The local weatherman did it again. They predicted a snowstorm of Biblical proportion, but we are experiencing blue skies and lots of sunshine. I report that since it is very hard to accurately determine when snow or ice is going to show up. In the ideal world, the snow melt system is in the idle mode, keeping the slab temperature around the melting point. If snow falls unexpectedly, it melts on contact, creating an oasis of clear pavement surrounded by the white stuff.
This job is located near Dayton, Ohio. The contractor sent me this wonderful picture in January, thanking me for my help. The homeowner wanted the system installed so his wife wouldn’t have to shovel the driveway. His health condition would prevent him from doing that chore in the future.
Design Work
Actually, I didn’t do the design work on this project. Like I like to do nowadays, I forwarded the contractor’s request to the Kid. He has a wiz bang computer program that takes all the fun out of figuring up all the details. In the past (or if I have to do a quick estimation) I have my routine of taking the square footage of the area to be conditioned to a material list. Pencils and erasures are required.
Hydronic design work requires a balance between pipe size and pump size. During my hitch in the Air Force as a radar technician, the balance was between wire size and current flow. In this case, it is tube size rather than pipe size. The smaller the tube is, the bigger the pump needs to be to overcome the increased resistance to flow. I like bigger tubing and smaller pumps.
I went over all this in last May’s column, Make Snow to Melt Snow. To update that project, the job was installed and started up this winter without incident, operating at a 23° F water temperature difference. Not far off from the 25° F water temperature difference used in that design procedure.
Obviously, the Kid’s computer design worked for this driveway. If you look closely, you can see the morning paper laying there with snow on top. The newspaper acts as an insulator between the heated driveway and the plastic wrapper, allowing a thin layer of snow to accumulate.
Reviewing this design, the area to be melted was 1,120 sq ft. The computer selected ¾” tubing spaced at 10” on center and split into 6 loops of 250 lineal feet or less. The pump was sized for 17 GPM at 15 feet of head, again based on the 25° F water temperature difference. It’s all good on paper and practice.
Down to Execution
Kudos go out to the contractor. He’s a guy I love to work with. He does a great job of giving us the details required and a better job of executing the plan. I can’t downplay how much it matters to follow the plan…
Recently I was on another snowmelt startup. It didn’t go so well. When this contractor called to arrange the job site visit, I wasn’t aware that the job was sold, much less already installed, or maybe I forgot. That seems to happen a lot also. Anyway, I meet the two techs that installed the boiler, but not the tubing.
This project is designed to melt the steps at a church, not too far from that ski resort in southeastern Indiana. I guess I had forgotten about it being sold, because they had a piping sketch of the boiler room only that I roughed out somewhere along the line. Like many other projects, I was disappointed in the location of the circulating pump. Installed on the return piping, rather the preferred location on the supply piping, pumping away from the expansion tank connection.
The water circulates with the pump in either location, but creates better air control when pumping away. This job wasn’t circulating well at all because the boiler water temperature was going up rapidly and cycling the boiler off. We purged the air out of the loops more than once.
But let’s go back to pump sizing, since that has the most effect on water circulation. If the pump is too small, the flow rate is low while the temperature rise is high. Exactly like what we were experiencing.
The water/glycol mix was circulating to a two loop manifold. We used ½” tubing to make it easier to make the bends in the steps. At this point I have no idea how long the loop lengths are, I just know it’s acting like there is a major restriction in each one. The flow indicators on the manifold are barely indicating any flow. It was a cold day and had been cold for a while, so the slab is cold and so is the water/glycol mix.
I mention that because a cold fluid is harder to circulate than a warm fluid. In this case, our water temperature at start-up was 36° F, while our design water temperature is 120° F. In ½” tubing at the same flow rate, the head loss is more than double at temperatures of a cold slab compared to the operating temperature.
"Little" Changes – Big Problems
Adjustments were made to the boiler’s firing rate and water temperature to keep it running before I left, but I knew I would be back. I checked the original quote back at the office and saw that I had quoted three loops for the job, however only two were installed. I had increased the number of loops while limiting the loop lengths in the design phase to keep the head loss down.
Unfortunately, whoever installed the tubing decided to decrease the number of loops while increasing the loop lengths. This “little” change increased the head loss dramatically, especially at low fluid temperatures. I guess they thought they were saving time or money, but the cost of installing another pump in series with the existing pump is much greater than the cost of a roll of tubing.
Patrick Linhardt is a forty-year veteran of the wholesale side of the hydronic industry who has been designing and troubleshooting steam and hot water heating systems, pumps and controls on an almost daily basis. An educator and author, he is currently Hydronic Manager at the Corken Steel Products Co.