Leaks in our hot water recirculation pipes

Just a few questions on domestic hot water recirculating system. Our building is 13 years old and we have experienced several leaks in the copper tubing of the hot water recirculating system.

The circulating pump is a grundfos UP15-18 SF. The output of the pump is in 3/4'' copper tubing and connects to the 75 gallon water heater.Then from the water heater there are 3 supply lines of 1'' that run in different direction of the building. The 3 supply lines at some point are reduced eventually to the 1/2'' return lines that all connect on the same 3/4'' line at the input of the pump. The building is 40' x 100' with 3 stories.

Over the years, we had several water leaks in the 90 degree elbows and tees on the 1/2'' return lines just before they connect to the 3/4'' line. We installed a bypass on the pump to slow down the flow of water in the supply lines since we thought that there might be a corrosion problem due to the high velocity of water and the fast changes of the water flow. Even with a portion of the flow going through the bypass of the pump we still got some leaks in the return lines.

Would anyone know if this is a problem due a bad design? Should the return piping be 3/4''? Would it help to install a smaller pump? Will a smaller pump give enough head to the system? Our water pH is 7.4 so I don't think acidity is a problem.

Kind regards,
Luc

When you say the leaks occurred in the elbows and tees, it makes me wonder if this is a case of bad solder joints.

Hello Gary,
I said leaks but they are pinhole leaks due to corrosion. The copper inside the elbows or near the tees was heaten away.

When you say bad solder joints, do you mean that some of the solder could have entered inside the pipe and caused a situation with turburlent flow? The turbulent flow would then have eaten away the copper of the elbows.
Regards,
Luc

May not be totally helpful but in my city (city only) copper is frowned upon. Something with the water and/or soil just eats copper. 99% of new construction is CPVC.

My thinking was just poorly made joints that failed, not anything that would cause holes in the fittings. Perhaps there is some mineral in your water supply, but that should affect all pipes and fittings everywhere in your area.

I would bet that the joints weren't reamed, because of the location of the leaks. In a recirc system with the water moving continuosly, it's a must that the joints are reamed. If it was something in the water, or stray electrical current the pinholes would be scattered around.

Seen this before in a 3 year old nursing home. Too much pump and unreamed pipe cuts make for pinholes downstream of the cuts.

I became involved in the situation as a third party contacted by the buildings owner. The original designers, engineers and plumber were all pointing to stray voltage/ground problems. Cutting out a section of the piping (3/4" type L) around a coupling showed heavy deterioration on the pump side of the fitting, while the pipe was nearly perfect upstream of the coupling. The burrs were clearly evident on the original pipe cuts, and the interior of the pipe downstream of the coupling could best be described as the grand canyon. That 6"-8" section of pipe on the downstream side of the coupling was about as thick as a pop can.


Short term, get a smaller pump. Install a timer to operate the pump, allowing it to sit idle at times of low demand.

Thanks everyone for your answers. As I see there could be several possible solutions.

How would I go about in finding out if there is something in the water that is eating away the copper? Can a water test give me this data?

I read on the web that it is reccomended to keep the flow rates bellow 4 feet/second in a hot water circulating system tom minimize corrosion. If I am not mistaking a flow rate of 3 USGPM in a 1/2'' pipe gives a velocity that is bellow 4 feet/second. If it is too much pump head/flow, how can I figure out the velocity or flow in the 3 return 1/2'' pipes?

I know that we can mesure the pressure difference between the input and output of the pump and mathematically convert this pressure into the pump head needed to overcome the system losses. Then using the pump's head and flow curve supplied by the pump manufacturer, I would be able to get the flow rate for the specific pump head. Finally, I could mathematically convert the flow rate to velocity. This would give me the flow rate at the output of the pump but not in the 3 different return pipes where the corrosion problem occurs.

Kind Regards,
Luc

Luc
Doublare hit the nail on the head. What your copper return line is suffering from is corrosion/erosion. It is caused by a combination of any of the following. Water temp above 140ºF and non water soluble solder paste cause corrosion. Unreamed tube ends, excessive flow rate (due to high water pressure, an over sized pump, or too small a return line [3/4" preferred]) and sharp changes in direction on the return line (long turn 90º's and 45º's preferred) all cause turbulence which contributes to erosion.

The Copper Pipe Asscociation knows this to be a problem. We have had many systems here that have had leaks at the elbows. The conclusion was a number of things, one being errosion corrosion as the above person said, the others were not reaming the pipe to the full bore, use of corrosive flux, which now has to be water soluble and having a pump that flows too fast. Hot water systems and recirc systems should be designed not to exceed 5 feet per second.