by Daniel R. Nichols AIA, homeowner

The following is a “diary” of three repair projects that I have undertaken at the Sweeton house since my wife Christine and I purchased it in 2008. Each project was a learning experience for me and an opportunity to possibly improve a condition that led to a system failure.

November 2013: Circulator Pump Replacement

On a cold Sunday morning in 2010, I suddenly heard a loud grinding racket coming from the heater basement beneath the house’s kitchen. I went downstairs to find that one of the house’s two water circulator pumps was very warm and making a metallic rattling noise. I immediately shut off the boiler and turned off the gas to the boiler. I called my heating contractor and left a message that I had a pump failure on my radiant floor system. I received a call back about 20 minutes later, described the issue and the contractor was at the house within another hour. The problem was diagnosed as a failed coupling in the pump motor. Luckily my contractor had parts on his truck to make the repair, and the system was running again after another ½ hour.

In our postmortem discussion, it was noted the two pumps (Bell and Gosset units dating from the 1970’s and 1980’s) were serviceable, but parts might not always be readily at hand. This means part of the system could be down for a day or more while awaiting parts. Despite the floor slab’s 36 tons of mass and slow rate of heat loss, the thought of a loop of the system being unheated and uncirculated for a day or more in the middle of winter got my attention. My contractor noted that Taco produces smaller pumps of the same capacity with sealed bearings that last 15 years on average that are readily available. When repairs are needed, the pump itself is simply removed and replaced. Given the age of the exiting pumps and their infrequent maintenance in the past, I opted for a planned change in the pumps within the next few years.

In November 2013, before the heating was turned on for the season, I scheduled a service call to replace the pumps. To prevent the draining down of the old, funky, air bubble and mineral free water from the system into the basement, I hoped to close the balancing valves. No such luck. They had become stuck after 60 years without being touched.  Instead small freeze jackets were applied to freeze the water in the pipe at either side of the pumps. Next, a ball valve was installed on either side of the pumps to facilitate future replacements without loss of substantial amounts of water (isolation valves). With the closed valves in place, the freeze jackets were then removed. The existing pumps could then be removed and replaced with new pumps.

Circulator pump and isolation valve with expansion tank behind

Circulator pump and isolation valve

October 2014: Living Room Window Wall Perimeter Pipe

During the course of a restoration of the J.A. Sweeton House’s (1950) window wall and living room roof structure, we encountered a radiant heat failure. The series of photos below start the story:

Exterior slab and 4" cinder block edge curb removed to expose mullion post pins, interior slab haunch, and foundation.

Slab edge exposed and foundation cut for placement of suitable concrete pier supports of mullion point loads.

Radiant heat pipe found 6" closer to the exterior wall and 3" higher (engaged with the slab) than indicated on Wright drawings. Some scaling found but not leaking....yet.

The day after the photos above were taken, the bottom surface of the pipe (exposed under the slab due to outfall of crushed stone during the dig) appeared to be sweating.  The following day, the exposed scaled pipe developed three pinholes on its bottom surface which were slowly, but visibly, dripping. I called my heating contractor who had assisted us with our conversion from oil to gas fuel, and had recently installed new circulator pumps. He turned on the pumps and the drip rate increased and a muffled intermittent hiss could be heard.

The solution was to carefully remove concrete around the pipe and down its length until solid sections suitable for mating splices could be found.

A 36″ length of pipe was removed and a new section of the same material was spliced in and painted with red oxide primer. Original pipe is schedule 80 steel pipe as is the splice. A welded connection of the splice was considered, but concern for super heating 60+ year old pipe whose condition for several feet beyond the splice was unknown, won out, and compression couplings were applied where the steel was full wall thickness. The pipe was left exposed for a day with the pumps running (20 lbs pressure vs 5-10 lbs at rest) to test the seal of the pressure fittings.

From what we have found, the heating contractor in 1951 chose to use a different detail than that shown on the Wright drawings: instead of placing the pipes in the stone beneath the slab, the pipes were set on the stone with some stone placed between to allow 2/3 of their surface engaged with the bottom of the slab. Welded wire fabric crack control was then wire tied to the radiant pipes. My heating contractor noted this was a common installation method with these systems in the ’50’s and ’60’s. The pipe corrosion at the leak location seemed concentrated at a bracket that was used to space the pipes. From what little we have exposed, these brackets appear to have been placed every 8 to 9 feet. If the bracket was a dissimilar metal, it may have corroded the pipe through galvanic action.

Another more likely possibility is that the sagged roof ridge concentrated sheet flow of rain water off the roof at this central section of the wall; dripped water then seeped in between the exterior slab and the porous cinder block (not concrete block) adjacent to and beneath the interior slab edge; capillary action via slab edge cracking could then have allowed water to repeatedly wet the pipe from its outside surface causing corrosion.

Other possible sources of leaks include:

  • Carbonation of concrete: This could lead to corrosion  of the pipe. This could be a factor in some cracking of the concrete along the path of the pipes where they are at their hottest.
  • Slab layers: The slab at the Sweeton House is a 3″  base slab with a 3/4″ to 1″ topping of very fine aggregate (coarse sand at best) onto which the colorundum pigment was applied. Cracking could be related to differing rates of expansion and contraction of the slab layers, but in this case cracking would probably be all around the pipe rather than on the bottom surface.
  • Seismic: In areas with seismic activity Seismic cracking could be due to this source.

April 2017: Bedroom 3 to Workshop Riser Pipes

This portion of the “saga” occurred over an extended period of time, so it is presented as a “diary” of sorts. We were afforded the luxury of time as this failure occurred toward the end of the heating season in New Jersey.

Friday, April 7, 2017 was not a good day at the Sweeton house.
I went into the workshop and found water in the unit grooves of the concrete…it was not storm water, and I determined it was not a water supply leak to the clothes washer or powder room. I shut down the boiler and its make up water supply, then I watched the boiler water pressure gauge…it slowly went to zero. When makeup water was turned back on, I could hear water going into the system and the gauge needle rose. I kept wet-vacuuming the workshop until the seeping stopped…about 3 gallons total came out before it stopped.

I called my heating contractor. We needed to find the bleed. The boiler had not run for a few days due to warm temperatures; it ran for the first time that morning. This is the first I had seen water coming from beneath the powder room, and I am in the workshop a few times each day. I’m not sure how long this may have been leaking…it couldn’t have been too long as I feel the supply pipe regularly where it meets the boiler to check if it is cold…it was cold today (a warm pipe has no cold fresh water running through it; a cold pipe when the boiler is running has been cooled by new water). Make up for a pinhole may not have been perceptible, however.

I suspect the water would not bubble up from beneath the workshop floor, but rather, it may be seeping down from the master bedroom, following the either the supply or return riser pipe to the 24″ lower workshop floor and seeping out through the wall. I don’t know where the 2 drop pipes are, but I suspect they are in or behind the common wall with the master bedroom in the vicinity of the powder room, as that is where the water seeped from.

We have backup heat from Mitsubishi mini-splits in the living, dining, and master bed areas…not efficient, but they can prevent freezing at least in New Jersey April weather. I don’t want to run the boiler on just the living room loop at this time because it is not isolated by valves…we may need to add those. Isolation valves are at the returns to the boiler and either side of each circulator pump.

Our pipes are welded steel.

Apr 10, 2017
I’m open to all ideas, but I’m reluctant to attempt a non-mechanical fix, fearing that it may be temporary and it’s eventual failure might occur when I’m not at home, potentially resulting in a 24″ deep flood, destroying my recent fit out of the workshop. I’m not happy about the current situation, but I am thankful I was able to address the situation almost as soon as it began.

My focus now is to map the as-built condition of the system as best I can from old correspondence, sketches, and my observations over the last 9 years, followed by infrared or ultrasound mapping as needed to pin point the leak. Once the leak is found, the floor will be opened and the pipe spliced with heavily primed like material and the floor patched.

I haven’t decided which is preferable: removing the powder room floor and its framing to cut a floor that will not need to be color matched when it is replaced, or removing a square or squares in the bedroom that will need to be carefully matched. Either way this will be expensive, but I don’t want to give up the pleasurable feeling and elegant simplicity of radiant heating in the house.

Apr 11, 2017

I have been in contact with Roland Reisley, he has been very supportive. Roland wrote an essay about his experience which is included as a part of this collection of articles on radiant heating systems:

Reisley House - Finding leaks by Ultrasonic Detector - and by Ear

Apr 12, 2017

From what I have learned about the system in my house and the steel or iron pipe systems in similar installations I have inquired about, is that the corrosion of the pipes rarely comes from within, but rather from the outside. Interactions with dissimilar metals or repeated wetting of the outer exposed surface of the pipes from groundwater, stormwater, condensation, or leaking sub-slab domestic or sanitary piping are the causes most often found for leaks. Old, funky, well air-bled, fully pressurized systems do not tend to promote corrosion of steel pipes, however if many tiny pinhole leaks begin to allow oxygenated or mineral laden make up water to enter the mix regularly, interior corrosion begins to be more of a possibility. Expansion/contraction cycling of imperfect welds (particularly at bends) over the course of many decades can lead to leaks as well.

Apr 14, 2017
Based on my conversations and reading of Roland Reisley’s experience, I’m tracking down ultrasonic testing firms local to us…there seem to be a number in and around NYC, though I may have a candidate in Trenton NJ. One would think with the petrochemical and chemical industries and their miles of piping around Philly and Wilmington, I’d find more…I’m beginning to think the industrial process companies do this work in-house or use a different process.

Apr 24, 2017

I had a leak detection company out to the Sweeton house on Friday. Rather than ultrasonic, the detection company used acoustic detectors (teacup sized mics set on the floor, wired to a belt mounted amplifier, heard through aviation type headphones) and the leak was located.
We did not turn on the pumps, just the water make up valve to start, as pump noise sometimes can be heard more than the leak. Once the pressure got near operating level, the leak became audible with the instruments. The sound was loudest and clearest in the southwest corner of the master bedroom. It sounded like a repetitive gurgle.

I have a concrete cutter scheduled for Saturday, May 6th to open a 2’x 4′ section of the floor (the loudest point detected is within a half unit at the south wall of the room) cutting along the unit line about 2″ deep, followed by chipping. We’ll see how it goes. Hopefully there will be good splice-able pipe within that removal section or another unit will need to be cut out. In any case, it will be interesting to see what the piping material and substrate looks like and to attempt to diagnose the cause of the leak.

As the cutting will seriously dust up the room, we have decided to move all of our stuff out of the MB and bedroom #2 and complete our restoration of them as part of the effort (ceilings, roof insulation, window screens) …lemonade out of lemons so to speak.

Apr 26, 2017

At our house the pipes are organized as two not completely isolated loops each controlled by its own thermostat and circulator pump. Contrary to the house’s drawings, which show 2 simple loops, each loop consists of a series of panels (zigzagged runs of pipe that approximate an area of about 60-100 SF depending on room configuration) that are fed and returned by leader pipes that run more or less at the perimeter of the house. (see the diagram below) This placement keeps the floors relatively uniform in temperature when the system is operating. All pipes are filled with water, but the flow of freshly heated water to each panel can be controlled by flow valves at the return manifold.

Mr. Sweeton wrote that he all but closed a valve for the panel in bedroom #2, where the bedroom wing thermostat is located, to prevent overheating of the master bedroom…..all rooms have been comfortable, but there is a chilly spot on 2 full units and 2 half unit floor squares in B2 and its closet that one notices when barefoot.

May 02, 2017

I just received my April water bill. Our water usage is always very consistent, but this month it was 3000 gallons over normal. It would appear there was a sudden change within this billing period…either a slow leak suddenly grew, or a leak suddenly occurred. The meter measures in units of 1000 gallons; water bills typically indicate 3 units or 3000 gallons…on a rare month,  if we have been away for a week or so, it will be 2 units. This month the reading was 6 units, or 6000 gallons.

May 03, 2017

We found the location, we just (yeah, just) have to assess the precise condition and how best to splice.

Our hydronic system is pretty typical…it is a closed system. The radiant floor piping is connected to and filled via the domestic water system, but separated with a backflow preventer. It has an automatic fill valve which adds water as needed to maintain pressure. Provided it is not leaking, the system will not need much if any water, but it may loose a little via the air bleed valves, which the auto fill tops off. The water in the system should be old and funky, not fresh, oxygenated, and mineral rich. If feeding from a well, a water softener or other filtration system should treat water going into the auto fill. As of 1963, there was city water at Sweeton.

When I discovered the leak, I shut down the boiler which cut power to the circulator pumps, AND I shut off the valve feeding domestic water to the automatic fill. Had I not shut the valve, the system would have kept adding domestic water to maintain operating water pressure thus perpetuating flow at the leak.

May 08, 2017

The floor was opened this Saturday to assess the condition of the piping and exact location of the leak:

The floor is covered except for the unit to be removed, and a dust enclosure is made to reduce dust travel.

Cutting starts...given the position of the pipe found in 2015 at the window wall, we are cutting no deeper than 2".

After cutting on the unit score line and chipping of the 1-1/2" top coat, the 3" structural slab is chipped out using a large electric chipper. A small chipper and the large chipper were used depending on how fine the spot to be chipped was. No pneumatic hammers were used.

Mostly open...the large pipe rounding the corner is a leader from which the field panel pipes branch off. The branch dropping is the supply to the work shop (24" below the main house floor). The pipe against the wall that comes up from the deep drop is the return from the workshop which is a home run to the bleed vent in the kitchen and back to the boiler. The master bedroom floor panel branches off horizontally then coils through the room to its return. The main level pipe is very clean with no sign of corrosion...that's a good thing.

The drops to the workshop are excavated, and corrosion is found starting at about 16" below the main level; not good.

We know from closing the house's original well that the water table is about 35' below grade, and that the surrounding soil is variable in its percolation rate...some drains rapidly, some drains very slowly. The septic system was done in 2008 as a 24'x20'x9'deep sand pit to account for slow absorption. In periods of long duration rain, or very heavy rainfall, a small amount of water (less than a pint) has appeared in the NW corner of the workshop floor...when we bought the house, any rainfall would cause significant infiltration...that was stopped by excavation of the exterior, waterproofing, adding drainage pipe, and placing a ground gutter to remove dripped roof water from the area. The water seen in the 24" deep hole seems to be trapped stormwater or leaked boiler water.

The leak is on the supply riser about 16" below the main floor. An 1/8" hole is visible where a scaled bit blew out.

When the leak was noticed on April 7, there had been two days of heavy rain, and boiler water appeared on the workshop floor. Two weeks ago when the acoustic leak detectors were on site, it had not rained for a week, and the boiler water was on for one hour with no appearance of water on the floor at all. This Saturday, after a 2″ storm on Friday, water appeared on the workshop floor 5 minutes after opening the make up water valve. This indicates perched water which accumulates rapidly and dissipates gradually….in other words, the lower radiant pipes have been subjected to repeated wetting and drying over the course of 66 years. The corrosion seen on the lower bit of the riser is likely present on some or most of the workshop in-floor panel tubing. I’m not going to mess with it; I will not dig up the workshop and dismantle what I just installed, nor will I overpour the existing slab, working around new wood partitions to install pex. Wright originally indicated a radiator beneath the windows…it was just a workshop on a tight budget. The Sweetons extended the in-floor during construction when a little more money was available…it worked for 66 years, but its time is up.

We will abandon the workshop in-floor piping in place, and reroute the water to the room using the same supply and return branches (which are in fine shape) to cast iron baseboard to be run along the south window side of the workshop beneath the powder room raised floor and on the base of the wall the behind the desks…(fin tube baseboard will not work efficiently with our boiler temperature). The main floor of the house will remain radiant floor, the workshop only will be radiant baseboard.

May 10, 2017

I checked the Sweeton Plumbing and Heating sheet, and was able (with eyeglasses and a magnifying glass) to read the originally spec’d radiator for the workshop: “Sterling-fin type 2″, 6′-0″ L, 3 units high”.
If “3 units high” refers to concrete block courses, that would be 24″ tall, or 24″ above the floor. Sterling was founded in 1946 and exists today. The company stated the appearance of their Versa Line is virtually unchanged from the product they offered in the 1950’s:

Wright was definitely going utilitarian in the workshop.
I’m going to stick with the cast iron baseboard as that is functionally the best choice for the operating temperature of our boiler water…it also has a lower profile to boot.

June 19, 2017
My heating contractor was on site cut loose the damaged pipe and set a temporary pex connector pipe to bridge between the old supply and return pipes that served the workshop.

The supply and riser pipes to the lower level are cut.

The remove riser pipes: the leak holes of supply pipe on the left are visible; the rusted though not yet leaking return pipe is on the right.

The return riser pipe on the left; the supply riser pipe on the right. The supply was rusted so thin it was just broken loose to remove it, hence the ragged edge. It is notable to see how thick these schedule 80 pipes are and how much they can rust down when exposed to repeated wetting. The inner surface of the pipes where not breached have no corrosion...damage came from outside in. Where pipes were set in crushed stone at the upper and lower levels, the corrosion ranged from non-existent to moderate...where set in raw soil (clay-sand called "marl" locally) the corrosion was most severe.

With the risers cut, a temporary flex (pex) connector was set to allow the water to be turned on to confirm no other unknown riser pipes exist feeding the lower level. With water turned on, no water bled from the lower level...this meant we have the only two risers accounted for. Pressure held and no further water was being added by the auto feed valve. The leak is isolated.

September 2017
After many calls and many “the job is too small for me to justify the trip, I found a welder. Valves supplied by my heating contractor were set in place. Pex shunt lines were run through the wall to connect to the new baseboard in the workshop. With this work completed, the forming of a valve box and patching of the floor could proceed. I sprayed the exposed pipe with rust inhibitor prior to the next steps.

The steel valve box is fabricated and set. Gravel is in the process of being placed in the excavation to the specified depth.

The concrete was mixed and poured. The color was applied as a powder in the final trowel finishing. Scofield Lithochrome Color Hardener was the product used.  (Scofield is now part of the Sika family of concrete products. Their website:

PLEASE NOTE: THERE WAS NO SINGLE “CHEROKEE RED” COLOR USED IN THE USONIAN HOUSES OF FLW. All houses and buildings for the era seem to have similar, and at times not so similar, color tints in their concrete floor slabs. In the case of the Sweeton house, our concrete finisher (and finish carpenter, and “concrete removal surgeon”, Ean Frank) mixed two or three Lithochrome red tints and made a few samples for us to make the closest match…the stock colors straight out of the container were not matches at our house.

As a finishing touch, Ean made an access panel to the valve box using fiber cement board with a metal angle edge to hold a gypsum concrete surface tinted with the Lithochrome used on the surrounding floor.

Following the setting of the new baseboard radiant heating in the workshop, it was time to restart the system. Though we had been cautious and filled up the system to get most of the air out and check for leaks when the temporary connector pipe was set, we had not wanted to tempt fate by running the pumps with the temporary connector in place. Now it was time to get the system running again. To do this we started up the system and I nursed the air bleed valves, which had a tendency to spit some water as the few remaining air pockets burped out. After about an hour all spitting stopped.

The bleed valves are set at the high point of the pipe array, in our house a small heater basement (8’x16’) locates the boiler under the kitchen, thus the bleed valves were set under a removable panel in the broom closet cabinet in the kitchen. The little caps on top of the valves should be loose: loose enough to let air escape, but not so loose to as to fall off. I turn the cap on to finger tight, then back turn 2 turns to keep it loose enough.

Please Note:The provision of this information or mention of a specific product or products does not constitute endorsement, recommendation, preference or approval by the Conservancy.

Radiant Heating 101 - Introduction Page

Posted April 3, 2023