Why's my pressure test losing pressure slowly, but no drop of water found anywhere?

[Rossn]

TLDR;
I'm air/pressure testing the water supply system snaking through one end of my house. The pressure creeps down slowly. I've checked every joint by hand at 20 minutes, and haven't detected a drop anywhere, but pressure dropped more overnight. More details below. What might be going on?


Additional Information
20 mins @ 60 psi loses 1/2 psi
10 mins @ 100 psi loses 1 psi
1.5 hrs @100 psi loses 4 psi
10 hrs @ 100 psi loses 11 psi

It's all copper, except the test setup. 36 viega propress fittings, 5 sweated fittings, 6 press ball valves (all closed for the test). Only 24" of pex in the pressure test setup.

 

[wwhitney]

If you are using a combination of water and air, it is possible the air is dissolving in the water. In which case you could wait a long time to see if it reaches a steady state. Or periodically repressurize with additional air to the same pressure, and see if the rate of pressure loss decreases with each cycle.

Certainly if the last 3 data points in the OP represent a single test, then the rate of pressure loss is going down, which is consistent with the air dissolving in water idea.

Cheers, Wayne

 

[Sponsor]

 

[Rossn]

If you are using a combination of water and air, it is possible the air is dissolving in the water. In which case you could wait a long time to see if it reaches a steady state. Or periodically repressurize with additional air to the same pressure, and see if the rate of pressure loss decreases with each cycle.

Certainly if the last 3 data points in the OP represent a single test, then the rate of pressure loss is going down, which is consistent with the air dissolving in water idea.

Cheers, Wayne

Thanks, Wayne! That didn't cross my mind... I will add some more pressure and report back, though it's tough being incremental with the air.

Yes, those last 3 points are on single test.

 

[Rossn]

Unfortunately, when I attached the air chuck, some air/water snuck out and dropped the pressure significantly. I assume air dissolved into the water will come out instantaneously. But, will see what this pressure test does.

I also sprayed a soap/water solution over the testing equipment, and couldn't detect any bubbles. Added a sealing cap over the valve.

 

[Reach4]

City water?

I think some or all of this can be explained by the pex expanding. To test for that, keep pressure high for a few hours. Then drop the pressure to maybe 20 or 30 psi. If the pressure then rises, expect that was expanded pex then contracting.

It could also be affected by the diaphragm in the thermal expansion tank relaxing.

Temperature can also affect things. I am not sure that a pressure test with no pressure tank has validity.

 

[Rossn]

City water?

I think some or all of this can be explained by the pex expanding. To test for that, keep pressure high for a few hours. Then drop the pressure to maybe 20 or 30 psi. If the pressure then rises, expect that was expanded pex then contracting.

It could also be affected by the diaphragm in the thermal expansion tank relaxing.

Temperature can also affect things. I am not sure that a pressure test with no pressure tank has validity.

Thanks, Reach!

I had thought about the pex expansion... but there's only 2' of it in the system. It could be, but given it dropped 10% over night, not sure.

I could potentially eliminate it using a piece of copper on those sharkbites.

Good thinking on the thermal expansion tank, though it is out of the picture here - who needs one??? J/k - it's on the other side of a closed valve... only on end of the house is under testing.

Is temp an influencer on a water + air test (all filled with water + topped off (1' of pex) with air)? I had heard about that as an issue on all-air tests.

If this test still drops, I will try the pressure drop check. With the air-water test, a tiny amount of air greatly affects the pressure... so fine tuned changes are difficult!

Edit to add: Yes, city water.

 

[wwhitney]

Is temp an influencer on a water + air test (all filled with water + topped off (1' of pex) with air)? I had heard about that as an issue on all-air tests.

Yes, to the same extent as on an all air test. To first order the water is incompressible, the pipe is fixed in volume, and you have some air space. If the system is sealed with no exchange between the water and air, the pressure in the fixed volume of air is proportional to the absolute temperature.

So if you were doing this in an unheated building, and the temperature changed from 30F (490R) at night to 70F (530R) during the day, you'd expect the pressure during the day to be 530/490 = 1.08 times the pressure at night.

Cheers, Wayne

 

[Reach4]

Air and water both contract with falling temperature, but will expand with rising temperature.

Two feet of pex will have very limited expansion, and that expansion is reversed after you drop the pressure. Looking at your photo I guess the only copper pipe that is pressurized by your test pressure is off to the right, correct?

 

[wwhitney]

Air and water both contract with falling temperature, but will expand with rising temperature.

The thermal contraction of the water is an interesting point. Changing my example of extreme temperature swing to 40F (500R) to 80F (540R), as 30F is below freezing of water, the air pressure change directly due to temperature would be 540/500 = 1.08. The water volume change would be very small, engineeringtoolbox.com says the specific volume of water changes from 0.01602 ft^3/lb to 0.01607 ft3^lb, or a ratio of 1.0031.

But if your system is filled mostly with water, with just a little air, then the impact on the available volume for air could be quite large. E.g. if it is 100 parts water and 1 part air (by volume) now the air space decreases from 1 part to 0.7 parts, or by 30%, and the pressure change attributable to that would be 1/0.7 or 42%.

Copper thermal expansion could also be a factor. The linear expansion coefficient for copper is apparently 9.4E-6 / F. For 40 degrees F that's 3.8E-4, and for volume it would be 3 times as much, or 1.1E-3, i.e. 0.11%. So now if we start with 100 parts water and 1 part air (by volume), we'd end up with 100.3 parts water and a total volume of 101.11 parts, so 0.81 parts air, and the air pressure change is only 23%.

Probably best to do this mixed air/water test isothermally, or switch to basically all air.

Cheers, Wayne

 

[Rossn]

Air and water both contract with falling temperature, but will expand with rising temperature.

Two feet of pex will have very limited expansion, and that expansion is reversed after you drop the pressure. Looking at your photo I guess the only copper pipe that is pressurized by your test pressure is off to the right, correct?

Slightly below grade with open slab, it's much more stable temp than upstairs.

You have a good eye, Reach. However, I did some trickery there! The system is to the left, and what you see to the right is a stub-out for the above-garage apartment, which could be more likely to freeze in the future. Hence the hose drain valve.

To do the test, I pulled the handle and rotated the spindle 180 degrees, so it's actually opposite of what is shown on the diagram. It's also a way I can drain the house, if ever needed, and why it's setup to gravity drain. However, really the valve being 180 from normal position is just for this test

 

[Rossn]

The thermal contraction of the water is an interesting point. Changing my example of extreme temperature swing to 40F (500R) to 80F (540R), as 30F is below freezing of water, the air pressure change directly due to temperature would be 540/500 = 1.08. The water volume change would be very small, engineeringtoolbox.com says the specific volume of water changes from 0.01602 ft^3/lb to 0.01607 ft3^lb, or a ratio of 1.0031.

But if your system is filled mostly with water, with just a little air, then the impact on the available volume for air could be quite large. E.g. if it is 100 parts water and 1 part air (by volume) now the air space decreases from 1 part to 0.7 parts, or by 30%, and the pressure change attributable to that would be 1/0.7 or 42%.

Copper thermal expansion could also be a factor. The linear expansion coefficient for copper is apparently 9.4E-6 / F. For 40 degrees F that's 3.8E-4, and for volume it would be 3 times as much, or 1.1E-3, i.e. 0.11%. So now if we start with 100 parts water and 1 part air (by volume), we'd end up with 100.3 parts water and a total volume of 101.11 parts, so 0.81 parts air, and the air pressure change is only 23%.

Probably best to do this mixed air/water test isothermally, or switch to basically all air.

Cheers, Wayne

Good points. On your point, perhaps the incoming cold water needs to equalize before the test. I can try air-only if the current test doesn't look good (haven't been down to check yet).