

If I go upstream of the smart valve, I'll be able to detect all events in the home, but I won't be able to detect when city pressure drops (or oscillates)… and sometimes it is important to know that city pressure dropped off. There might be other events I'm not thinking of?

If I go downstream of the smart valve, then I won't be able to detect pressure rises within the home. For instance, if the expansion tank stops functioning. There might be other events I'm not thinking of?

What do you all think?

hot_rod

do you have a water meter in the house or out in the yard in a meter pit? Most of the water meters, or the yoke they go into have backflow devices built into them.

If your water department receives federal funding, BFDs are required.

The .25 is the pressure difference across the check for it to open. Called delta P. So as you open any faucet the flow movement will cause a pressure drop(dynamic pressure) allowing the check to open. Nothing to do with Cv.

The Cv is the gpm that flows through the valve with a 1 psi drop. So if you have a check with a Cv higher than your anticipated flow, there is on pressure drop penalty.

As far as the PRV , most PRV sizing is around a 7 psi drop at full flow. This is known as fall off pressure. The Caleffi PRV have an excellent flow passage, so the pressure drop or fall off is very low at higher demand gpm. The correct way to size a PRV is by the flowrates/ fall off pressure. An example attached.

Of course your long piping runs after the PRV will be the cause of the pressure drop in the building

So high fall off could be your main line from the city is undersized or partially plugged. You could run a flow test before you PRV to get this info.

I think current plumbing codes requires no less than 20 psi, not to exceed 80 psi in the home. So officially is the most distant fixture never drops below 20 at full demand on the system, you are covered.

Most PRV have a strainer in them, check that ocassionally as it will be a pressure drop concern.

I would install a check before the MOEN.

How does the MOEN detect flow, could it be sending false signals?

IMG_9503.jpeg

rossn

Hi Bob, it's a meter pit at the street (I want to say around 300' from the house)… we have a frost-free line defined at 36". I think I recall only seeing a yoke in there, don't recall any other fittings. I couldn't find my pic, so I'll have to check again. Also, the water manager mentioned "Water valves in a home typically are not airtight so you can get some backflow when pressure dips outside in the watermain.  Residential service lines are so small it doesn’t pose a risk to either the resident or the system, but those backflow events are why commercial properties, multifamily buildings, fire suppression systems, and some irrigation systems are legally required to have a backflow preventor installed to protect the water system from contamination. 

So, I don't think there is backflow prevention downstream of my city tap.

Don't have and definitely will not want to have any PRV, as I'm already pressure challenged, with minimal risk of pressure above 60 psi. As to the low pressure side, I can't imagine a shower in the 20's that actually feels like a good shower or will operate a sprinkler, but that's just my personal outlook. Since I'm not a contractor aiming to 'meet code' I'm seeking an all around good experience. I plumbed the house with proper sizing tubing and the long and contorted run in mind (about 225 equivalent feet), but even so, my calculated pressure drop at the furthest, highest fixture (a shower head) is around 18-20 psi drop (relying on IRC flow rates). Likely IRC conditions won't be met very often.

The downside with the check valve upstream of the Moen Flo is that I won't be able to tell if there are city water pressure issues going on without getting out a manual gauge. Sometimes I literally do see city water presure issues.

Based on the troubleshooting I've completed, I'm quite confident Moen isn't sending false signals. It uses a turbine design (with a magnet, I believe), I think flow direction is not indicated. I suspect that given the short cycle, water is literally heading in and out of the house on a relatively small scale, and triggering the unexpected flow events.