Our involvement in Castle Drogo began in 1994. We had recommended that any solutions should be based upon a long-term investigation programme into the causes of the problems affecting the castle. This approach was considered to be more conducive to reaching strategic solutions than a simple report summarising the castle's condition. In 1996, together with a team of consultants, we began on-site testing.
We conducted a series of tests on an area of the south wing's west elevation, with a two-fold purpose: to ascertain the extent of the water problems and to test the efficacy of existing and new types of mortar. We used non-destructive testing techniques to study the arrangement and condition of materials within the masonry structures. The process consisted of a series of monitoring periods of one and two years, punctuated by intervals of analysis leading to the publication of a report recommending next steps.
This first set of investigations focused on the defective waterproofing envelope. We used thermography and pulsed radar - which shows differences in moisture content - to measure how much water penetration was occurring through the roof, elevations and windows, as well as finding out how much moisture was already in the walls. Moisture levels within the west wall were measured with wooden and metal sensors placed at varying depths. The resulting data was uploaded to a computer every three months and the results analysed after a period of two years.
For the stonework, Lutyens used an early cement mortar with properties akin to hydraulic lime. This was replaced at a later date with a Portland cement mortar with a waterproofing additive; it failed due to shrinkage relatively quickly, which led to water build-up behind and hindered the evaporation of moisture. 'Free lime' from the mortars was precipitated on the face of the stonework and poisonous lead salts developed from the lead damp-proof course.
Through an iterative process of on-site trialing and testing over six years, we identified an Italian lime mortar which was suitable and which did not leave an unsightly calcium carbonate precipitate.
Lutyens' construction build-up was technically awed and the subsequent repair solutions over the years were largely ineffective.
Inspections by GB Geotechnics uncovered cracking due to differential movement at the base of the parapet and the roof.
Criteria for restoration were laid down in a series of documents, and the following policies were put forward for repairing the roofs:
Policy 8: the principle of authenticity in matching of Lutyens' construction technology should not be adopted. This is justifiable on the grounds that the fabric remains at risk for the future and the burden of maintenance is not sustainable. It would not be possible to hand down the building in good or better condition;
Policy 9: to respect Lutyens' original design intentions for the granite paved roof terraces and parapet wall; and Policy 10: to seek an improved sustainable solution to the thermal and moisture problems.
Lutyens' roofscape could be considered as very 'modern' for this period. However, Lutyens designed a 'cold' roof, without insulation or ventilation and without accommodating thermal movement at abutments of concrete, steel and granite. Inspections of the mastic asphalt concluded that deterioration was due to interstitial condensation and lack of movement joints at abutments. Significant interstitial condensation was found to be present and surface condensation had promoted mould growth.
Lutyens covered the roofs with granite paving slabs to achieve a monumental unity with the granite walls, but this had an adverse effect of enhancing cracking of the asphalt and water penetration through the construction. The Drewes disposed of the slabs but the roofs continued to leak for obvious reasons. This was compounded further when the parapets were dismantled and a lead damp-proof course inserted.
Current works have employed modern roofing technology from Bauder for the selection and build-up of the waterproofing membrane, insulation and ballast to ensure the performance and longevity of the construction and to minimise maintenance requirements - to resolve the inherent design aws, we adopted an inverted warm deck roof to eliminate interstitial and surface condensation. The choice of a membrane is determined by its resistance to differential movement. The choice of insulation (closed cell) above the membrane is determined by its ability to resist interstitial condensation damage and cold bridging problems.
Reinstatement of granite slabs would then prevent wind uplift of the insulation as well as restoring Lutyens' design intentions.