My next little adventure has popped up out of the ether. For undisclosed reasons I started a rapid “BIM sketch” of Fallingwater, the famous ate project by Frank Lloyd Wright. This is a “quick and dirty” model purely based on publicly available image from the Web, and not worrying too much about dimensional accuracy. Rather I’m trying to capture the spirit of the design and get the layout of the house into my head, identify points of interest and difficulty, start to think about the significance of this project in the long story of humans building homes.
Like Le Corbusier’s Villa Savoye, this is a vacation home, eminently impractical in some ways, but stunningly creative and inspiring to generations of architects and lay enthusiasts. It straddles a fast-flowing stream with a small waterfall. Overhangs might be a better term, but the overhand is so extreme that straddles seems apt.
Four main levels, with just a viewing platform and undercroft at stream level, a main living area above that and bedrooms on the top two floors. So in a way it’s quite a vertical building, but being Wright the emphasis is on the horizontal, with broad cantilevered terraces, and bands of glazing rather than conventional windows. This last attribute is shared with Corb, but otherwise it’s a very different aesthetic for a very different site.
About a day and a half’s work in the images shared here, and these are 75 yr-old days of maybe six working hours, not the ten or twelve I regularly but in a few years ago. No complaints, I still have a wonderful life and the slower pace has its benefits.
For these early sketches I usually set up sheets with several working views on them, to make it easier to see the “big picture” and to jump between open views in a simple and intuitive way. So far I just have the plans set up, with jpegs I found online as an underlay. A couple of these had scale bars to get a rough idea of the size of things. I also looked at some 3d modes on sketchfab. No guarantee at all how inaccurate these are, but it helped to have something I could orbit.
The real proof of the pudding is in the many photographs available online. Once you have the model roughed-out it’s possible to compare these to what you have and go round troubleshooting. Bit-by-bit the uncertainties pop up and get resolved. It’s a process I’ve been through multiple times not and something I find really invigorating.
Elevation view of Fallingwater by Frank Lloyd Wright. For the windows I started with a thin glass wall, that’s the simplest way to model large areas of glazing in the earliest days of modelling, for me at least. Then you can go one of two ways. Use that wall to host window families in such a way that all that is left is a sealant gap between the units, or maybe change the wall type to a slightly thicker metal “subframe” so that the gaps between windows appear as metal posts. I try to follow what I believe to be the construction method used or to be used.
Method two is to convert the basic glass walls into curtain walls and set up glazing patterns. This has its drawbacks. You can’t schedule with the “normal” windows for example. But for now, this is what I have done with Frank’s little gem. I have 3 or 4 different types of curtain wall set up to suit the various conditions. No mullions needed at first but you can get quite a long way with the standard rectangular mullions. I guessed the sizes. Possibly I will convert back to a glass wall once I have time to model window families with the right kinds of openings. This is always more stable than a curtain wall where the sizes of panels are always instance-based and subject to unintentional change while editing the rest of the wall.
I need more information to take the next step. But that is probably coming down the pipe.
