For a complete evaluation a cumulative smoke measurement may be needed in addition. Although the cone calorimeter is a valuable research tool, and may be used frequently in industry, it needs to be made more robust and user friendly to avoid excessive downtime and to reduce calibration time. If the instrument is to be used for routine development and quality assurance, only one or two heat flux levels should be selected. These could represent a developing and a developed fire and should depend on the product application.

We should like to reply to the letter from Dr Brabauskas (Fire Muter. 15, 101 (1991)) concerning the APME paper' on the cone calorimeter, which appeared in this journal.Our remarks on the robustness of the cone calorimeter were not aimed solely at the instrument at Fire and Materials Ltd where our test programme was carried out. Indeed, the comments were largely the result of experiences of APME members who bought commercial instruments and experienced difficulties during the commissioning period and beyond with various components which needed to be replaced or adjusted by the manufacturers. The laser system and its necessity for fairly frequent realignment is still, in our opinion, a weak point. Frequent calibration checks are obviously needed at the present state of development of the commercial instrument but they are time consuming, using a significant proportion of the operator's working day. We look forward to the development of a more robust instrument which could eventually be used by skilled industrial staff without the present painstaking 'setting up' procedures. With regard to the operating software and the Fire Data Management System we agree that these represent a major step forward but we understand they are unfortunately not widely available at present. The use of such systems does not obviate the necessity to rationalize large quantities of data.The effect of thickness is a very important consideration for plastic materials. In fact for many applications in building, furnishings, transport and electrical products plastics are used at thicknesses below 6 mm (typically 1-3 mm). We agree that this means that the substrate upon which the material is used can have a large effect but we feel it is, on balance, more important to test the material at its end-use thickness on a representative substrate than to test at a thickness which is seldom used in practice. The composite sample (i.e. thin plastic and representative substrate) should then be wrapped in foil and placed on the fibre blanket. However, there are certain plastic applications such as films and surface coatings for which the thickness (50-250 mm) is too small for the cone calorimeter, in its present state of development, to produce reliable and accurate data collection.We obviously agree that the testing of expanded polystyrene and other thermoplastic foams creates a problem in this instrument. The mechanism of fire retardancy of such very low density ( < 20 kg/m3) foams includes a very rapid retraction of the foam away from the fire source and any true evaluation should attempt to model this. The small falloff in external flux of 8% claimed by Brabauskas probably does not represent the behaviour in a real application such as mechanically fixed or glued tiles.Our suggestion on guidance on heat flux levels applies to situations in which the end-use application of a product is known. For exploratory testing of materials we agree that the EUREFIC recommendation referred to by Brabauskas is a valuable starting point.Our concern about the effect...

A two-storey test rig was used to investigate how an insulated facade would react to the impingement of a simulated room fire. In particular, whether or not a facade insulated with polyurethane foam would promote vertical flame spread. Wooden cribs weighing 40 kg were used as fire sources. One set of tests was run without any steel cladding to study the behaviour of the polyurethane foam alone. In a wall-configuration test as well as under a corner-configuration one limited vertical and horizontal flame propagation were found. The degree of damage was, to some extent, greater under the corner-test conditions than with the wall tests. A ventilated facade construction with the profiled steel cladding fixed to vertilcal support-work was used for the second set of tests. In order to simulate repair conditions, the outer metal cladding was completely removed from the lower 1.5 m of the facade. In both tests a strong chimney-effect behind the cladding was observed. This intensification of the flame impingement led to a flame spread up the top of the facade. The tests indicated that the vertical flame spread would continue unless the method of construction incorporated vertical fire stops.