Code requirements for domestic hot water return
systems with mixing valves and circulating pumps
to prevent scalding
By Ron George, CIPE, CPD
President, Ron George Design & Consulting Services
The incident
Back in early 2002, an attorney contacted me about an article I wrote in Plumbing Engineer magazine earlier that addressed problems with oversized thermostatic mixing valves. He told me there was an elderly man in an assisted-living center that had died as a result of burns he received in a shower. He said several experts that had investigated the incident had produced the my recently-written article, which warned of the inability of oversized mixing valves to adequately control the mixed water outlet temperature when only one shower was flowing in a large building. The article mentioned that oversized thermostatic mixing valves could contribute to a scald incident. Oversized mixing valves can contribute to scald incidents when there is not a minimum flow maintained through large thermostatic mixing valves. The temperature-sensing element needs to have enough flow in the blended water flowing over the temperature-sensing element to allow the thermal motor to sense the temperature and react to maintain the outlet temperatures to the tolerances allowed in the industry standards.
The standards for thermostatic mixing valves are published by the American Society of Sanitary Engineering (ASSE). The proper standard for this application is titled ASSE 1017 — Temperature Actuated Mixing Valves for Hot Water Distribution Systems. There is also a Canadian Standard CSA B-125.3 that covers this type of thermostatic mixing valve. This minimum flow requirement is a function of the size of the valve body and is what led some valve manufacturers to offer a high-low mixing valve arrangement or parallel mixing valve arrangement with smaller valves and a pressure reducing valve or spring loaded check valve with a small pressure drop in one branch to force the mixed water through a smaller mixing valve. This parallel or high-low piping arrangement allows the smaller valve to blend the water more accurately at low flows. Other mixing valve manufacturers simply use a note that requires a minimum circulated flow rate through the temperature actuated mixing valve to assure enough flow over the thermostatic element to allow the valve to mix properly. It should be noted: when a circulating pump is used on a large valve to meet the minimum flow requirement, there should not be a timer or aquastat used on the circulating pump. If a timer or an aquastat shuts down the circulating pump, then the large temperature actuated mixing valves may not be able to mix properly for a late night shower.
In the above mentioned scald case, the attorney said the scald incident happened about two years earlier and they were finishing up the investigation and deposition stages. He said several of experts believed the scald was the result of an oversized mixing valve based on my article. There were attorney’s and experts for the building owner, the contractor, the water heater manufacturer, the valve manufacturer and the victims family. The experts included: three building inspectors, three master plumbers, two doctors, a scald expert, a valve manufacturer, a sales representative, a water heater expert, a couple of professional engineers and an engineer with a PhD. Many of these experts had been involved in investigating the case for almost a year and several of them had testified that they believed that the scald was as a result of an oversized mixing valve. One expert blamed it on thermal layering or stacking in the water heaters. I quickly dismissed that theory when I asked if there was a circulating pump, and they said, yes, there were three circulating pumps. He said at least two of the experts had produced the article concerning oversized thermostatic mixing valves and they believed that was what happened in this case. After exhaustive inspections and testing, they could find no other explanation for the terrible scald burns that led to the man’s death. At that point one of the attorneys decided to contact me and have me look at the system since the building was just a few miles from my office.
The investigation
Upon inspection, I found it was a three-story building with three wings and about 120 assisted-living apartment units with a kitchen, dining room, hair salon, laundry room, community room, offices and a mechanical room on the first floor. The hot water system consisted of three 120-gallon water heaters all set at 150 degrees Fahrenheit. The three water heaters served the entire building and there were no thermal expansion tanks on the system. There was a two-inch hot water pipe routed from the three water heaters to an industrial-type thermostatic mixing valve with an ASSE 1017 standard listing with two-inch inlets and a two-inch outlet. The two-inch outlet from the thermostatic mixing valve connected to a two-inch hot water pipe leaving the mechanical room and routing to each wing through a two-inch hot water pipe. The entire piping arrangement was not necessarily a good layout, but upon inspection and flow testing, the piping arrangement was working as required and did not appear to be causing the problems that would lead to the scalding incident. There were piping connections for a former water softener that was no longer in the cold water line. The softener may have been removed when the mechanical room was renovated. The city where the building was located had recently connected to the adjacent city’s water supply, so if prior to that connection the city used municipal wells, there would no longer be a need for a softener with the new water supply because that water supply generally did not have hard water. There was a temperature gauge on the outlet of the thermostatic mixing valve that read about 120-123 degrees F. During several flow tests the measured temperature stayed relatively the same. It appeared the mixing valve was working properly. Mixing valves are generally at least one pipe size smaller than the mixed water outlet pipe but in this case it was the same size. Again, upon inspection and testing, the mixing valve was maintaining temperature and the system was operating fine. So, it did not appear to be an oversized thermostatic mixing valve because there was water circulated through the mixing valve that exceeded the minimum flow requirement of the valve. Each of the three wings of the building had a circulating pump near each of the three water heaters. The water heaters were piped with a pyramid piping manifold arrangement that was awkward and used a lot of two-inch soldered tees and elbows unnecessarily in lieu of a reverse return manifold arrangement, which would have used fewer fittings and equally distributed the water flow to the cold water inlets and the hot water outlets of the water heaters. The manifold arrangement was a clue the work was being done by an inexperienced plumber.
The hot water return pipes were not connected to the manifolds. Each circulating pump was located adjacent to each water heater. One circulating pump could have been used, with balancing valves in lieu of three separate pumps, but again, everything seemed to be working properly upon my inspection. The hot water return piping was peculiar in that it was routed to each circulating pump in a downward direction; the outlet pipe was routed down to near the bottom of each water heater with a horizontal swing check valve in the downward position. (This was another sign of inexperience because the swing check valve will hang open in this orientation and not work properly.) The hot water return pipe made a 180-degree turn and went back up to above the water heaters and collected across the top of the water heaters. Near where the hot water return pipe made its 180- degree turn, there was a valve and cap on a pipe leaving the cold water connection to the bottom side of each of the first two water heaters. I noted this and took photos and video, as it looked like alterations had been made to the piping. The hot water return piping was split after the third water heater and one line went to the cold water inlet to the third water heater and the other line went to the cold water pipe connecting to the thermostatic mixing valve. This piping arrangement was awkward and caused the third water heater to work harder to make up for temperature losses in the circulated loop, but it met the intent of the manufacturer’s installation instructions to allow flow from both sides of the mixing valve during periods of non-use and the system was working properly.
A quick check of the manufacturer’s catalogs for the mixing valve and the circulating pumps showed the mixing valve had a published minimum flow rate of 18.5 gallons per minute. The minimum flow rate is what is required to maintain the allowable temperature fluctuations in the mixing valve standard. If the mixing valve had been installed by itself without a circulating pump this would definitely have been an oversized mixing valve problem if only a single shower was flowing through the mixing valve. In this installation there were three circulating pumps and each one had a published flow rate of 10 gallons per minute at 16 feet of head. Depending on the actual head loss in the system, the circulating pumps should have been producing about 18 to 30 gallons per minute through the mixing valve. The piping arrangement looked odd at first, but the system was maintaining temperature when it was tested and monitored.
The chronology
Later in the week the attorney for the building owners stopped by my office to discuss the site visit and we examined the photos of the hot water system in the mechanical room. We also discussed the flow and temperature measurements, which showed it took over 41/2 minutes for the water temperature to stabilize during the flow test. This indicated that was how long it took for the hot water to get from the mechanical room to the room where the incident took place. I asked the attorney for the chronology of what happened in the scald incident. He gave a description of what he understood to have happened in the scald incident and then he produced the maintenance records, and the maintenance records told the story. Apparently, a few months before the scald incident, one of the old water heaters developed a leak. A plumbing contractor was called to make repairs.
The plumber was called out to look at the leaking water heater. He closed the isolation valves to the leaking water heater and gave the owner a price to replace the water heater. He also gave them a price to replace everything in the mechanical room with new piping, heaters, circulating pumps and a new thermostatic mixing valve. The owner, fearing the other two heaters were near the end of their life expectancy, also agreed to have the contractor replace all three water heaters and all the associated piping, pumps and valves in the mechanical room.
The maintenance records showed that shortly after the replacement project there were numerous complaints of water that was too hot in rooms near the mechanical room and water that was not hot enough from rooms further away from the mechanical room during off peak hours. There were numerous service calls to the plumber to correct this problem and every time he came out it was mid-morning and the system appeared to be working fine. The maintenance records also showed numerous circulator pump failures and replacements over a very short period of time following the mechanical room renovations. These were signs of the circulating pumps being dead headed. I asked myself how they could have been dead headed. The only logical answer was the circulating pumps must have been piped back to the water heater only. This would have caused the circulating pumps to try to force hot water through the mixing valve. During periods of no-hot water use, the thermostatic mixing valve would have tried to close down the hot water port and it would have deadheaded the circulating pumps. That would explain the circulating pumps being replaced on a regular basis. The dead headed circulating pumps also would have caused very little or no flow in the hot piping during periods of non-use or at night time hours. This explained the complaints of not enough hot water at the ends of the hallways. (over a four minute wait in the mornings) The thermostatic mixing valve cannot be manufactured with tolerances that are so tight that the piston cannot move, so there is a manufacturing tolerance that allows hot water to leak or flow around the mixing valve even when it was trying to shut all the way down. This caused a slug of very hot water to develop downstream of the mixing valve during periods of non-use and it explains the calls for water that was way too hot from the rooms near the mechanical room and calls for no hot water at rooms near the end of each hallway. The victim in this case was an elderly man and his room was located near the end of the hallway away from the mechanical room. I asked how the incident occurred and I was told he got in his shower and turned the shower to full hot and he only got cool or ambient temperature water so he began showering in the cool water. About halfway through his shower, after he was covered with soap and shampoo, the very hot water (near 150 degrees Fahrenheit) arrived from the slug of hot water near the outlet of the mixing valve. The hot water caused a sudden rise in temperature which scalded him on his face, chest, torso and legs. His daughter said when she talked to him in the hospital he said he had soap in his eyes so he grabbed the shower curtain to try and shield himself from the hot water. Initially his burns did not look too bad so he proceeded to lunch after his shower where a staff person and some other residents in the lunch room commented on his redness and a couple of blisters on his forehead. He went back to his room and the burns became worse. Later the pain, blisters and loss of skin were so bad an ambulance was summoned. He died of a sepsis infection a few days later.
I then reviewed the photos of the water heater installation and the clues were there that alterations to the piping system had been made. There were the capped-off pipes near the bottom of each water heater and pipe and fittings were added routing the hot water return from the bottom of each water heater back up to a pipe that ran across to the last water heater and to the cold water inlet of the mixing valve. The revised piping had corrected the piping problem that caused the scald incident. Upon further review of the maintenance records, I saw where there were several complaints about the temperature after the scalding incident. One of the maintenance logs indicated the same plumber was going to replace the thermal element in the mixing valve and a system shut-down was scheduled to replace the cartridge in the mixing valve. The contractors invoice also indicated that the hot water return piping near the bottom of the water heaters was a trip hazard and it was re-routed to eliminate the trip hazard. There was no trip hazard and this seemed strange so we went back to examine the return piping closer. The pluming contractor claimed he had the piping installed properly in his original installation and he said the piping in the room was all done during the initial installation several months earlier. Photos showed the hot water return piping connection to the cold water inlet of the mixing valve had a different color and type of solder than the remainder of the piping. Upon a closer examination, it appeared the hot water return piping was tied into the system at a different time using a different type of solder. A call to the mixing valve manufacturer revealed there had never been a replacement thermal motor sold for that mixing valve model anywhere. The manufacturer said we could tell if it was the original thermal motor by the manufacturing date stamp on the thermal motor part. We scheduled an inspection of the thermostatic mixing valve. The re-inspection photographs revealed UPC retail price tags on the back of a couple of pieces of the revised hot water return piping. The revisions were made with type “M” tubing with red markings and the original piping was type “L” with blue markings. There were also manufacturing dates on the piping that showed the type “M” piping was manufactured after the original mechanical room renovation project. The contractor could not explain how the piping was modified but he said he piped it correct when he made his original installation and he said someone else must have changed that piping.
Non-union plumbers
The plumbing contractor was a non-union contractor and he did not belong to any industry trade organizations that could have offered training or education. The owner of the plumbing company was a young man that had worked for the former owner of the company doing residential plumbing long enough to get his masters license. The depositions showed he had no formal training in plumbing other than working on residential plumbing for his company. The former owner of the company decided to retire and sell his company to the young plumber after he passed his licensing exam. After the younger plumber bought the company, he did a little more advertising and that is how he got a call for this commercial job at the assisted-living center. He had never worked on a commercial project or a project with a mixing valve or a circulating pump, and he admitted he did not read the installation instructions.
Union plumbers
This case highlights the value of using a union-trained plumber. Union plumbers go through hundreds of hours of formal union apprentice classroom training and they get extensive shop training in installation, service, and joining for various joints, materials and plumbing equipment. Union plumbers also get backflow training and they learn about plumbing code requirements and the health and safety aspects of the plumbing systems. In addition, they also get thousands of hours of on-the job training.
Lost evidence
After the deposition at the attorney’s office, I asked the attorney if there were any photos of the mechanical room immediately after the scald incident. He said yes, but they no longer exist. Then he paused and thought for a moment and said, “I think I might have something, follow me.” We went to a file room where there were rows and rows of file cabinets. There were at least five file cabinets with two years worth of files, depositions, photos and documents for this case. We spent the next hour or so digging through the file cabinets and then he found what he was looking for. It was a deposition of the insurance company’s inspector with the exhibits.
The insurance company’s inspector had taken photos of the piping in the mechanical room a day or two after the scald incident as part of their investigation. The photos were photocopied and attached to the insurance representative deposition and distributed to the various attorneys in the case. The photocopies were poor quality and would not be admissible in court but they clearly showed I was right. There was no hot water return piping across the top of the heaters connecting to the cold water inlet to the mixing valve. Unfortunately, the original photos were sent to the insurance company’s headquarter offices in the World Trade Center in New York city in the summer of 2001 and were lost in the September 11, 2001 world trade center disaster.
There were still the capped pipes, added pipe couplings and rerouted piping manufactured after the original installation date that told the story of what happened. I called the local representative for the mixing valve manufacturer and he said he remembered getting a call from the contractor complaining about a defective thermostatic mixing valve. When the valve representative asked the contractor if he piped it properly, the contracted asked what he was talking about. The rep asked him if he had split the tempered water return pipe and routed it to the cold water inlet to the water heater and the cold water inlet to the mixing valve as it was shown on the installation and maintenance paperwork. The valve representative said the contractor tried to argue with him and indicated he did not read the instructions. The representative faxed the instructions to him and followed up with a call to make sure he understood how it was supposed to be piped. The manufacturer’s rep said he never heard from the contractor again. I obtained a copy of the fax and the transmittal from the manufacturers representative. Within a day of the phone call and fax between the manufacturer’s rep and the contractor, the contractor called the assisted-living facility and explained that he believed their fluctuating temperature problems were the result of a defective mixing valve and he needed to replace the thermal motor in the mixing valve. The owner agreed to the changes and scheduled a shut-down of the hot water system a couple of days later. The thermal motor in the thermostatic mixing valve was never replaced. The contractor went in that night and corrected the piping mistake and billed the client for changing the thermal motor. The note about moving the pipe to avoid a trip hazard was to cover his tracks if anyone questioned why that work was done. The case settled and the family donated a large sum of money to the local hospital burn center to develop a website to educate the public on how to prevent scald burns in memory of their father.
The code change
Shortly after this scald case was settled, I assisted the hospital with developing information for their website and then I submitted a code change to the International Plumbing Code to require hot or tempered water systems that are circulated through a mixing valve to be piped with a tee after the circulating pump and the mixed water return is required to be routed to the cold water inlet of the water heater and the cold water inlet or tempered water return connection of the mixing valve.
The code change addressed recirculating pumps used with thermostatic mixing valves and first appeared in the 2003 edition of the International Plumbing Code with language that addressed specific piping requirements for hot water systems when a mixing valve and a circulating pump are used together in a hot water system to circulate hot or tempered water. The code change was submitted because the piping arrangement must be installed exactly as described in the following code text otherwise the tempered water system will either rise to the full hot water storage temperature and become a scald hazard or if the return is piped to only the cold water inlet of the mixing valve the system will drop to ambient temperature during periods of non-use and cause sudden temperature rises. The code language is as follows:
607.2.3 Recirculating pump. Where a thermostatic mixing valve is used in a system with a hot water recirculating pump, the hot water or tempered water return line shall be routed to the cold water inlet pipe of the water heater and the cold water inlet pipe or the hot water return connection of the thermostatic mixing valve.
This code change came about because of the above referenced scald death from an improperly piped thermostatic mixing valve. Upon further investigation, I found that not all manufacturers listed this piping requirement in their literature. I proposed it as a code change to allow inspectors to be able to look for this piping arrangement and help prevent this tragedy from occurring again.
Mixing valve and circulating pump piping arrangements
When piping a hot water return system that has a mixing valve, there are three basic designs: |
1. The Master Mixing Valve conforming to ASSE 1017 located at the water heater (See Figure 1).
2. The point-of-use or local mixing valve conforming to ASSE 1070 used at a group of fixtures (See Figure 2)
3. The Dual Temperature System. (See Figure 3.)
One of the most common mistakes a young designer or installer makes is connecting the domestic hot water return to the water heater in a system with a mixing valve. If the hot water return line is connected directly to the water heater, circulated flow is pumped or forced through the water heater. When there is no water flowing from a faucet in the system the mixing valve closes down the hot water inlet port and opens the cold water inlet port to try and mix the water. When there is no flow from a fixture, no cold water would be able to enter the mixing valve to blend with the hot water. In the piping diagram in figure 1, the hot water return is split and routed to two locations.
This is the piping arrangement required by the International Plumbing Code and by manufacturers in order for the mixing valve to work during periods of no-use. By splitting the flow and allowing the flow to go to both the cold water inlet of the water heater (which is the hot water inlet of the mixing valve) and cold water side of the mixing valve, the mixing valve can mix hot water at 150 degrees Fahrenheit with the cooler tempered water return line water at 100 degrees Fahrenheit to get tempered water at 120 degrees Fahrenheit. Most mixing valve manufacturers require a minimum temperature differential from the hot to the cold side of the mixing valve in order for the valve to function properly. A 20-degree differential from the mixing valve setting to the hot water return temperature is recommended, but check with the manufacturer for your application. Some manufacturers valves can operate with a closer temperature differential. If you follow the hot water return pump sizing as outlined in the ASPE Plumbing Engineering Design Handbooks, you should be fine because it is based on a 20 degree temperature differential.
Local Mixing Valve Mixing Valve Located near the fixtures (Figure 2)
The piping diagram shown in Figure 2 is the easiest to pipe and maintain. The thermostatic mixing valve in this application must conform to the industry standard for point of use mixing valves which is ASSE 1070 — Performance Requirements for Water Temperature Limiting Devices. A thermostatic mixing valve that conforms to the industry standard ASSE 1017 — Performance Requirements for Temperature Actuated Mixing Valves for Hot Water Distribution Systems cannot be used in point-of-use applications because the ASSE 1017 standard has a wider allowable fluctuation in temperature swings. Therefore, the scope of the ASSE 1017 standard requires the valve to be located at the source of hot water or near the water heater to allow the temperatures spikes to stabilize in the turbulent flow of the hot water distribution piping.
Designers should try to keep their systems as simple and easy to maintain as possible. This is a very good design and there is no recirculation through the mixing valve. I have found this system to be cost effective because you can pipe only one HW temperature out and mix down to the various usage temperatures near the fixtures.
Dual Temperature Systems (Figure 3)
Dual Temperature systems offer the greatest challenge to the plumbing designer and installer. These systems typically have a master mixing valve conforming to ASSE 1017 at the hot water source or near the water heater and two hot water return temperatures. Notice that there are two hot water return circulating pumps. These pumps need to be independent of one another because if the two hot water return lines were combined, the hot water return temperature at the cold water side of the mixing valve could exceed the set point of the mixing valve. For example: If the mixing valve was set at 120 degrees and the water heater was set at 150 degrees, the hot water return from the 150-degree line would be about 130 degrees if it is sized per the ASHRAE or ASPE sizing method. If the circulating pump flow rate is higher or if there are peak periods where ht e hot water is drawn out to the end of the system the return temperature will be higher. This is why these pipes must be kept separated until after the pumps and check valves. Notice how the tempered water return line splits, as we discussed earlier, and connects to the cold water inlet to the water heater and the cold water side of the mixing valve. You should always check with the manufacturer of the valve for recommended piping arrangements. Following these design configurations will keep everyone out of scalding hot water.
Ron George is president of Ron George Design & Consulting Services. He has served as Chairman of the International Residential Plumbing & Mechanical Code Committee. He is active in plumbing code and plumbing product standard development committees with ICC, IAPMO, ASSE, ASME, ISEA and ASTM. His company specializes in plumbing, piping, fire protection and HVAC system design and consulting services. He also provides plumbing and mechanical code consulting services and he provides investigations of mechanical system failures and litigation support. To contact Ron, email: rgdc@rongeorgedesign.com.