For those who were not present at the opening – and unable to witness our extravagant performance – I will describe for you some of the events that transpired last evening:
With a few small miracles in finishing some difficult sections of the vault and some very rapid, large-batch mortar rounds, we managed to complete the construction of the primary section of the vault in time for the exhibition opening. Our vault was completed with a small oculus at the pinnacle of the dome, both as a testament to its thin-ness and to our tight time-table, as well as a gorgeous teaser of interiority for a vault lit from within in the late evening. I myself had no time to shift identity from construction worker to architect, and so attended the opening clad head-to-toe in wet clay and plaster, passing somewhat precariously through crowds of tidy, black-clad young architects. I always love this moment, when roles may be reversed and the unspoken (perhaps latent) class divisions between the design and the building culture are subtly undermined – except normally I like to be the construction worker who morphs to architect/ intellectual in slightly more classy attire.
The risks that we have taken in the construction of this vault, in our eyes, have been absolutely proportional to the gains in our design and construction education. The task of translating into construction design (and the actual building itself) of Philippe Block’s complex curvature, as well as the task of building with a very low carbon footprint – all in an extremely short time period – are numerous and self-evident: 1. The choice of an unfired brick (never really meant to be used for such purposes), which resulted in both risk of water damage and the difficulty in adhesion of masonry to mortar, 2. the decision to build at our full scale of 21′ x 15′ x 10′ h, 3. the unresolved design issue of how to best approximate such complex curvature for the brick-layers and avoid local regions of negative curvature, 4. the very real significance of building such a complex structure with student architects (and not more experienced construction workers or masons trained in the eye and hand techniques of thin-shell vault construction), and 5. the timing of the project, which beset us with the task of resolving our material learning-curves even as we built the full-scale structure.
And so – there it was at the opening, finished. Would it stand or fail? Which variables of risk might govern our failure mode, if it were to occur? We decided that the wagers taken to realize this vault only begged for the greatest gamble: to de-center the structure with a full audience. We had removed one centering carefully before the exhibition, and – with the un-impeded view of the underside of one perimeter arch and the curvature of our vault springing – we gained confidence that our public de-centering would indeed demonstrate vaulted splendor. Our greatest error was that the time, and the pressure of public viewing, did not allow us to more carefully coordinate as a team in the de-centering process. Rather than de-center slowly and carefully the way the centering had been designed for removal, we opted for the more rapid and perhaps ‘ta-da’ method of the whole form removal – again, with much greater chance. During the removal of the formwork for the large end arch, the masonite surface (upon which the brick arch was bearing) was pinched by the OSB panels of the main formwork, pulling on part of the arch which had bonded to it. The result was a local failure of that eyebrow arch, followed by a crack along the opposite side of one groin, and then, perhaps 4 seconds later, by the collapse of that springing and the whole vault with it. Despite all of the risks involved – it was nevertheless a shocking occurrence for us.
My dear friend Mary Fillman, who graduated from MIT Department of Architecture in 1956, came to the opening and reminded me of the long history and celebration of such prototype risk at MIT. She said that Bucky Fuller would have loved this test (and the failure of the vault), and talked about Félix Candela’s frank conceit that he had learned the art of building such complex hyperbolic paraboloid shells through the collapse of several early prototypes. As you will remember, it is David Billington’s exhibition of Candela’s work that inspired the first proposal for our group’s MIT Museum grant submission, and John Ochsendorf, our advisor, who was a student of Billington. Such precedented cases of risk and collapse are irrevocably linked to innovation in the field of structural masonry. Indeed, the flavor of the exhibition itself seemed in some way to celebrate this history of innovation risk at MIT and our own part in the development of experimental thin-shell masonry structures. The time-lapse sequence of our vault’s construction, projected on the wall in the museum, animated the shattered shell in the courtyard with a sense that we pushed the limits of what we knew to be possible in both our design and construction capacity. And though the spectacular failure of our vault has been somewhat disappointing for us (and even perhaps, for a moment at least, just a hair demoralizing in such a public context), well, frankly… it’s pretty cool to watch a vault come down!
Risk and gain are bound together in innovation, however, only when a commitment to further innovation is present. This entire process has offered us a profound learning experience in which our design, material research, structural analysis and construction techniques have been tested. This vault has taught our team tremendously – even more so through its collapse. And I will speak for myself here, in saying – I look forward to implementing these discoveries in the next vault.
– Lara
We'll leave the light on for ya'
Despite some set backs in recent days, we are ON for tomorrow’s grand party in the courtyard. There are still bricks to be laid and other logistics to iron out, but with an extra energy and morale boost, we’re moving forward.
See it for yourselves tomorrow:
MIT Museum
265 Massachusetts Ave.
6pm open for students, 7pm for the community at large
With any experiment, there are obviously risks taken, with the expectation for a great return in the end. We were well aware from the get-go that building a thin brick masonry vault with freeform geometry would be pushing the boundaries of design, material and construction, and the last couple days have tested that theory and shown it to be a reality.
On Tuesday during one of John’s visits, we experienced a local collapse at one of the footings. One moment we were discussing the construction, and the next moment there was a crash as a small section caved in. Thankfully, no one was on the inside of the vault at the time, but it was still a shock for all of us to see hours of work crumble away. We started analyzing what happened and noted that the vault had failed at an area where the bricks had not been cleaned of their red dust coating. The dust was keeping the mortar from properly sticking to the brick surface, leading to easy separation between joint and masonry.
We were pretty bummed after this happened, but we cleaned up and started laying bricks in that area again, this time making sure to scour the masonry down to the clay base so that there would be clean adhesion between brick and mortar. We also thought there might be an issue with having too much thrust on one side of the spanning arch (between the main vault area and the cantilever), so to counter that we started building up part of the footing for one of the ramping arches.
We thought we were in the clear, but then yesterday around 10:30 pm, another collapse occurred at almost the exact place where the first one had. Much of the work we had done to patch up and build upon the previous day’s mishap fell through again, temporarily shaking the stability of the structure around. After some lengthy discussion with John and the rest of the team, we determined that the real culprit was local negative curvature (aka concavity instead of convexity) in that part of the vault. We knew the importance of double curvature for the structure’s stability, but this incident made it crystal clear that we would have to be quite careful and have sufficient guides to dictate the geometry.
The area of collapse - yet again
Showing the negative curvature in the vault - a big NO
Three of us stayed behind to build an additional layer over that footing, giving it the structural reinforcement necessary. It was a very long night, but we made some good headway and could move forward with construction.
Although we won’t be able to complete the entire structure for Friday’s reception, we’ll at least have a good work in progress to show our efforts. This has definitely been a testing ground for masonry innovation, and I think from here on out we’ve learned the lessons necessary to make the rest of the project successful.
- eLo
The most exciting news of the week: not “we’re completely done with the formwork!” but
“we finally have bricks and mortar!”
Yes, ladies and gentlemen, we have moved from the world of wood to the realm of masonry, living up to the name of our vault and the main purpose of our project. The bricks were actually first laid on Friday night (the 4th) – a moment we couldn’t wait to see happen (sorry for the delay) – and since then, we’ve been trying to plow full steam ahead. The rest of the guidework was also completed, as well as the formwork for the two ramping arches that will support the cantilevered leg on the building wall.
To prepare for the work, we first swept up the ground beneath to have a clean surface to work on. To prepare the bricks, we immersed them in a bucket of water to make sure they’re well hydrated and allow for the plaster to adhere well to the surface. Lara and Cynthia were the first team to get the footings going. At John’s suggestion, we had one person mixing the mortar and the second laying the bricks:
It takes teamwork
In the beginning progress was a bit slow, since everyone was pretty new to either brick laying in general or this type of technique (based on Guastavino’s catalan vaulting methods). But after a few rounds, we each got into a rhythm, leading to several pairs of people being able to work on the vault at the same time.
Some things we found helpful in our process:
- We had a range of mortar ratios to achieve the desired thickness, typically a 1:3 ratio (1 portion of water, 3 portions of plaster) for a thinner mix to a 1:2 ratio for thicker ones needed for bricks that began to cantilever outwards.
- Because the mortar sets faster when mixed excessively, we had to learn to be patient and let the mix set a bit on its own before use.
- These bricks are not only weak, but also extremely dry. Coupled with a layer of dust on the surface that kept the plaster from sticking properly to the masonry, we had to remedy this by scrubbing the bricks free of their red dust layer and then soaking each brick for at least 3 seconds in water prior to buttering them (aka applying plaster).
- Although we have some great tools (ex. rectangular trowels over pointed ones), sometimes it just works the best when you use your hands.
Progress!
- eLo
- – -
* Update: As with any experiment, complications occur…and ones that came with a bang. More soon…
Yesterday, Steve Rudolph of the Civil Engineering department tested our unfired clay bricks, to see what sorts of loads they would take. Because we’re working with unbaked masonry, we know it’s more brittle than the typical fired brick and so wanted to get some more definitive numbers on its performance before building an entire vault with them.
The first test oriented the brick vertically under the crushing machine:
I did take a video of the crushing action, although there wasn’t much to see. I had witnessed load tests before in my building technology class last year, so was expecting some spectacular bangs and possibly brick pieces flying across the room. My wishes were in vain – because the bricks are unfired and so soft, they actually have quite a bit of internal give and aren’t stiff enough on their own to snap or crack in any theatrical way. Rather, these were “stealth” failures: we couldn’t tell from looking at the bricks that they had reached their maximum load capacity, but instead the machine stopping indicated the material had failed. Only after looking at the back of the brick did we see what had happened:
Evidence of local crushing
In this direction, the brick took a load of around 1500 lbs before “failing,” or becoming structurally unstable. Just to do some further tests, we did another round with a second brick but in the long-ways direction. Since the bricks would be positioned in that orientation in the vault, it seemed to be more readily-relevant data. We found that, due to the increased cross-sectional area that the force would travel through, the brick could carry over 5000 lbs of force before failure. This time, the cracking was even more subtle than the other test. No pieces fell off until I picked up the brick from the machine and started prying at the cracks that had developed on both sides.
In the horizontal direction, the brick displayed even subtler hairline cracks.
Given the self-weight of the vault, which doesn’t exceed the maximum load of the bricks, there should be no structural issues with making this structure stand. [ sigh of relief ]
- eLo
As we take our first steps with our brick vault, it is the feet of the vault with which we are concerned. The ‘footing’ of the vault is the position from which our vault will spring, and each subsequent brick placed will take its cue from the geometry and orientation of the very first bricks placed. In liu of the tremendous complexity in the geometry of our vault, it is most critical that our first bricks are cut, positioned and set properly. To add to the importance of this task, these bricks must be cut from fired brick – not our unfired, adobe-type brick that we will use for the rest of the vault – in order to ensure that the slender springings of the vault are protected from water damage and not reconstituted into wet clay.
The photos below may appear similar, yet show the distinct differences in curvature and angular orientation at the end-condition of each footing. In each example, one may see the ‘eyebrows’ of catenary arches along the formwork, a second layer to be built only at the edge to stabilize the vault during construction, and the beginning of the vaulted surface itself in between. Footing 1 & 4 will have one more set of ‘eyebrows’ that will begin the coursing for the rampant arches to be built in the second phase of vault construction. These bricks have been cut and placed, but have not yet been set with plaster mortar – we will wait until our final diagonal groin guides are in position to begin.
Okay, less time for chatting as we move along…
– Lara
It seems that every day, we put in quite a bit of work but continually come across new issues that we either had never thought of before or that just came sliding onto our plates. I guess this goes to show that, in any project, there are unpredictable aspects that we just have to take in stride and not let them keep us back.
Just about all of the formwork is done, but the missing links lay in the diagonal guidework that defines the curvature of the vault and helps us visualize the shape while laying bricks. We looked at the profiles of these slices from the Rhino model…and that reminded us that yes, it is just a Rhino model – a sketch and not reality, given the highly-faceted nature of the computer model itself. Actually printing the templates, cutting, and assembling them lead us to see what we needed to modify in order to get smoother curves to work with.
Lara works to sand down some of the bumps in one of the guides...
...while Scott attaches the pieces for the other diagonal guide.
We had some hiccups in our calculations, but at last, we knew that these pieces would fit nicely in with the formwork we had already built. There’s still the issue of erecting these thin, spindly guides and securing them to the OSB panels, but we’ll be able to accomplish that tomorrow morning. We ended the day in a bit of a dismal mood, only because we’re getting antsy about not actually starting the laying of the bricks. Today’s visitors raved about the size of the project, but still – it’s all in wood and we want to turn it into masonry!
Tomorrow will be the day! Keep your fingers crossed for us…
- eLo
Yeah! The main formwork for laying the bricks was finally completed last night. We also laid out the bricks at the four corners. Our schedule today will be to start laying the bricks along the formwork and to work out on the diagonal guidework, which will help us follow the correct curvature of the vault.
We spent four days setting up the main formwork. To be honest, four days is a bit more than we have expected to spend on this task, but it took additional time because the translation from a 3D Rhino model to the real construction is not easy. Here is our preliminary take on the formwork, initially developed by Scott and Flo:
We wanted to be on the safer side to ensure that the formwork was secure while building the end arches (one of the first steps in brick laying). Although the entire structure is structural and will be stable once all the bricks are in, the worry was keeping the masonry well-supported during construction and accurately defining the geometry from the get-go. There was particular concern about the side arches, which are not completely vertical but actually tilt inward. The initial design called for quite a bit of wood, but in the realities of building, we’ve managed to minimize some of it by increasing the spacing between the studs and devising a different way to support the tilting arches.
There were so many issues that we figured out on the site while constructing the formwork. One of the important issues, realized later on, was the exit strategy for the formwork. We spent a lot of time finding a solution that would make it easy to take out the OSB* formwork and the masonite* lying along the edge of OSB. It was important because, if not taken into careful consideration, we could damage the brickwork while removing the wood. The exit strategy that we have finally agreed on is to
- Unscrew the studs from the OSB.
- Unscrew the blocking that support the masonite.
- Slide out the masonite and leave the vault standing on its own.
Let’s have a preview of the stop-motion movie, recording the past four days of our work:
- Cynthia and Emily
- – -
* OSB = oriented strand board, an alternative to plywood
* Masonite = a type of thin, flexible hardboard that we used to line the curvature of the formwork and support the bricks.
We’ve gotten a few visitors in the last few days who came by to check out what we’ve been up to (Huang), give commentary (John O.), jog their memories about funicular shapes (Debora), and bring tasty sweet treats when our blood sugar was running low (Juliet). Although we haven’t yet put mortar to bricks, all the formwork is up and it’s now very apparent just the scale of what we’ve been working on.
“Wow, it looks like a lot of work!”
Well…yes.
We’re trying to keep up the steam as we barrel into this last week or so of work, and as you can see by this blog, we’ve also been making sure to keep all our dear readers (actual and theoretical) entertained and up to date on our goings on. In fact, this little guy gets the most preferential treatment on site:
the king of the site
Thanks to Cynthia’s camera, we have a great record of our progress so far. After every milestone,which can be up to every half hour or less on particularly productive days, someone tries to remember to run over and capture the moment. Not only does this help to give others an idea of what we’ve been working on, but it also shows us how far we’ve come. In the end, we compile a stop-motion animation like previously seen. To give a preview of where we were close to the beginning of the week:
With two simultaneous clicks, the formwork went up.
Stay tuned to see where we are now…
- eLo
As our vault formwork is nearing the full assembly stage, we have been multi-tasking our material tests in preparation for the actual vaulting work: plaster consistency, time/ sequence and adhesion tests, as well as the first fitful attempts at vault construction. As a necessary early experiment (in both the material and the sociological domain), I put together some simple formwork for a small catenary arch*.
The idea was simply to begin training our facilities in handling the brick material (with its very unique set of properties per unfired brick, not least of which presently is water solubility), and the plaster mortar (which, depending on the proportions and volume of the mixture, may set very rapidly). These factors require very good time/ sequence coordination between team members mixing mortar, preparing surfaces, and applying mortar to lay the coursing, as well as coordination of the types of tools used in mixing, handling and applying mortar to bricks.
The actual construction sequence goes something as follows:
1. Generating the structural form: Hanging chain model
2. Constructing the formwork: Quick-release designed in for formwork stability, which also allows the form to be ‘dropped’ and pulled out without disturbing the finished arch.
3. Setting the Mise-en-place: Formwork, stack of bricks, mortar, water, mixing bucket, palette tools, sponge, etc.
4. Vault construction:
Ps. Don’t do this – Building a vault in tension, rather than building up from either side in compression.
5. Removing the formwork
6. A loading test, anyone?
Though I would not say that our first test was particularly facile, it was a good opportunity for us to establish the primary tool, timing and material parameters in the brick-laying process. Now with our first test completed, our hopes are to work more confidently with our timing, and to allow our team to conscientiously slow the pace to be accurate, clean and thorough in our bricklaying.
– Lara
- – -
* catenary arch = the ideal curvature of an arch under the conditions of its own weight
