Good Intentions (John Siegenthaler)
If you've been in the radiant heating business a few years, chances are you've seen at least one thermally challenged installation. Perhaps it was a system you were called in to correct. A job designed and installed by a P&H contractor with good intentions and considerable experience with traditional hydronic baseboard systems, but lacking training in modern low temperature radiant systems. With an air of self-confidence justified by his tenure in the industry, this contractor probably visualized the pending radiant job as little more than a slight variation on the baseboard systems he'd been installing for years.
I recently got called to look at just such a job. The location and installing contractor are irrelevant. Based on what I've seen and heard, this job could show up anywhere in North America. It could be built with anybody's tubing, boilers, pumps and controls. The installation errors are not product specific. Instead they fail to recognize some fundamental limitations imposed by thermodynamics.
The system I inspected is less than a year old and, at least for the time being, delivers barely tolerable levels of comfort. It does so while consuming copious amounts of fuel, and making far less than a consummate impression on the many people who occupy the building on cold winter days.
What follows are the forensics of a system that, unfortunately, is not the first nor the last to create its share of radiant heating skeptics.
Who Needs An Actuator?The photo in Figure 1 shows some of the near boiler piping as I found it this past winter. The two four-way mixing valves were neatly installed one directly in front of the other.
Note that neither valve has an actuating motor nor controller. It's likely the installer felt the valve handles could simply be set at a given position based on the water temperature supplied from the boiler at the time, and henceforth the system would simply operate at the same water temperatures. It's also likely he thought the four-way valves were guaranteed to protect the boiler from sustained flue gas condensation. After all, that's why there are four ports instead of three, right?
Although such valves are specifically made for coupling conventional boilers to low temperature distribution systems, without a controller and actuator they're totally oblivious to what's happening with supply and return water temperatures.
If the water temperature from the boiler, as well as that returning from the floor circuits, remained constant, a manually set, four-way valve would work fine. For that matter, so would a pair of manually set, two-way globe valves. Unfortunately, the water temperatures coming from the boiler and return side of the distribution system change almost constantly as the system operates. Since the valves can't "see" what's going on temperature wise, they can't react to it. Neither can they protect the boiler and floor circuits from "out of range," possibly even damaging, temperatures.
The correction for this problem seems obvious. Install the actuators and controllers and let the valves do their thing. Too bad the installer didn't think of this contingency when piping the valves. Take another look at Figure 1.
It would be easy to mount the actuator on the first mixing valve, but the second valve is piped about 3 inches directly behind the first. The actuator will not fit in the narrow space in between. The only way to create the needed space would be to repipe the rear four-way valve.
If this were the system's only shortcoming I would quickly recommend doing so, adding the controls, and being done with it. Regrettably there're more glitches to deal with.
Radiant Potion No. 9I recently listened to a consumer advisory radio program where a caller asked the host's opinion on a newly released paint that supposedly adds an R-70 (that's right ...70!) insulation factor to existing walls. There were no independent lab tests, only the manufacturers assertion that the product would deliver such performance.
Fortunately the host pointed out that without independent tests to support such a claim, there's ample room for doubt. By the way, R-70 is roughly equivalent to 22 inches of fiberglass insulation. Just think, with a couple of coats of this paint we could heat our houses with a 100-watt light bulb!
Perhaps the same company sells an equally far-fetched potion that enhances heat transfer. When applied to PEX tubing suspended between floor joists, it allows heat to profusely radiate into the rooms above, even though the water temperature in the tubes is a mere 105 degrees F.
The job I inspected could sure use a gallon of this concoction if it were available. You see, a portion of the building consists of a kitchen and restrooms built over a basement. The floor is framed with wooden I-joists spaced 12 inches apart. In this area 1/2-inch PEX tubing was clipped to the underside of the subfloor - without heat dissipation plates, and without underside insulation (see Figure 2).
The circuits were routed to a manifold station piped in parallel with the manifolds serving circuits in the much larger slab-on-grade portion of the building. This means all supply manifolds will get the same low temperature water. The space above the suspended floor has a northern exposure and a relatively high heating load per square foot of floor area.
Let's just say the refrigerator in that kitchen doesn't run much during the winter.
The "staple-up" portion of this system needs significantly higher water temperature, better heat dissipation detailing and underside insulation. Given the present detailing, I'd be surprised if the heat output even reaches 5 Btu/hr./sq. ft.
Ample AfflictionsSo what else could possibly be wrong with this installation? How about:
- A unique method of enabling boiler operation: A jumper wire across the (T T) terminals on the limit control (see Figure 3). As long as the building has electricity, and there's fuel in the tank the boiler is firing until the lack of load allows it to climb to its limit setting of 190 degrees F.
During warm weather, the boiler needlessly fires two or three times each day only to replace its own heat loss to the building. Sadly, mounted about a foot away from the limit control are two zone relays, each with a set of dry contacts for independently enabling the boiler upon a call for heat.
- No boiler return temperature protection: There is some type of bypass pipe between the supply and return side of the boiler, but no temperature sensing mechanism to prevent the boiler from operating below the dewpoint of its exhaust gases.
While I was in the mechanical room, the boiler ran non-stop with its temperature indicator showing a steady outlet temperature of 105 degrees F (remember the limit control is set for 190 degrees F). All floor heating zones where on. The large slab-on-grade floor was dissipating all 240,000 Btu/hr. the boiler could send it without need of higher water temperature. Only when we turned off one of the two large zones did the boiler climb to its limit temperature and turn off.
Whoever opens that boiler for cleaning is likely to find some pretty nasty scaling and corrosion. There's also a good chance the galvanized flue pipe will be toast in a few more months.
- Although not really a hydronics error, the installer selected a 10-gallon electric water heater to handle a kitchen and two restrooms in this building.
Perhaps I would concur with this selection if the only load was a couple of restroom lavatories, but factor in the dish washing demand of the kitchen and that little tank wimps out pretty fast - a fact already noted by those who've used the kitchen. Given the way the boiler was wired for perpetual heat demand even a tankless water heater would have been a major improvement.
The MakeoverThe schematic in Figure 4 shows a piping schematic of the original system; Figure 5 shows the modifications I recommended.
You may be wondering why I recommended changing the system over to variable speed injection mixing rather than reusing the existing four-way valves. Given the necessary repiping, as well as the cost of the additional hardware, it turned out to be significantly less expensive to convert to injection mixing. The line-sized mixing valves are also larger than necessary and thus likely to experience limited stem travel in this application. Better to salvage them for use on other systems.
The two injection pumps will be piped to a split primary loop ensuring the same injection water temperature to each zone. Each injection control will be equipped with a return water temperature sensor to prevent the boiler from operating with sustained flue gas condensation.
All controls will be wired so the boiler is fired only when there's a demand for heat.
The staple-up circuits, after being equipped with heat dissipation plates and 6-inch fiberglass underside insulation, will be controlled as a separate, higher water temperature zone.
An indirect water heater is shown to keep up with the demands of the kitchen.
I also recommended a zone of baseboard to allow certain offices to be maintained at normal temperatures when the majority of the building is in setback.
Pay Now Or Pay LaterWhen we first discussed the system, the owners felt the $25,000 charged by the contractor was too much for providing floor heating to 14,000 sq. ft. of space. Most owners of marginally performing systems feel they've been overcharged.
I pointed out that the numbers worked out to a unit price of about $1.80 per square foot. How many of you can properly install a hydronic floor heating system (boiler, chimney, tubing, fuel tank, pumps ... the works) for such a price? If you think it's possible, perhaps you should consider another line of work - I'm sure the company selling the R-70 paint needs more sales people.
My speculation is that the installer started realizing his true costs as the installation progressed, and then omitted some of the necessary hardware and installation details in an attempt to cut his losses. Fortunately for all parties, this happened after an adequate amount of tubing and underslab insulation were already installed.
The changes I recommended can all be made in the boiler room. They'll probably cost in the range of $5,000-$6,000 - a cost that will likely pay for itself in fuel savings in five years or less (not to mention the improved comfort and longer equipment life).
Most of the changes are add-ons. Many components, including circulators, flow-checks and even some pipe fittings, can be reused. Still, costs would have been lower if the system was properly detailed from the beginning. By the way, anybody looking for a couple of slightly used 1 1/4-inch four-way mixing valves?
Get Learned Or Get BurnedWhy write about a system like this? I hope the forensics we've discussed will reduce (but not eliminate) the number of repeat performances by other well-intentioned installers over the coming months.
Radiant heating isn't rocket science, nor is it Soldering 101. Those who won't avail themselves of the numerous opportunities now available for learning the fundamentals of radiant heating are destined for at least one experience similar to what I've described. Perhaps one with a couple of attorneys mixed in to elevate the feelings of animosity, and costs of resolving the situation. I only wish those of you who think this doesn't happen could listen to some of the unsolicited inquiries for "expert witnesses" left on my answering machine.
Everyone, including heating pros, err on occasion. Reputable professionals correct and learn from their mistakes, and remain in constant pursuit of perfection. Few who repeat the same mistakes out of ignorance, arrogance or an unwillingness to learn will find much sympathy in this, or any other industry.
Good intentions are a necessary but insufficient condition for installing a successful radiant heating system. Learn the fundamentals. Find an approach that accommodates them. Don't let your customer design systems by giving them a la carte hardware options and pricing. For the sake of yourself and the entire radiant heating industry do it right or don't do it at all.