For almost 200 years, New Court in Trinity College, Cambridge, has been a place of study and contemplation. Its neo-Gothic walls have been home to the likes of the poets Tennyson and Hallam and even the current Prince of Wales. However, after two centuries as a student residence, the accommodation in the four-storey courtyard fell far short of current regulatory standards and present-day expectations of comfort and amenity.
The college wanted to refurbish the notoriously draughty Grade I listed block so that it could continue using it for another 200 years, but the block’s listing meant that any changes to the structure would require listed building consent.
‘The conservation-as-normal approach would have been to do very little in terms of improving performance and sustainability because the integrity of the historic fabric was more important than tackling the building’s shortcomings,’ says Oliver Smith, a director of architects 5th Studio.
New Court retrofit (1) Sustainable systems (2) Renewal of façade (3) Refurbished rooms
Sustainability fabric and systems: a. Photovoltaic panels b. Fresh air intake and outlets c. Extract air and heat exchange d. Fabric upgrades – air tightness, insulation e. Underfloor heating f. Ground-source heat boreholes
To its credit, Smith says, the college decided that ‘it had a responsibility to work out if it was possible to do something exceptional in reconciling heritage and sustainability through the refurbishment of New Court’. It appointed 5th Studio and engineer Max Fordham and together the team got to grips with one of the most radical refurbishments of a Grade I listed building ever attempted.
‘Fundamentally, the building had a heat loss problem,’ says Joel Gustafsson, senior engineer at Max Fordham. Adding insulation to the listed exterior was not an option, so the team set about exploring options for insulating the inside face of the exterior walls.
Aside from the issue of obtaining listed building consent for the intervention, there were serious concerns that improvement of the wall’s thermal performance would also create moisture and condensation problems.
‘When you insulate internally, you change the metabolism of the building; the risk can be anything from cold bridges leading to condensation and mould growth through to timber joists rotting,’ Gustafsson explains.
As luck would have it, Max Fordham had just completed a Knowledge Transfer Partnership (KTP) on moisture movement in the fabric of buildings. The New Court project offered the perfect opportunity to employ the theoretical expertise gained on a real project, using the industry standard hygrothermal modelling tool WUFI to characterise how heat and moisture moved through New Court’s historic walls.
Refurbished interior at New Court, Trinity College, Cambridge
Using this modelling tool, Max Fordham was able to undertake numerous WUFI appraisals with varying thicknesses of insulation and vapour barrier locations. On new build schemes, the vapour barrier is installed on the inner face of the wall to stop moisture inside the building from entering the wall, while using rainscreen cladding to stop moisture entering from outside. ‘The problem we had with New Court’s wall of brick, stone and render is that you get solar-driven inward vapour diffusion, which drives moisture into the building. This can lead to a build up of moisture on the cold side of the insulation,’ Gustafsson says.
New bathroom at New Court, Trinity College
The extensive modelling showed the most promising solution was to do away with the vapour barrier entirely to create what Gustafsson terms a ‘vapour open strategy’. This solution would reduce the heat losses, while still allowing vapour movement through the insulation.
This approach was not without its challenges, because it relies on keeping the moisture as a vapour by limiting the quantity of insulation added to the wall. ‘ You are slightly limited on what your U value can be because, if you restrict the heat too much, but not the moisture, you will have a condensation issue in the wall,’ says Gustafsson.
Fulfilling a seven-year watching brief
Monitoring installations were repeated during the main contract to provide a minimum seven-year ‘watching brief’, with E4 and E9 locations shown below as examples with slightly different time spans.
E4 (first floor) faces north into the courtyard and is well shaded by building geometry; with a rendered external finish this is perhaps a worst-case scenario.
E9 (second floor) faces north onto Garret Hostel Lane and, although in close proximity to other buildings, receives significant direct solar exposure; with the external finish as exposed brick this is perhaps a best-case scenario.
In both cases, the overall trend over time of relative humidity (RH) at all through-wall nodes is downward, although for the two outer nodes (3 and 4) in E4 this is much slower.
RH at the wall/insulation interface node 2 falls below the important 80% value at the end of February in both cases, indicating that drying of construction moisture is taking place.
The difference in solar exposure between E4 and E9 is clearly visible. The volatility of RH in E9 node 4 is a signature of ‘vapour openness’ – the rendered face of E4 clearly being more closed than the exposed brick face of E9, although solar exposure will also be playing a part. The effect of solar exposure on the drying of the outer brick wall in E9 is clear. At the end of May, there were a few grey days that were accompanied by significant rain on 31 May, causing increased RH at node 4 and, in turn, a little later at node 3 – Area A.
Solar activity thereafter affects rapid drying seen at both nodes 3 and 4 – a small residual effect may still be observed at node 2, the masonry/insulation interface sensor.
Comparing relative humidity at four nodes within external walls – two locations
Precisely how much insulation could be added was dependent on various parameters including the properties of the brick, stone and render used to construct the wall. ‘These are all natural materials, so their properties tend to vary considerably,’ Gustafsson explains. Nevertheless, because the properties of the materials from which the wall was formed were fundamental to the success of the scheme, samples of each material were sent to testing laboratories at Glasgow Caledonian University.
At the same time as the materials were being tested, the college employed the building performance research practice ArchiMetrics to measure the conditions internally and externally to the block and the moisture and temperature at various points within the wall, using hygrothermal probes. A weather station was also installed. ‘This was a key piece of work because it allowed us to calibrate our model,’ says Gustafsson.
When, however, the team did compare measured data for the wall with the results from the model, calculated using the measured material properties and the actual weather data for the sample period, there was, according to Gustafsson, ‘a small but repeatable difference between the model and the measured data’.
Stone facade at New Court Credit: Graham Copekoga
The surface convection coefficient was considered the most likely cause of the variation because it was affected by wind turbulence. The initial modelling had used the CIBSE semi-urban value for the coefficient. However, Gustafsson says the site is actually quite exposed and close to open landscape so that once the convection coefficient had been adjusted to between semi-urban and rural, the model aligned with what had been observed during the monitoring period. ‘This gave us confidence in our prediction,’ he says.
Using the calibrated model, 5th Studio worked with the consultant to put together a proposal to insulate the walls internally. ‘We worked with Max Fordham to work out what type of insulation was best and at what thickness,’ says Smith. ‘For parts of its life, this was as much a research project as it was a building project.’
The solution eventually decided upon was based on levelling the wall’s inner face with a 4mm-thick lime plaster skim, then attaching a 72mm-thick sheet of wood fibre insulation board. The inner surface of the wall is finished with a 15mm-thick layer of gypsum board attached directly to the insulation board.
Because the insulation is attached to the inner surface of the wall, Smith says, there was ‘a lot of angst about what to do when it gets to the cornice’. It was decided to leave the cornice exposed in its original position. ‘It makes it very explicit, what we have done,’ he says.
’Upgrading
SDC upgraded the windows
In addition to insulating the walls, the team also had to come up with a way to minimise heat losses through the building’s combination of sash and wooden casement windows.
Max Fordham appraised the effectiveness of 15 different options, which included fitting new triple-, double- and single-glazed units as secondary glazing or into the existing frames, only refurbishing the existing – and even doing nothing.
English Heritage would not permit replacement with new triple-glazed windows, even though the existing timber frames are already replacements for the original metal windows. Instead the solution decided upon was to take out the old frames, refurbish them, fit draught-proofing and new slim vacuum double glazed units. The team even found rippled heritage glass to add an aged look to the glass.
Along with cutting heat losses, the new windows also help reduce air leakage. ‘We aimed for an air permeability of 3 m3.m-2.h-1 @ 50 Pa and achieved 3.7,’ says Gustafsson.
To help limit moisture build-up in the student rooms, the proposed solution also includes an mechanical ventilation with heat recovery (MVHR) system. ‘The MHVR is jointly for energy saving and to help improve air quality and occupant comfort, but is also a key component of the vapour control strategy,’ explains Gustafsson.
The MHVR units are hidden in the tiny roof void of each apartment block. At the college’s request they are designed to be ultra quiet; in fact they meet Noise Rating 20, the same criteria used for many recording studios. From the roof space, supply air is ducted down the old chimney flues and discharged from the fireplace in each room. The air is either extracted via the ensuite WCs and communal kitchenettes or it passes under the doors before being extracted from the head of the staircase.
The college and 5th Studio applied for listed building consent. The BRE appraised the team’s work on behalf of the local council and concluded that the exercise was about as robust as it could have been, but that there was still a residual risk of moisture problems.
When listed building consent was granted, Trinity College decided to commit to the scheme. To manage the residual moisture risk, a condition of being granted listed building consent was that Trinity College has to undertake to monitor and report on moisture levels within the fabric for the next seven years. This obligation will be fulfilled by ArchiMetrics.
Position of sensors
As part of the refurbishment, the student rooms were given all new services, including new heating and lighting. The existing radiator-based heating system was installed in the early 1960s when Max Fordham was a student at Trinity and living in New Court. Now, 50-odd years on, the eponymous consultancy he founded has devised a heating scheme for the insulated rooms, replacing the radiators with a low temperature underfloor heating system hidden beneath the original, refurbished wooden floor.
The maximum flow temperature of the underfloor system is designed to be 45°C. Currently the heat is provided from an existing plantroom, but in the not-too-distant future, heat will come from a ground source heat pump. ‘We’ve run two pipes ready to pick up the ground array when it is installed,’ says Gustafsson.
To keep energy consumption to a minimum, particularly during college holiday periods, room heating is controlled by an occupancy sensor. Under normal use the rooms are to be maintained at 21°C. However, if the room is unoccupied for more than four hours, the heating temperature will set back. And if the room remains unoccupied after 24 hours the temperature will set back further still. Additionally, sensors in the window frames register when windows are opened for any length of time during the heating season, turning down the heating until these are closed.
The college required that the rooms also incorporate an electric heated towel rail, which Gustafsson has set to be off by default. ‘If you want it hot you push a button and it will come on for an hour,’ he says.
Aside from the towel rail, the remainder of the services are incorporated into a clever series of lining panels and attached to the inner surface of the wall. The lining panels hide new LED uplighting – they are a neat solution that eliminates the need to chase cabling into the walls. The panels do, of course, incorporate vents to ensure air circulates behind them to prevent any moisture build-up.
Electrical supply to the panels is from a new courtyard distribution system. This is routed up the building via risers tucked into the central stair well and then out through conduit concealed beneath the floorboards.
The mechanical services, heating flow and return, and domestic hot water flow and return follow a similar route. All the services are routed from the courtyard, under the main entrance doorway to a service pit hidden beneath removable stone flags in the entrance lobby. The space in the risers was so tight that the mechanical and electrical systems are distributed differently. The electrical systems rise up through the building from the pit, whereas the mechanical services go up to the roof void in larger risers from where they divide, and return back down the building through smaller risers. Each and every riser is different. ‘Design coordination meetings were many and long,’ recalls Gustafsson. However it was a worthwhile exercise because he says ‘everything pretty much went in as designed’.
Oliver Smith is happy too because the building is performing as expected. ‘At the beginning of this week we got the first monitoring report, and it backs up what we said we’d deliver,’ he says.
Project team
MEP and building physics modelling: Max Fordham
- Architect: 5th Studio
- Contractor: SDC
- Structures: CAR
- Cost consultancy: RUA
- CDM coordinator: Gleeds
- Building performance research: ArchiMetrics
- Specialist sub-contractor: Munro – M+E Service
- Building products: NBT
- Windows: Mat Bateman
- Lime render: AVV
- Joinery: Cousins