OUR LADY UP THE HILL

This is the second of (probably) several posts on Le Corbusier's famoul pilgrimage chapel on the hill overlooking the village of Ronchamp. As source material I had quite a lot of photographs and several drawings, in jpeg format. These can be dragged into Revit and scaled up to life size. I estimated the size based on Google Earth & started tracing over a plan.

The curved walls are generally irregular and asymmetrical, For simplicity of construction I have compromised a little, but to maintain the feeling broken them down into short lengths with different radii. I'm using real walls, not extrusions, because eventually I want to cut sections and see the thickness of the roughcast plaster. The South Wall is an exception. This tapers and curves, rising to a sharp point at the South East corner so the wall system tool is not an option. I used in-place massing, starting with a blend, then adjusting with push-pull.

The west wall is curved in plan with an irreglar curve along its top edge. Edit profile won't work on curved walls, so I made a wall hosted family with a void cut, using a spline for the lower edge, and adjusting it by trial & error. The same method was used on the east wall.

I made the chapel walls symmetrical about an axis so that the half-domes could be in-place revolves, picking the top edge of the wall to make the profile. This is not quite accurate. The tops of the domes seem to be slightly flattened. But I think its an acceptable trade-off.

Not everyone realises that this building is based on a regular grid. Actually Corb nearly always used a rational grid. I set the grids up and used them as reference planes for a series of splines within an in-place mass. I read a post last week where someone wanted to abolish in-place families. I didn't find much use for them when I first started with Revit, but on a project like this they are indispensible. Select the splines and hit "create" to make a surface.

One of the splines is placed on a reference plane (not a grid). This can be rotated to fine-tune the angle. The whole surface can be adjusted while in edit mode by tab-selecting the splines one by one and adjusting the control points. I set up several perspective views to match photographs & just kept tweaking until it looked OK.

There is a void cut to form to remove the excess around the North-West corner. I tried to use roof-by-face to create an element with thickness, but it refused. This is unfortunate because I was hoping to attach some of the walls to the roof.  Strangely enough, wall-by-face worked, so the roof is actually a wall as far as Revit is concerned.

At this point I had a light-bulb moment. I don't know why it's taken so long to think of this. For over 6 years I have been hiding and unhiding section boxes in the view, simply because I find the solid lines annoying and confusing. My light bulb said, go to object styles and make them grey dash. No longer obtrusive and annoying. This is going into my standard template. The other penny that dropped was to do with the roof construction. The more I looked at my reference material and thought about the structural properties of this roof, the more I realised it must be a hollow double skin construction with vertical ribs along the lines of my splines.

The ribs must line up with the concrete columns. There are photos taken during construction which show the column and beam arrangement for the south wall. The infill is semi-coursed stonework salvaged from a previous building. I made the columns as a loadable family with parameters to adjust the base width either side of the point of support.

I realised now that my first attempt at the south wall was inaccurate, so I rebuilt it, this time as a series of vertical profiles hosted on the faces of the columns. I gave them a 50mm offset to allow for the thick roughcast plaster finish and that worked fine. By setting the wall to ghost surfaces in a view you can see the relationship between concrete framing and masonry infill.

Creating the tapering voids for the windows is not so easy. They are blends of course, but because the wall itself is tapering they tend to get distorted out of square. I tried to solve this by drawing on the plane of the wall, but the wall is not a flat plane, so you have to create reference planes that approximates the interior & exterior surfaces of the wall. This worked fairly well, but I need to go back and set it up more carefully when I have more time.

The North Wall was much easier. I modified one of my simple window families so that I could control width & height by instance. (every window is different, so no point in creating types) Then I set up an image and a revit view side by side on my screen and eyeballed the whole thing. Good enough for first pass.

By this stage I had become interested in the site context and created a toposurface around the building.  This was all based on guesswork, but later, (after I discovered the google earth trick) I created topography for the hill and surroundings.  I was also adding more and more detail.  The floor inside actually follows the slope of the site, then steps up to create a low platform for the altar. 

As an illustration of the way that making a BIM model leads to a deeper understanding of a building, check out the stair on the North Wall that leads up past a green door to a red door. 2 storeys up.  There is a cutaway perspective that appears in many text books, copied repeatedly by different artists including the redoubtable Francis Ching (No offence, I love his work).  Because of the cut plane, only the lower half of the staircase is showing.  Nobody seems to have noticed over all the years that this drawing has been copied that the stair is actually going in the wrong direction !

Lots of fascinating details in this building.  The rainwater spout that looks like a bull's horns in section and discharges into a sculptural feature with a storage tank below.  The play between paired forms either side of the East Wall.  Two altars, two crucifixes, two choir balconies, two pulpits.  This creates two churches - one indoor & one outdoor.  It also gives Corb the perfect chance to play his little games with the relationships between the pairs.
I'm not 100% sure of this but I think there is a bridge at first floor level in the N.E. corner so you can walk throught the green door, assemble in the upstairs room, then across the bridge to the indoor choir & through another door to the external choir. 

These internal subtleties were modelled during my second weekend on Ronchamp, and in-between I had been exploring Gaudi and discovered the topography trick.  So once again I grabbed the terrain using the plug-in for AutoCAD, linked this mesh into Revit and created a toposurface.  Then I save an image of the same area, also from Google Earth.

The trick now is to crop the image very carefully so that the corners match the rectangle of the toposurface.  You also need to know the size of the rectangle.  From there it's fairly easy to make a material which scales up the image and maps it onto the terrain.  Probably you will have to mirror the image (no idea why).  You may also have to adjust the offset values for the image (materials /appearance /click on image / transforms) 

Someone has asked me to do a video of this process, and I would love to, but I haven't yet ventured into that territory.  Something for the future. 

Didn't take me long to adjust the topography I had previously made by guesswork so it fits on top of my new hill like a little hat.  New insights almost immediately.  Pathways through the trees that surely must be pilgrimage routes.  A visual connection between the church in the village below and the chapel up on the hill. 

If Yves Gravelin is reading this he might care to comment on some of my guesses.  He joined the blog not long ago, and happens to live within driving distance of Ronchamp.  He has a cool website featuring a Revit model of a theatre designed by Ledoux.  Nice work Yves !  http://theatreledouxbesancon.blogspot.com/