Snow melt system field guide: hydronic and electric

What a snow-melt system is and why people install one

A snow-melt system is heating elements buried in a paved surface that melt snow and ice as they fall, so the surface stays clear without a plow, a shovel, or salt. The elements are one of two kinds. Hydronic systems run warm fluid through PEX or rubber tubing set in the slab. Electric systems run resistance heating cable or pre-spaced mat in the slab or under pavers. Either way the heat comes up from inside the pavement and the snow never gets a chance to pack down into ice.

The reason owners pay for it is usually not comfort. It is the slip and the liability. A clear entry, ramp, or stair that never ices over takes a whole category of fall claims off the table, and it keeps a building accessible in weather when a shoveled path would not be. The wheelchair ramp that has to stay usable, the hospital entrance that cannot close, the loading dock that has to keep running through a storm: those are snow-melt jobs because manual clearing cannot be trusted to keep up.

The other driver is salt. Chloride deicers eat concrete, corrode rebar and vehicles, and kill the landscaping the runoff touches. A heated surface needs none of it. That matters most around parking structures, new architectural concrete, and any place where salt-laden meltwater would track inside. Snow-melt is a sibling of in-floor radiant heating, and it shares the tubing, the manifold, and the low-temperature hydronic ideas covered in the radiant floor design guide, but the load and the controls are a different animal because the slab sits outdoors in the cold.

Where snow-melt gets used

Snow-melt goes where clearing by hand is unsafe, too slow, or impossible to keep up with. Steep driveways are common, because a plow struggles on the grade and an iced slope is dangerous to walk or drive. Wheelchair ramps and accessible entries are a frequent target, since a ramp that ices over is no longer accessible no matter what the drawing says it is. Walks, entries, and stairs near a door get heated so the first and last few feet a person crosses are never the slick part.

Commercial and institutional work pushes it further. Loading docks stay clear so forklifts and trucks keep moving through a storm. Hospital and fire-station approaches get it because the access cannot fail. Helipads and exposed pedestrian bridges get it because they are wind-swept, hard to reach, and unforgiving if iced. No-salt zones around parking decks and signature concrete use it to keep chloride off the structure.

On the industrial side, snow-melt shows up at data-center generator yards and equipment ramps, where crews have to reach gear in any weather, and at backup-power and switchgear approaches that have to stay walkable during an outage. The pattern across all of it is the same: heat the surface where a fall, a delay, or a closure costs far more than the energy to keep it clear.

What is the difference between hydronic and electric snow melt?

Hydronic snow-melt circulates a heated glycol-and-water mix through tubing in the slab, fed by a boiler or another heat source through a heat exchanger. Electric snow-melt passes current through resistance cable or mat in the slab, with no fluid, no boiler, and no pump. That one difference, fluid versus current, drives every other decision: the size of area that makes sense, the install effort, the equipment, and above all the cost to run it.

Hydronic wins on large areas and on operating cost. Burning gas to make heat is cheaper per unit than buying the same heat as electric resistance, so a long driveway or a big dock run by a boiler costs less to operate over a winter. The price is complexity: a boiler, a heat exchanger, glycol, a pump, a manifold, and the mechanical room to hold them. Electric wins on small areas and simple installs. A set of stairs, an entry pad, or a short walk in electric is a cable, a sensor, and a circuit, with no fluid to manage and nothing to freeze. It costs more to run because resistance heat is expensive, but at a small size the simpler install and lower first cost usually settle it.

The rough rule the trade uses: small and simple leans electric, large and run-often leans hydronic. The crossover is a money question, not a physics one, because both deliver the same heat to the slab. Run the operating cost over the expected hours, not just the install bid, before the choice is locked.

The hydronic system: tube, boiler, glycol, heat exchanger

A hydronic snow-melt system is four parts working together. The tubing, usually oxygen-barrier PEX or a rubber hose rated for burial, is laid in the slab in loops off a manifold. The heat source, a boiler or another supply, makes the heat. A heat exchanger sits between them, keeping the outdoor glycol loop separate from the boiler water. And a pump and manifold move and meter the glycol through the loops so the whole surface heats evenly.

The reason the heat exchanger is there, and not just a direct connection, is the glycol. The snow-melt loop has to be antifreeze because it lives outdoors in freezing conditions. The boiler and the rest of the building hydronics usually do not, and you do not want to fill an entire heating plant with glycol to protect one outdoor loop. The exchanger isolates the glycol to the slab loop only, so the boiler side stays plain water and the antifreeze cost and maintenance stay contained to the part that needs it.

Tube layout, loop length, and equal loops follow the same hydronic logic as an indoor radiant floor, covered in the radiant floor design guide, and the same loop carries an expansion tank and an air separator sized per the expansion tank and air separator guide. What changes outdoors is the load and the fluid. The supply temperature runs higher than an indoor floor, commonly in the range of 100 to 150°F at the slab, because the system is fighting an outdoor design temperature instead of warming a room. Confirm the supply temperature, the loop length, and the delta-T against the design and the tubing manufacturer's data.

Do snow melt systems need glycol?

A hydronic snow-melt loop needs glycol. This is not a place to save money or simplify. The tubing sits in an outdoor slab that is, by definition, below freezing whenever the system matters, and any plain water left in that loop will freeze, expand, and split the tube or a fitting. Then you have a burst loop inside a concrete slab, which is a demolition job to fix. The antifreeze is the one thing on the system that has no acceptable substitute.

The fluid is almost always propylene glycol mixed with water, not the ethylene glycol used in cars, because propylene is the low-toxicity choice for anything near people, food, or potable systems. The concentration is set for the climate. A common mix lands near 50 percent glycol, which protects well below any winter design temperature, but the exact percentage is a design call: too little and the freeze protection is short, too much and the fluid gets thick, pumps harder, and carries less heat. Use an inhibited heat-transfer glycol, set the concentration to the manufacturer's freeze-protection table for your design temperature, and confirm it with a refractometer, not by eye.

One more reason the heat exchanger earns its place: it keeps the glycol in the small outdoor loop instead of the whole plant, so when the fluid has to be tested and eventually changed out, you are servicing one loop, not the building. Treat the glycol as a maintained fluid with a service life, not a fill-and-forget. The inhibitors that keep it from turning acidic and corroding the system wear out, and spent glycol is how a loop rots from the inside while it still has antifreeze in it.

The heat source and the heat exchanger

The heat source can be a dedicated boiler for the snow-melt loop or a shared boiler that also serves the building, and either way a heat exchanger separates the glycol from the boiler water. A dedicated boiler is simpler to control and stage for the big, intermittent melt load. A shared boiler saves equipment but has to be large enough to carry the snow-melt demand on top of whatever else it heats, and that demand is not small.

Sizing is where snow-melt surprises people. The melt load per square foot runs far higher than comfort heating, so a modest-looking driveway can ask for more output than the house it sits in front of. The boiler, the heat exchanger, and the pump all have to be sized for that peak, and the heat exchanger in particular has to move the full design heat across the glycol-to-water boundary without choking the approach temperature. An undersized exchanger starves the slab no matter how big the boiler is.

Size the heat source to the calculated melt load for the area, the class, and the local design conditions, then add the exchanger and pump losses on top. Confirm the boiler output, the exchanger capacity, and the supply temperature against the design and the manufacturer's data. The companion radiant guides cover the pump and the loop accessories that ride along with the heat source.

The electric system: cable, mat, voltage, and circuit

An electric snow-melt system is resistance heating cable embedded in the slab, or a pre-spaced mat that fixes the cable to a set spacing so the crew just rolls it out. Current through the cable becomes heat, the heat goes into the slab, and the snow melts. There is no boiler, no glycol, no pump, and nothing to freeze, which is exactly why electric is the easy answer on small areas and retrofits.

The catch is the electrical load. Snow-melt cable draws real power, so the circuits are large and there are often several of them. Residential systems commonly run 120 V for small areas and 240 V for larger ones, with 277 V and 480 V used on commercial work to carry more area per circuit and hold the current down. A driveway-sized electric system can need a dedicated subpanel and a contactor, not a spare breaker. The watt density, the voltage, and the breaker all come from the cable manufacturer's layout for the specific area, and the wiring follows the electrical code for the circuit.

Electric does not usually need under-slab insulation the way a hydronic slab does, since the cable puts its heat right at depth and the response is faster, but the down-loss is still real and insulation still helps in cold ground. Confirm the watt density, the spacing, the voltage, and the circuit sizing against the manufacturer's design and the adopted electrical code before the cable goes in, because once it is in the slab there is no adjusting it.

How much heat does a snow melt system need?

Snow-melt asks for a high heat output, much higher than heating a room, because the surface is fighting outdoor cold, wind, and the heat it takes to actually melt falling snow and evaporate the meltwater. Where a comfort floor might run 20 to 40 Btu/hr per square foot, a snow-melt surface is often designed for well over 100 Btu/hr per square foot, and cold, windy, or high-snowfall sites push it higher. Electric systems express the same thing as watt density, commonly in the range of 30 to 50 watts per square foot at the slab depending on spacing and climate.

The load is not one number. It depends on the snowfall rate, the outdoor design temperature, the wind exposure, and how much of the falling snow the owner wants melted on contact. A windy helipad in a cold climate and a sheltered residential walk in a mild one are not the same design, and sizing both off a single rule of thumb is how you end up with a surface that cannot keep up in a storm or one that costs a fortune to run.

Treat the heat-flux and watt-density figures here as starting ranges, not design values. The real number comes from the ASHRAE snow-melting load method for the class and the local climate, or from the cable or tubing manufacturer's design for the specific site. Hedge to those sources, and size for the storm the surface actually has to beat, not the average day.

ASHRAE snow-melt classes

ASHRAE sorts snow-melt designs into classes by how completely and how fast the surface has to stay clear, and the class drives the load. The common framing runs Class I for residential surfaces, Class II for commercial walks and entries, and Class III for critical, high-priority surfaces where snow cannot be allowed to accumulate at all. The higher the class, the more of the falling snow has to melt on contact, and the more output the design carries.

The class is a design decision, not just a label. A Class I residential driveway can tolerate some snow sitting briefly before it melts, so it can be sized leaner. A Class III surface, a hospital entrance or a critical ramp, is expected to stay nearly clear all through the storm, which means higher heat flux, often some idling to keep the slab warm ahead of the snow, and controls tuned for fast response. Picking the class sets the load, the equipment, and the operating cost all at once.

Match the class to what the surface actually has to do, then pull the design heat flux for that class and the local climate from the ASHRAE snow-melting method or the manufacturer's design. The class names and the load percentages they imply are ASHRAE's framework, so confirm them against the current handbook rather than a remembered figure.

How does a snow melt system know when to turn on?

An automatic snow sensor turns the system on, and it is the part that makes snow-melt affordable to run. The sensor watches for two conditions at once: moisture, and a temperature low enough for that moisture to be snow or ice. Only when both are true, it is precipitating and it is cold, does the controller fire the boiler or close the contactor. Rain in October and a cold dry night in January both leave it off, because neither is a snow event.

There are two sensor styles. A slab sensor sits flush in the pavement and reads surface temperature and moisture right where it matters, but it has to go in before the pour. An aerial sensor mounts on a post or wall above the surface and reads air temperature and precipitation, responding fast to weather and not requiring embedment. Some designs use both. After the snow stops and the moisture element dries, an adjustable hold-on timer keeps the system running a while longer to finish drying the surface so the last of the meltwater does not refreeze into a sheet of ice.

The alternative to a sensor is a person flipping a switch, and that is how energy bills run away. A manual system left on for a forecast that misses, or left on for days because nobody remembered to shut it off, burns far more than the storm needed. The automatic sensor is what limits runtime to the hours that actually have snow on the ground, which is most of why a well-controlled system costs what it should and a manual one does not.

Idling versus cold-start, and slab response time

A snow-melt slab is heavy, and concrete takes hours to warm from cold. That mass is the central tradeoff in how the system is controlled. Start the system cold when the snow begins and it is slow off the line: the slab has to come up to temperature before it can melt anything, so snow accumulates during the warm-up and a fast storm can get ahead of it. That is acceptable for a residential Class I surface and a budget that favors low standby cost.

Idling is the other approach. The controller keeps the slab held at a low warm setpoint whenever cold weather is likely, so when snow falls the surface is already primed and melts on contact with no lag. Idling buys fast response, which is what a critical Class III surface needs, but it costs energy to hold the slab warm through cold spells that may never bring snow. The control strategy is the balance point between those two, and a smart controller will use weather data and slab temperature to idle only when a storm is genuinely likely rather than all winter.

Match the strategy to the class. A residential walk can cold-start and accept a little lag. A hospital ramp idles because it cannot. The controls and the snow sensor together decide how much of the slab's response you pay for in standby energy versus how much snow you tolerate at the start of a storm.

The slab: depth, spacing, and fastening

The tube or cable goes into the slab at a controlled depth and spacing, and both decide whether the surface heats evenly. Set the element too deep and the response slows and the surface temperature drops. Too shallow and you risk hitting it with the saw or a fastener, and you can get hot spots. The element is fastened down so it cannot float up during the pour, tied to the reinforcing mesh or set in clips, with the leads and the cold ends brought out where they can be terminated and the splices kept out of the heated field where they belong.

Spacing controls the even melt. Run the tube or cable too far apart and you get cold stripes between the loops, where snow sits unmelted in lines that map the layout exactly. Tighter spacing gives a more uniform surface but more material and, for electric, a higher watt density. Electric systems trade spacing for watt density directly: closer cable means more watts per square foot, which is why a colder climate calls for tighter spacing. The spacing and depth come from the manufacturer's design for the class and climate, so follow that layout rather than eyeballing it.

Document the as-built layout before the pour covers it. A photo of the loops and cable in place, with spacing and the location of leads and sensors, is the only record anyone will have once the concrete goes down and the element is invisible.

Insulate under the slab or you heat the ground

Put insulation under a snow-melt slab. Without it, a large share of the heat you are paying for goes straight down into the ground instead of up to melt the snow. The system is trying to heat one surface, and an uninsulated slab heats two: the top you want and the earth you do not. That down-loss is wasted energy on every hour the system runs, for the entire life of the installation.

The loss is worst on hydronic systems and on cold or wet subgrades, which pull heat away faster. Rigid insulation board under the slab, and often around the slab edge, turns the heat back up where it belongs and cuts the down-loss sharply. The edge is easy to forget and it matters, because the perimeter of an outdoor slab loses heat sideways to the surrounding cold ground and frost. Electric systems are more forgiving since the cable sits near the surface and heats faster, but insulation still pays back in cold ground.

This is the energy decision that is invisible after the pour and impossible to fix later. Skip the under-slab insulation to save on board and you pay for it in higher operating cost every winter the system runs, with no way to add it short of tearing out the slab.

Zoning the surfaces

Heat the surfaces that need it on their own zones rather than as one big circuit. A driveway and the walk beside it have different priorities, different areas, and often different schedules, and putting them on separate zones lets each one run only when it has to. On a hydronic system that means separate manifold loops with their own actuators; on electric it means separate circuits and contactors.

Zoning also keeps the load manageable. The whole melt load hitting at once is what sizes the boiler or the electrical service, and being able to stage zones, or prioritize the critical entry over the broad driveway, can hold the peak down. Lay out the manifold so each zone is balanced and the loops within it are close to equal length, the same equal-loop logic the radiant floor design guide covers, so no single loop starves while another runs short.

Asphalt, concrete, and pavers

The surface material changes how the element is installed and how the heat behaves. Concrete is the most common host: the tube or cable ties to the reinforcing and the slab gets poured over it, giving a solid, even thermal mass that holds and spreads heat well. The mass that makes concrete heat evenly is also what makes it slow to respond, which loops back to the idling-versus-cold-start decision.

Asphalt is workable but harder on the element, because hot asphalt goes down at a temperature that can damage tubing or cable if the wrong product or sequence is used. Asphalt snow-melt typically uses a protective sand or base course and products rated for the placement temperature, and the install sequence matters more than with concrete. Confirm the element is rated for asphalt and follow the manufacturer's placement method.

Pavers are the retrofit-friendly option and the trickiest to detail. The element, usually electric cable or mat, sits in a sand or screed bed under the pavers, and the paver thickness over the element is limited so the heat can reach the surface. Manufacturers cap how thick a paver can sit over the cable, commonly a couple of inches, because a thick paver insulates the heat away from the top. Keep the pavers off bare cable, embed to the specified depth, and hold the paver thickness to the manufacturer's limit.

How much does it cost to run a snow melt system?

Operating cost comes down to how much heat the surface needs, what you pay for that heat, and how many hours the system runs. The heat load is high, so the cost per hour is real on any snow-melt system. The fuel is where hydronic and electric split: gas-fired hydronic heat is cheaper per unit than electric resistance, so a large area run often is markedly less expensive to operate on hydronic, which is the main reason big surfaces go that way despite the heavier install.

The hours are where the controls earn their cost. A system on an automatic snow sensor runs only when it is snowing and cold, which on most sites is a modest slice of the winter. The same system on a manual switch, or idling all winter, can run an order of magnitude more hours and cost accordingly. This is why the sensor and the control strategy matter as much as the fuel: the cheapest hour is the one the system never ran because there was no snow on the ground.

There is no single dollar figure, because it scales with area, climate, fuel price, class, and idling strategy. The honest answer to a client is a range tied to those inputs, run from the design load and the local rates, with the point made plainly: insulate under the slab, hold the right class, and run it on a snow sensor, and the cost stays reasonable. Skip those and it does not.

New pour versus retrofit

A new pour is the easy case. The element is laid, fastened, and documented before the concrete goes down, the insulation goes under it cleanly, and the depth and spacing are exactly what the design called for. If snow-melt is on the table at all, getting it into a new slab is far cheaper and better than adding it later.

Retrofit into an existing surface is harder and the options are limited. You can saw-cut grooves into the existing pavement and set cable into them, you can put down a heated overlay on top of the old surface, or you can tear out and repour. Saw-cutting and overlays both compromise the ideal depth and the under-slab insulation you would get in a fresh pour, so the response and efficiency are usually a step down from new construction. Pavers are the friendliest retrofit, since the bed can be opened, the element laid, and the pavers reset. Price the retrofit honestly against a repour, because the cut-in version sometimes costs nearly as much and performs worse.

Commissioning and startup

Commissioning a hydronic snow-melt system starts with the fluid and the pressure. Confirm the glycol concentration with a refractometer against the design freeze protection, pressure-test the loop before and after the pour to catch a tube damaged during concrete placement, and purge the air out so every loop flows. A loop with trapped air or a buried leak found after the slab cures is the worst-case discovery, which is exactly why the pressure test happens before the pour as well as after.

Then prove the controls. Trigger the snow sensor, manually or with the test function, and confirm the boiler fires or the contactor closes and the right zones energize. Check the hold-on timer runs the system past the end of the event. On electric, verify each circuit draws its rated current and that the contactor and any subpanel are not overloaded.

Finish with an even-melt check in a real or simulated event. Watch for cold stripes that map the spacing, cold corners at the slab edge where insulation may be short, and any zone that lags badly. The surface should clear uniformly. Stripes or cold patches at startup are the layout or the insulation telling you something, and that is the time to find it, while there is still a record of what went in.

Maintenance

The glycol is the maintenance item people forget. Test it on a schedule, because the inhibitors that keep the fluid from going acidic wear out over years even though the antifreeze protection looks fine, and spent glycol corrodes the loop and the heat exchanger from the inside. Check the concentration and the inhibitor condition, top up or change the fluid per the manufacturer's interval, and treat a changeout as routine, not an emergency.

The sensor and controls need a look before each season. Clear debris and leaves off an aerial sensor, confirm a slab sensor still reads and responds, and test that the system fires on a simulated event before the first real storm rather than during it. A sensor that fails dirty either runs the system constantly or never runs it at all, and both are expensive in different ways.

On the heat source, the boiler and pump get the same service any hydronic plant does, plus a check that the heat exchanger has not fouled on the glycol side. On electric, there is little to service beyond the controls and the connections, since the cable has no moving parts, but a circuit that has started tripping is worth tracing before winter.

Safety, accessibility, and the slip

The point of most snow-melt is keeping people upright. A heated entry, ramp, or stair removes the ice that causes falls, and on an accessible route that is the difference between a ramp that works in winter and one that is a hazard the moment it snows. An ADA ramp or an accessible entry that ices over is not accessible, and snow-melt is one way to keep it usable through a storm when a shoveled path cannot be relied on.

The liability follows the slip. A surface that stays clear automatically takes the fall claim, the closed entrance, and the failed accessible route off the table in the weather when they are most likely. That is the case that justifies the operating cost on a commercial or institutional building: the energy to keep the entry clear is cheap next to one serious fall or a day the building cannot open. Keep the heated zone covering the full path a person actually walks, including the landing at the top of a stair and the bottom transition, not just the obvious middle, because the edges are where the ice that trips people survives.

Field checklist

Common mistakes

• Filling a hydronic snow-melt loop with plain water instead of glycol, so it freezes and bursts inside the slab.
• Leaving out the under-slab and edge insulation, so a large share of the heat goes into the ground every hour.
• Sizing the heat source or circuit for comfort heating instead of the much higher snow-melt load.
• Skipping the automatic snow sensor and running on a manual switch, so the system runs for days nobody needed.
• Picking the wrong idling or response strategy for the class, so a critical surface lags or a residential one idles all winter.
• Spacing the tube or cable unevenly or too wide, leaving cold stripes that map the layout.
• Choosing electric on a large area where hydronic would have been far cheaper to run, then living with the bill.
• Setting pavers too thick over the cable or resting them on bare cable, so the heat never reaches the surface.

What to document

A snow-melt system disappears under concrete the day it goes in, so the record made during the install is the only thing anyone has to work from later. Capture what was buried, what fluid is in the loop, and how the controls are set, because every one of those is invisible once the slab cures.

Standards and references

The design framework for snow-melt is the ASHRAE Handbook, HVAC Applications, in the chapter on snow melting and freeze protection. That is where the load method lives: the heat-flux calculation for the surface, the snow-melt classes, and the climate data that drive the design output for a given site. The heat-flux ranges and class definitions in this guide are that ASHRAE framework, so size from the current handbook for the actual location rather than a remembered figure.

The equipment numbers come from the manufacturers. The boiler and heat exchanger output, the tube and its supply temperature and loop limits, the electric cable's watt density, voltage, and spacing, and the snow sensor and controller's setpoints all follow the specific product's design and instructions. The glycol concentration follows the heat-transfer-fluid manufacturer's freeze-protection table for your design temperature, and electric circuits follow the adopted electrical code for the conductor, breaker, and grounding.

Three things carry the system, and they are worth stressing because they are the three most often shortchanged. The glycol, which is non-negotiable on any outdoor hydronic loop. The under-slab insulation, which decides whether the heat goes up or into the ground. And the automatic snow sensor, which decides whether the operating cost is reasonable or ruinous. Confirm the loads, the classes, and the glycol percentage against ASHRAE and the manufacturer's data, and let the project specification and the adopted code control where they are stricter.

Units, terms, and conversions

Snow-melt borrows units from both the hydronic and the electric worlds, so the same design can read in heat flux on one sheet and watt density on another.

Hydronic load is given as heat flux in Btu per hour per square foot, sometimes written W per square meter in metric sources. Electric load is watt density in watts per square foot. Supply temperature is in °F or °C, and delta-T is the temperature difference between supply and return. Glycol concentration is a percentage by volume, and the freeze point that goes with it comes from the fluid maker's table. Class refers to the ASHRAE snow-melt classification that sets how completely the surface stays clear.

Heat flux

Snow-melt design load in Btu/hr per square foot, the heat the surface must deliver

Watt density

Electric snow-melt load in watts per square foot, set by cable spacing and climate

Glycol

Propylene-glycol-and-water antifreeze that keeps the outdoor loop from freezing

Heat exchanger

The barrier that isolates the glycol snow-melt loop from the boiler water

Idling

Holding the slab at a low warm setpoint so it responds fast when snow falls

Snow-melt class

ASHRAE classification (I, II, III) for how completely a surface must stay clear

Snow sensor

Control that runs the system only when moisture and cold are both present

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• How does radiant floor heating work?
• What does a hydronic expansion tank do?
• What is a unit heater?
• How do you balance a hydronic system?
• Why do boilers need water treatment?