Individual Study of Roof Tiles for Our Grid shell

The Origin of My Grid shell Roof Tiles Study

I have been fascinated by triangles since a child. Therefore, I thought I would research more about it.   Triangles are difficult to use for buildings since they only have three angles.  They basically provide a living area that gradually reduces itself area towards the other angles.   It may be disguised fully if the triangle is a large block such as the Flatiron Building in New York.


But, despite building to be perfect for the site, It is my understanding the shape of the Flatiron buildings caused it to have some offices that have a odd geometric space.

It may work best if the triangle is part of the roof or of the elevation, rather than the plan – this will make sure the plan is roomier.   This would be the case with the Swamp Huts by Moskow Linn Architects in Massachusetts.


This last week, we decided to study the roof tile of the Grid shell more closely.   I got interested in the triangle, and so did several other students groups.

I did my initial model with some laser cut basswood strips; stretch magic jewelry bead wire, and some old museum board or card stock from previous semesters.


I initially started small squares attached to each joint.  This would have made possible for us to pre-install the tiles while the grid shell is retracted.    I left the bead wire extra long to facilitate tying, untying, and further adjusting.   I may consider cutting the extra slack – but since I am still experimenting – I will leave the strings long for now.

But, the little square tiles, began to move and rotate all over the place producing a random appearance that looked less than architectural.


They looked interesting in the model – but a few days retracting the structure and reopen it – even though it worked well – it did not look appealing and they seem without a concept.

Afterwards, I did try with two right triangles – and the result was much better.   It did retract with no problem, and it did open smoothly.   It seemed as if I was up to something.


Then, I decided to attempt, a rectangular shape without the edges, trying to make it look like a hexagon or an octagon– I placed each rectangle on each joint.  But it was also behaving unpredictably – and I decided to increase its size.


I made a large octagon and attached it to two joints.  The large octagon worked well – but it hid the structure.  Most of the students (including myself) think the grid shell is a nice looking structure and it is important to display it.


Last but not least – was another experiment with some medium size rectangular octagon?  They worked well – but they hid the structure a bit.  Anyway, it did not produce any excitement when I presented it.   So, I tried a scored version of an isosceles triangle – but that did not work as well either.    The triangle was too large, and the top angle hit the other triangle every time I tried to retract and reopen the structure.   So, the only tile that worked was the second one from left to right – with the right triangle.

Each right triangle attached to two joints on each extreme worked well when retracting the structure and extending it.

These were my experimental models from Monday through Wednesday.   By Friday, it was requested I tried my designs with a grid of at least 4 to 6 arrays to see how the different tiles would behave.

I decided to experiment with all sort of triangles:



I began using the Equilateral triangle and it worked well – as long as you tiled the roof every other row.

Then, I began to work with the Right Triangle.   This time I placed the tiles on every row.   But, it overlapped too much when retracted, became rather bulky when extended, and very annoying since it got stuck on the ends, and it also hid the structure.

The Right Triangle may work best if you skip every other row, just like the Equilateral Triangle.



After the Right Triangle, I proceeded to experiment with the Obtuse Triangle.   An obtuse triangle has one angle that is over 90 degrees.   I skipped every other row, and the structure retracted and extended well.    At this time, this is my preferred shape, the obtuse triangle.



I revisited the Isosceles triangle again, but you need to skip every other row.  But, I found the top angle too high and it did conflict with the other pieces. So, no – I don’t suggest we use the isosceles triangle.



Last and not least is the Scalene Triangle.  In this triangle, all of the angles have less than 90 degrees, i.e. they are all acute angles.    We call it in Spanish, “el triángulo escaleno o agudo”.



The Scalene triangle works well if it is only in one direction, and maybe if I made a bit smaller it will work.

Characteristics of the Roof Tile and Suggested Materials

We expect our roof tile to have a rigid yet have a moisture (mold proof) flexible characteristic.    We are thinking about corrugated plastic, a material marketed under the name Coroplast ©.

The Evolution of Our Minimum Surface construction throughout the semester

By providing a roof to our structure that is thin, lightweight, we continue with our quest to build with thinness in mind and with minimum surfaces construction techniques.


Coroplast © is the trade mark for corrugated plastic.  For more information please visit the vendor through their web site:

The Triangles:

 Scalene – All of the sides have different lengths, and the angles are different.

  1. I made the Scalene triangle also an Acute triangle
    1. Because I made all the angles less than 90 degrees.

Obtuse – One of the angles has more than 90 degrees.

Equilateral – All sides have equal length.  All angles are equal as well.

Isosceles  – Two sides are of equal length.

1.  You can make a right triangle that is also an isosceles.

Right Triangle – At least one of the angles has 90 degrees.

No other types were used because I thought these were the best shapes.

Description of Scale

I calculated each cell to represent 1 ft. by 1 ft. at a scale of 1 inch equals 1 foot.


 A Minimum structure format can be adapted to the roof cladding as well by using lightweight, and thin materials.  The week before Spring Break, I was in a group that experimented with fabric.   This week I decided to represent thin plastic at a scale of   1 inch equals 1 foot


The Scalene and the obtuse triangles worked best.    Although, my favorite was the obtuse triangle – it tiled well and it showed plenty of the structures.     Left to do would be to double layer the structure, just like we did in the other models, and see how it would work or look.  My project is beginning to look too much like my other classmates, and it is very likely I will jump on the bandwagon and do a coalition with the other triangular folks in class – they have created a model with a scored center square that is attached between the wood members on two sides.

They created this square by scoring it diagonally in the center, while still attached to two sides of the square.   Since they are scored in the center – it forms two triangles on each side – giving it a pyramid shape – the roof tile also has punched holes.  It retracts well, and it extends well.

Furthermore, we may not need to worry much about retracting or extending the structure too much.   From our conference call with the structural engineer on Friday – it seems the deformation of the structure will be permanent.   So, we may not need to worry much about the tile be compliant with the retraction or extension of the wood members.   More than likely, our structure will remain in one position.

Thank you.   Andrés Mulet


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: