300 Litres of Oil a Winter Saved by a Nest Learning Thermostat on a 1930s Detached House

A 1930s detached house of roughly 130 square metres burning around 1,800 litres of oil a season is a familiar pattern for solid-feeling brick construction with an unmodulated boiler. When the annual delivery figure falls by close to 300 litres after a Nest Learning Thermostat is installed, the control looks like the obvious cause. In this case the Nest stopped the boiler heating the whole house to 21 degrees for hours while only the bedrooms were being used, then learned that the fabric held enough heat to shut the burner down 40 minutes before the setpoint was reached. Those two behaviours account for most of the saving.

The rest came from fabric and reliability work completed within the same six months. Cavity injection, draught-stripped sash windows, insulated reveals, and a condensate fix all changed how the house behaved during the heating season. The oil tank recorded one result from one boiler, with no separate meter for each measure.

Where the Nest earned its litres

The Nest infers return temperature through cycle timing and uses its True Radiant algorithm to anticipate the setpoint. On an oil boiler with no weather compensation, that anticipation has value because the burner fires at full output every time it lights. Cutting three or four avoidable firings a day on a system consuming maybe 2.5 litres per hour of burn is enough for the oil level to move differently over a winter.

Auto-Schedule also caught a standing habit: the heating was being left on through a midday absence. Once that pattern disappeared, the boiler spent fewer hours trying to hold a comfort temperature in empty rooms.

A wall’s U-value remains governed by its construction. A 300mm unfilled cavity in 1930s brick loses heat at a rate beyond the reach of a schedule change, so the Nest optimised boiler runtime against the heat loss that remained. The same device in a poorly insulated house of the same age often returns half the saving. The measured 300 litres here sits on top of cavity injection completed the same autumn.

Cavity injection and the condition of the wall

Cavity wall injection on a detached house of this footprint typically costs between 700 and 1,500 pounds, with access and fill material driving the spread. The 1930s build here had a real two-leaf cavity of around 50mm, wide enough for blown mineral wool, the standard fill used for that era.

The installer drilled a grid of 22mm holes at roughly 1 metre centres. Each port was injected until the cavity registered full pressure, then the holes were made good with mortar-matched plugs. The detached shell took most of a day.

The saving from that fill varied by elevation. The north and west walls, exposed to prevailing wind and receiving the least sun, saw the largest drop in heat loss.

A borescope inspection before the work confirmed that the cavity was clean of rubble bridging. That detail matters in houses of this age because dropped mortar snots across the gap are a common failure point. Where the cavity is bridged, injected fill can sit against a mortar bridge, allowing moisture to track across to the inner leaf.

The pre-inspection was the difference between a clean installation and the kind of cavity fill that produces damp complaints two winters later. In this house the fill went in cleanly. The inner-leaf surface temperature on the north wall rose enough to stop the seasonal condensation that had been marking the corner of the front bedroom.

The condensate run that froze

Before the fabric work, the boiler locked out twice in one cold snap. The cause was an external condensate run in 22mm waste pipe. At the first sustained sub-zero night, warm and damp condensate moved slowly enough to form an ice plug at the trap or the first outdoor bend.

The fix was a replacement external run in 32mm pipe, weatherproof foam lagging, and a shorter horizontal section so the fall to the drain was steeper than 2.5 degrees. A 32mm bore needs a much larger mass of ice before it blocks, and the steeper fall keeps condensate moving instead of pooling.

Each lockout left the house cold, required a manual reset, and encouraged recovery heating at 23 degrees. The Nest then had to relearn after those spikes. Removing the freezes also removed a failure mode during the coldest week of the year, exactly when the boiler was carrying the heaviest load.

Window reveals as cold bridges

The cavity fill left one visible weakness: the window reveals. In 1930s construction, the reveals are often solid brick returning into the cavity, which means the injected fill stops short and the reveal stays cold. A thermal camera showed the reveal plaster running several degrees below the adjacent wall, the first place mould appears even after a successful cavity fill.

The correction was thin internal insulation on the reveals using laminated board. A 20mm to 25mm Kingspan Kooltherm K7 board was bonded to the reveal, taped at the junctions, and re-plastered, raising the reveal surface temperature above the internal dew point. The trade-off was a slight reduction in daylight opening and a fussy detail around the existing timber sash box.

The decisive workmanship detail was the junction to the frame. The board had to sit tight without an air gap behind it. An air gap convects behind the board, leaving the cold bridge active under a warmer surface finish.

Draught proofing the sash windows

The original timber sashes were the largest uncontrolled air path in the house. A sash leaks at the meeting rail, down both staff and parting beads, and under the bottom rail; on a windy night the combined gap can equal a small open vent.

Proper draught proofing meant routing a groove into the beads and inserting a brush pile carrier, then fitting a compression seal at the meeting rail so the two sashes closed against each other. Done well, the windows still slide and still open. That is the point of keeping working timber sashes in service.

After the work, the air change rate in the front rooms dropped enough for the Nest’s recovery times to shorten. Less heat was leaving through the window perimeter, so the thermostat reached its target with fewer and shorter burner cycles.

Why the air source heat pump was deferred

An air source heat pump retrofit was quoted during the same survey and then left for later because the existing emitter sizing constrained the design.

Radiator output at lower flow temperatures was the limiting issue in the quote, which made the fabric measures and control changes the immediate work.

What the oil notes showed

The 300 litre saving was real and was measured against the previous three seasons of delivery notes on the same tank. The notes record the combined effect of the Nest, the cavity fill, the reveal insulation, the sash draught proofing, and the condensate repair. A single boiler gauge reports total demand from the heating system, so the litres appear as one blended figure.

With the fabric holding heat longer, a lower boiler flow temperature became plausible. The house had been running at 70 degrees, and around 55 degrees would be the target if the existing radiators could still hold comfort across the rooms. The delivery notes established the oil saving; they left the radiator margin unanswered.