What the fins are for

The fins on the south elevation of the building in this issue's cover study are 520mm deep and spaced at 820mm centres. Nobody chose those numbers for how they look. They are the output of a cooling load calculation run in the second month of the project, before the practice had drawn an elevation at all, and they are the reason the plant room is one storey rather than two.

This is the part of the brise-soleil that does not survive the magazine photograph. A shading device is an instrument for moving money from the facade budget to the mechanical budget, or the other way round, and the exchange rate is set by the sun angle at the hour the building is fullest. Everything else — the rhythm, the material, the way the shadows rake across the wall at four in the afternoon — is what you get to design once the depth is fixed.

Work it the other way and the absurdity is obvious. Draw a beautiful screen, then ask what it does at the summer solstice, and you will find it either shades the glass at noon and blinds the room at five, or it does neither and costs the same. The section is where the answer lives, because the section is the only drawing that shows the sun arriving at an angle.

There is a second calculation that almost nobody runs, which is what the fins cost to clean. The practice that designed the cover study specified an access rail at the head of each of the 14 bays, and it was value-engineered out in the final package. Four years in, the south elevation is grey at the top and pale at the bottom, in a gradient you can read from the street, and it is the most honest thing about the building.

Ask who designed the screen and you will be told the name of the practice. Ask who set the depth and the spacing and the answer, on this building and on most of them, is a services engineer in month two, working from a cooling load and a shading target, in a spreadsheet nobody kept.

This is not a complaint about engineers. The calculation is the right way to arrive at the number, and the number is good: the screen cuts peak solar gain on the south face by a bit over half, which is what took the plant room from two storeys to one. The complaint is that the output of that calculation arrived in the architectural model as a geometry with no note attached, and from that moment it was treated as a given rather than as a result.

You can tell, because of what happened next. The depth survived three facade revisions unchanged. The spacing survived a change of structural grid. Both are quantities that should move when the thing around them moves, and neither did, because nobody in the room knew what they were a consequence of.

Ask three practices for a shading strategy and you will get three elevations and no numbers. Ask for the cooling load with and without, and two of the three will send it the next week, which tells you they had it all along and did not think you wanted it. The third will not send it at all.

The elevations are not useless. They are just answering a different question. An elevation is a statement about composition — rhythm, proportion, where the eye rests — and on a screen like this one it is the drawing that cannot show the thing the screen is for. A fin seen square on is a line. Its depth, which is the entire mechanism, projects toward you and collapses to nothing.

That is the gap this journal keeps returning to. The quantities exist, they are produced early, and they decide the thing the photographs are of. They are simply not what anyone publishes.

Which is the limit of a page. An elevation flattens the 520mm to a line, a section cuts it, and neither performs the move the argument needs, which is walking round the thing. The last plate is therefore not a drawing.

Plate 03model · 1:50

Scale

1:50

Printed size

341 × 190 × 224 mm

Parts

19 parts