This video was produced was part of The Carpenters’ Company 2011 Master Builder Dialogues, “Preserving America.” It was developed in collaboration with Independence National Park and Drexel University.

Robert: Thank you so much, Charlie. It’s so nice to be here and so wonderful to talk to all of you who have this great interest in preservation and to talk about this great iconic structure.

As the lights dim, let me just ask, how many of you have been to Fallingwater? That’s a good chunk, even though it’s out west in Pennsylvania.

I hope what I’ll be able to tell you today may be some things that would amplify what you already know and enlighten your already passionate interest in the building.

Just historically, Frank Lloyd Wright got a good reputation for himself building prairie houses at the turn of the century. Then he had this personal quirk of running off with a client’s wife, with Mamah Cheney, and going off to Europe and being hounded by the press and being vilified and all of his clients deserting him, coming back to the United States, building Taliesin for Mamah. She was in Taliesin with her 2 children when the terrible disaster struck, as you probably all know, and a crazed servant burned the house down and murdered Mamah and her children. That was the end of Frank Lloyd Wright’s career for 20 years.

From that 1911 disaster, he had no work for 20 years except for the Imperial Hotel. He was down and out. He needed money.

In 1932, he got this brilliant idea of starting, what he termed, a school where people would pay tuition to come and study with him, but it was really to work for him in the apprenticeship or the fellowship. I was privileged to know Edgar Tafel, in his last 20-something years of his life, very well. Edgar was one of the original apprentices in 1932 who was the chief apprentice who did the drawings of Fallingwater. A lot of what I know about this building comes from my friendship with Edgar, who just died this year at almost age 99 in January.

Let’s just go through a little bit. Again, as everybody has said, when you work on something like this, it’s not the work of one person. There were other consultants that we had, although we were the prime consultants – somewhat unusual for a structural engineer to be the prime consultant. There was no architect insulating us from the client. We worked directly with them. The house is now owned by the Western Pennsylvania Conservancy, the largest land conservancy in the United States, and its executive director, Lynda Waggoner.

We had consultants for post-tensioning – Schupack and Suarez. Mario Suarez helped us on that. We did non-destructive testing with GBG, an English company that now is located in the United States as well. The post-tensioning was done by VSL. Other structural work was done by Structural Preservation Systems, and that’s a sister company of them.

There were architectural repairs that were necessary that was done on a separate contract, parallel work done by WASA in New York.

Let me start off by giving a little background and a description of the problem. This is, of course, the iconic picture of the house taken from down below, where the waterfall is. These are drawings that are taken from the book that was written at the 50th anniversary of the house, 25 years ago. Today, it’s the 75th anniversary of the house. It was built in 1936. A new book is out now, which has more spectacular photos.

SilmanThese drawings done by Lou Astorino, a Pittsburgh architect, have an idea of how the house is actually constructed. You can see it’s really built on top of the waterfall. There’s a big massive boulder here, which Mr. Wright instructed would be the first floor of the house. That’s how they determined where it would be. It is the hearth of the fireplace actually.

There are 4 main structural supports coming out of the water, 3 of them are concrete. 1, 2, 3. The fourth one is masonry. These are called bolsters or piers. They’re just sitting on the bedrock in the stream bed up on top. You can see some of the projecting boulders here. These 4 elements are one of the main parts of one storey and what sits on top of those.

On top of the four bolsters is the first floor. This is an upside down structure. The concrete slab is actually on the bottom. When you stand underneath the house, on the falls, and look up, you see the concrete slab.

From the top side, you see an open egg crate kind of a thing. What you’re seeing here is, on top of these 4 bolsters, you see one beam here with a funny cutout, 2, 3, and then the fourth one here, which is on top of the masonry wall. Here’s that boulder for the fireplace that projects through.

These are 4 main concrete beams, 3 feet wide by 2 feet deep. Then going across on the other direction are these concrete joists, as it were, 4 inches wide and 2 feet deep, spaced about 4 feet apart. The finished floor is on top of that, which we’ll see in a minute. You can pick up the finished floor and get at this concrete work.

In the end wall, the south parapet – a very important part of our story – are 2 window frames. They’re rectangular, steel window frames made out of little 2 1/2-inch Ts. Remember that. A big part of our story.

Then you can see all of the masonry that’s built up there, the piers here, here, here, here – all made out of local stone, stuff that came out of the site. It’s a local sandstone.

It was built with wood scaffolding, coming up out of the stream. Classic kind of stuff. You just hammer it together and nail it together until you get it strong enough.

The second floor, which is very much like the one below it, has this master terrace sticking out here. Then the master bedroom’s in here. This is an open trellis work here. Now you can look down at the floor below and see the stone covering that covers up that egg crate kind of construction.

That’s essentially how the house is put together on these various trays. It is built on top of the waterfall, as you see.

We got a call one day in 1995 from the executive director, Lynda Waggoner. She said, “How would you like to work on a problem we have at Fallingwater?” If you’re in the preservation business and you get a call like that, you kind of gasp because I don’t know if you know this, but Fallingwater has been voted by the National AIA as the Best All-Time Work of American architecture.

If somebody asks you to work on the number one building in America, it’s a staggering responsibility. It’s very scary in some ways and very flattering in others. Of course, we jumped at the chance.

“What’s the problem?” we wanted to know. Lynda described that … This is the view that you see when you come in across the bridge and that the area above here, which there is a masonry pier under here, has some cracks. I’ll show them to you in more detail, bigger cracks here.

These have been apparent ever since the house was built. Every time they patch the cracks, a little while later, they open up again.

This is what sets off an alarm bell in an engineer’s mind. It’s a concrete building, and always expect cracks in a concrete building. You would think, after 60 years, that the cracks would have stabilized. This seems to indicate that there’s a cantilever and it’s deflecting. As it deflects, it cracks.

I don’t know how many of you are engineers here, but many people think that concrete is an extremely rigid, solid material, which it is, and that it doesn’t move. That second part is not true. It does move.

There is a phenomenon in concrete which we call creep, or the real name for it, I guess, is plastic flow. It goes something like this: if concrete is under continuous load – being compressed or squeezed together – it will, in fact, compress itself slightly and respond to that load by shortening and getting smaller. If it’s a column, it will get shorter. If it’s a cantilever beam, it will deflect a little bit.

90% of that creep happens in the first year of the life of the building. The last 10% comes out after maybe 10 years. Then it stabilizes and you don’t get creep anymore.

The question here is what’s happening? The load on this building is not changing. It’s a dead load building. It’s the weight of the building itself. The people in here don’t weigh very much. They’re insignificant.

Why is it cracking every time they repair it? This is a question that we say to ourselves. There’s some investigation.

You can see the cracks again here. You can see them in reference to the masonry pier behind. Here’s the masonry pier. The cracks are right at the point in the cantilever, where the stress is the highest, if indeed this thing is acting as a cantilever hanging over the waterfall.

Down below, you can see the edge of the bolster here, in this triangular thing here. The masonry pier sits on the edge of that. There are concrete beams over it.

So far, the evidence has been anecdotal. They tell me it cracks every time they fix it, but what does that mean? We wanted to know is it really ongoing movement in the building or something else happening.

Silman2We suggested that they install some kind of monitoring system. On the left is something called a crack meter. On the right is something called the tilt meter. These are electronic gauges that measure movement. You can see the blue wires coming out of them. Inside that metal tube is a vibrating wire, and the technology is such that if it shortens or lengthens, we can record it to the nearest 1000th of an inch.

What we did is hooked it all up to an electronic system. On the right side is one week’s worth of data. We have the elongation, we have the tilt, and we have the temperature and humidity at the bottom.

If you count carefully, I think there are 7 peaks and 7 valleys because the movement is diagonal. As the temperature goes up during the day and the building expands, the crack closes. As it gets cold at night and the building shrinks, the crack opens. Same with the tilt.

There is a pattern to that. We would expect that to vary on a daily basis, but the question is, on a long term basis, is this crack really getting bigger? If it opens and closes every day, that’s not so serious. If it’s getting bigger and bigger and bigger all the time, that is something we want to know.

We monitored this for 17 months. The information was fed into data logger, which you can see on the floor here, downloaded into a computer, and the data sent to our office. If you really want to get fancy, you can have a telephone dial-up system and get it instantly. We didn’t need that, but they would send us the disks. We analyzed it over 17 months. Indeed, we found that the building was still moving and deflecting downward.

This is an alarming conclusion because when we did a survey, we found that one corner of the terrace was down 7 inches another down 5 1/2 inches over a short cantilever of 15 feet. This was a huge amount.Sillman1

When you’re in the house, you’d experience it. I don’t know if those of you who were there, whether they pointed out how the doors are all wrecked and they’ve been adjusted to take care of that. You can’t play marbles on the floor here. It’s all downhill. There’s nothing level.

The house is way out-of-whack. The question is it’s going more and more that way. The conclusion is someday it’s going to wind up in the river.

Did Mr. Kaufmann, the owner of this house, ever measure it? We should go back and say a little bit of the history of how the house got built.

Edgar Kaufmann was a department store owner in Pittsburgh. His son, Edgar Jr., was a painter and an artist, who decided he would like to study architecture for one year, not to be an architect but to get an idea of what it was about. He heard about what was going on at Taliesin, contacted Frank Lloyd Wright, and was accepted in the fellowship because he could afford to pay.

He went for a year, met Mr. Wright, and introduced his father to him. Mr. Wright did his father’s office in the Kaufmann Department Store in Pittsburgh because Edgar Kaufmann Sr. was a department store owner. Then, eventually, the site of the land which had been a summer camp for the women employees of the Kaufmann Department store and now became their vacation home, as a family. Mr. Kaufmann went to Mr. Wright and said, “Would you design us a weekend house there?” That’s how it all started.

After the house was built, Mr. Kaufmann realized that it was moving a little bit. He had his caretaker, Earl Friend here in the blue jacket, go out and put a stick between the second floor and the first floor and measure. Earl would come back and say, “There’s no difference. The stick fits tights all the time. I can get it in and out.” Of course, what Earl didn’t realize is that the 2 were moving together.

I reported this in the new book on Fallingwater that just came out this year. It has a chapter on the strengthening of the building that I wrote. I reported that anecdote.

If you go on to Amazon and you look up reviews of the book, Earl Friend’s son has criticized me very strongly for that statement. I really wish I could take it back now because I didn’t know.

According to the son, Earl said, “That’s the wrong way to measure. We should be measuring from the house to the stream.” Of course, Earl was right. He should have, but he had been told to do it this way. I don’t want to blame Earl for this. I blame whoever told him to do this.

It wasn’t that they completely ignored it. On the other hand, they really did, because over the 25 years or so that they lived in the house, the only had about 3 sets of measurements and nobody ever paid any attention to them.

Our monitoring was done over an intensive 17 months’ period. We found out the building was still moving. We made a report to Fallingwater. They had an advisory committee on the building.

Of course, we sounded a rather alarming note that this building is in danger. It could wind up in the river. It’s the Best All-Time Work of American Architecture and you’re the stewards of the building.

Their first question was: is it safe? How would you answer that? What would you say?

First of all, there was no change from day one in this building. Whatever was going on then was going on now. It was not a question of any deterioration. There was no rust in the building, the concrete is painted, the water doesn’t get at any of the concrete or the reinforcing, the load hadn’t changed, yet it was still moving.

We said, “We better do a little analysis.” We went back and we started to look at drawings. We found this drawing, which is the original reinforcing drawing that shows this upper terrace. The cracks that we noticed are right here, just to the left of this masonry pier, the maximum point of cantilever bending on it. This big thing is a cantilever. That’s its maximum point.

The story goes something like this: on the day that they decentered the formwork from this building – that is they knocked the props out – it deflected an inch and three-quarters. Now that’s about right if you were to do a calculation for what it should be. They should have cambered it up an inch and three-quarters. When they took the formwork out, it would have come back level. They didn’t so it deflected and, of course, it cracked here right away.

Everybody was alarmed. They called back to Taliesin. They said, “What’s going on here? Would you check your calculations?”

There were 2 engineers for this project. One was Frank Lloyd Wright’s son-in-law, Wes Peters, who was an MIT graduate, both an engineer and an architect, and a very good engineer. the other was Mendel Glickman, a Madison, Wisconsin engineer, also a very good structural engineer.

They called Glickman back who had done this calculations. He said, “Wait. Let me look.” the story is that he goes back and he looks at his drawings and he looks at this point. He said, “Oh, my God. I forgot the negative reinforcement.”

You can see this bar here, it just stops with a hook. It should have been continuous all the way out to the end. There isn’t any reinforcing here. Of course, there’s a crack. Not to be surprised about that.

On the other hand, did anybody ever do anything about that? We don’t know, but, “Oh, my God. I forgot the negative reinforcements.” There was some awareness that something was going on here.

We had these other drawings that, indeed, showed other kinds of reinforcing, including the metal window frames, which are here, which are going to play an important part in the story because you’ll see that they actually do serve to hold up the end of that cantilever.

We had the original drawings from the Frank Lloyd Wright archive out in Taliesin West. We decided what we needed to do is find out how much reinforcing is in these beams. When we realized that the window frames were also supporting the end of the floor, we realized we needed to look at reinforcing not only in that massive terrace upstairs where the cracks were, but also down on the main floor.

We took up some stones. They had done this before, by the way, because of the outside terrace is they had picked up stones so they could waterproof it. We got down to the top of the concrete beams here and we put chalk marks on them because we run across it with radar.

We didn’t want to start chopping in there and doing probes because we had the sense that the stress in these beams was really high. Any diminution of the strength in that beam by making a probe is something I didn’t want to chance, at least before I knew what was going on.

We sent one of our engineers – a small guy, this is John Matteo; he’s very nimble – down into the hole there. It’s only 2 feet deep. He was looking for cracks. He found some cracks. He had a flashlight down there with him.

Silman5We also got radar technicians out there, GBG. Here on the left slide, you can see them running the transducer across the beam in both directions. They are calculating where the reinforcing is. They can get a rough idea of how big it is as well. Up on that master terrace where the cracks were, where “Oh, my God. I forgot the negative reinforcement”, they’re testing that as well to see how much reinforcement there is in there.

When we got all done, we tried to compare the amount of reinforcement that we found in the output – and this is the output, by the way, which is something that if you were to read it, you wouldn’t know what it is, and I don’t know what it is, but these guys do – they know that that means it’s a reinforcing bar.

We found that there was a little bit more reinforcing in there than the original drawings had shown. We backed to the archives and we find a letter written by the reinforcing bar suppliers who were also engineers. They wrote this letter to Mr. Kaufmann, who had hired them. The letter said, “Dear Mr. Kaufmann, we are your suppliers of reinforcing. We’re also engineers. We don’t think there is enough reinforcing in the main floor beams of this house. We suggest that we put in more.”

Mr. Kaufmann is a department store owner. What does he know? He takes the letter, sends it out to Taliesin, to Frank Lloyd Wright, and says, “Frank, what do I do about this?” Frank Lloyd Wright replies in one of the most classic letters that any architect has ever written. It goes something like this: “Dear E.J., I have done more for you than any owner has a right to expect. If that’s not good enough with you, to hell with the whole thing. Signed, F.L. Wright.”

Here he is, Edgar Kaufmann, in this dilemma. Is his engineer right? Is his architect right? He turns to Frank Lloyd Wright and said, “If you say it’s good enough, that’s fine with me.” He instructs the engineer to put in just the steel that is on the drawing.

It turns out that they didn’t do that. They sneaked in a little bit extra steel. I think they probably did it at their own expense because they couldn’t charge him. Good thing they did that. It still wasn’t enough, we’re going to find out.

Now we know how much reinforcing is in there because the radar tells us, and we’re pretty sure of that. We’re also sure that the letter that Metzger-Richardson wrote was in here, too, and they put that extra reinforcing in. We now know the dimensions of the building and we know the reinforcing. We can actually do calculations and see what’s going on here.

We start out by doing a conventional analysis of that master terrace, the 2 big, long beams, as if they were acting as cantilevers. This is where the crack occurs here. Right under here is this masonry pier.

We do an analysis and we find out how much bending moment. It’s a simple problem, by the way, that anybody in Concrete 101 could solve, even in 1935. It’s not like it’s some kind of advanced calculation that nobody understood in those years because concrete was new. They very well understood how to make that kind of calculation. It’s a simple one.

We calculate the bending moment and then we find out how much reinforcing we would need. Of course, we know there isn’t any reinforcing because the drawings don’t show it and Glickman says he forgot it. There’s a big crack there. We know that the structure cannot resist this kind of bending. The analysis, therefore, of this as a cantilever can’t be accurate, but we’d do it anyway just as an exercise.

Remember the story of the 2 window frames. This is a picture of the living room looking at the south parapet. Four of these window mullions are bigger than any other. 1, 2, 3, 4.

If you were Frank Lloyd Wright, I’m sure you would have preferred the window mullions to be only this big, not this fat. He must have recognized that the cantilever was too big and that he would have had to prop up the end of that second floor.

We took off the covers of these windows and, indeed, in each of the 4 frames, there were little tiny steel Ts, 2 1/2 inches by 2 1/2 inches. We calculated how much load was in there and they just worked. They were just on the money that they could carry that much load without buckling. Even though they’re 2 1/2 inches, they were braced at the parapets. It worked.

We say to ourselves, “Okay. Now we can go ahead and feel comfortable analyzing this thing by putting these 4 supports in there.” We’ve exploded this thing vertically now so it’s easier to see.

We now do an analysis with relative stiffness of the various components and say if we put the load on the building, how would the moments distribute? We still get a very high level of moment here. We say, “That can’t be,” we don’t have enough reinforcing here to support that kind of moment. It can’t resist. We know that can’t be right.

Let’s assume that that section cracks, as a way of analyzing concrete. There’s a cracked section. We do that and we still get quite a lot of bending moment in that area, and we have no reinforcing there. Normally, it can take no bending moment.

We go ahead and we do a three-dimensional analysis. This is like a contour map of stresses because not only is it participating as a two-dimensional structure, it’s really participating as a three. We get pretty sophisticated.

We come out, finally, with an analysis that shows that there is no real bending moment left up here. There’s a tiny bit out here. That’s a simply supported beam. All of the load is now going down into the first floor.

Now we have to concentrate all of our energy on are these girders strong enough? Remember there were 4. This fourth one is not a player because it’s supported on a column out here, where there’s a staircase that goes down to the stream. That’s not a cantilever. We’re really only dealing with 3. We have these 3 beams that have a huge bending moment in them.

We now know the reinforcing in there. We do a calculation. I can remember the day that we’ve finished this and my engineer comes to me and he says, “Wait until you see this.”

The concrete is stressed to 95% of its ultimate capacity. The concrete is very good. We tested some of it. 5,000-pound concrete. It was [a site 26:16] next to concrete and it’s 5,000 PSI. Really strong. Again, thank God.

Its working stress is 4500. There’s no safety factor left. We don’t consider 500 PSI out of 5,000, 10%, as a real safety factor. It’s approaching failure.

The steel, which they had put extra in, is beyond the yield point, which means that if you bend it and then you take the load off, it doesn’t come back; it stays bent. These are horrendous kinds of numbers.

This is what we finally reported to the advisory committee. We said the house is moving, it’s going to fall on the water, and the stresses are at this huge magnitude. That’s when they said, “Is the house safe?” I pose the question to you. How do you answer that? That it isn’t any different from day one. They come back, “Is it safe?” I said, “I’m not even going to answer that question. I want to ask you a question. Are you going to fix it?” Of course, their answer was unanimously, “Yes, we have to fix it.” I said, “All right. When we go to fix it, we’re going to have to shore it up.”

“Instead of waiting,” – because they didn’t have any money; they have a fundraising drive – “let’s shore it immediately, then it will be safe, and you can take your time analyzing and deciding what to do, fundraising, and all of that, but let’s get shoring immediately.”

Silman4That was the next thing that we did. A system of temporary shoring was put in underneath those cantilevers. Again, in the stream bed, we had to divert the water. This is a historic stream. It’s got some kind of federal designation.

We set anchor bolts in the stream. We cored out 2-inch diameter rock cores, we saved them, and oriented them and marked them. We set the anchor bolts in grout. When we were all done, we drilled the anchor bolts out and we re-grouted the cores so all you see is a little angular ring. You can’t see it because there’s water in there, but you talk about being a Viennese teaspoon maker, as a preservationist. We really were.

On the left slide here, you can see the single line of shoring, which is thrust [light 28:25]. On the right is a common view in the sunlight, where all that goes into the shadow. You can see one shoring post on the left here. Most of the shoring disappeared, so people, still, were able to take pretty good camera shots of it. It was interesting because the shoring stayed up for 6 years. It wasn’t so horrible. It wasn’t great.

The shoring out of the end here was on the edge of this 10-foot thick rock cantilever. I didn’t want to chance the fact that we would put enough load in there to break it off and make Fallingwater back into a babbling brook. That would have been a disaster. We shored up the rock from underneath as well. That was a very inexpensive thing to do. That was also part of our shoring.

You can see what the shoring looked like here. Remember I said that last beam wasn’t a player because there’s a steel column here that supported it. That was one, but then there are the other 3 bolsters – 1, 2, 3 – that needed support.

This staircase comes down from the house so they could get into the stream. This is at low water in the summertime. We actually did this work in high water. They put sand bags and Jersey barriers in here and diverted the stream off to the left side while they worked.

All right. Now the house is shored. They’re raising the money. How do we fix it? We have some opportunities here to examine.

Word got out that Fallingwater was going to be repaired. Every architecture and engineering school in the country assigned it as a problem to their class. It was great.

We got responses from Saskatchewan, from the university of this, from the college of that. They were terrific ideas. Some of them was so wonderful. Some were fanciful, some are very serious. There was a whole series of propositions.

The hardcore preservationists said, “Let’s not do anything. You’ve got shoring in there now. Leave it in. Admit that the house failed. It will be part of the interpretive story. Everybody who goes through Fallingwater goes with a docent. The docent will describe that there was a mistake made, there wasn’t enough reinforcing put in this house in the beginning and we have to prop it up.”

This is an interesting solution, but when we ultimately arrived at our solution, we had a forum in the Carnegie Museum in Pittsburgh to explain what we were going to do. It was a 500-seat auditorium. There were 600 people in this auditorium. It was wild.

One of the judges, as it were, of our scheme – besides Nick – was Eric Lloyd Wright, grandson of Frank Lloyd Wright. Eric Lloyd Wright said, “My grandfather was interested in spaces. In this case, he was interested in the space inside the house. He was interested in the space outside the house. The space he wasn’t interested in was between the ceiling and the floor. Do what you want in that space but leave the rest of it as it was and emphasize the spirit of this house, which is cantilever. If you leave the permanent shoring in place, that concept is destroyed.” He was quite eloquent in his presentation of that position. I think he shot down the preservationists who said, “Leave the shoring in.”

The second was the most popular solution we got, which was, “Put in supplemental steel. On each side of those concrete beams where the joists frame in, cut the joists back, put in some steel to make it strong enough, reattach the joists, and put the floor back on it.”

That’s a perfectly valid and viable solution. It’s one that we seriously thought about because you could put a big enough piece of steel in there to act as a cantilever. It would work.

What it wouldn’t do is change any of the existing stresses in the house which were so high already. They were high enough to make me nervous even if we had supplementary framing in there. In addition to which it wasn’t a terribly reversible solution because you had to cut every joist. It was a lot of intervention work, but it was certainly feasible and certainly possible. We probably got 20 schemes from universities that suggested that.

The next 2 are schemes of actually bonding things on to the concrete. In one, you can bond steel. In one, you could bond fiber reinforced plastic that is carbon fiber or fiber glass. Both of those are very viable solutions and have been used a lot in concrete repair work, particularly FRP repairs now – you use a lot in seismic strengthening and seismic repair work. Even bonded steel, we have such great adhesives now. The epoxies are wonderful.

Those were possible solutions. We could have worked either on the sides of the beam or even on the top of the beam, where you really wanted to get that reinforcing because it’s a cantilever, because we had 4 inches of space where the old wood sleepers were and then the stone floor. We could have done something on the top.

That, too, would not do anything to reverse the high stresses in the concrete already. It would only take additional load. Any additional load that went in there would feed into the new stuff and it will be shared on this relative stiffness spaces. We weren’t too keen on that because, again, it would introduce even more stress into that concrete that was there.

The final solution was the one that we actually came up with the first. It was my original idea. As soon as I saw this thing, I got this neat idea of can we do some kind of external post-tensioning. You’ll see what that is in a minute. That has the advantages of not only adding reinforcing, but of being able to reverse the stresses that are in the concrete and the steel now. That’s the solution that we came up with. Let’s see how that works.

The top slide here shows the existing condition. You have this inverted T-beam. This upside down T-beam. You can see the reinforcing bars in here. They’re in the top of the beam.

As the load comes down on here and it bends down and deflects – this is the point that goes down 7 inches over the 15 feet – cracks open up on the top here as this thing stretches. You can feel the top of that beam in tension, I think. Concrete, of course, is not able to resist tension. That’s why we put reinforcing bars in it. That’s why it’s called reinforced concrete.

Our idea was, on the bottom, to somehow attach steel cables to the sides of the beams. We would have to drill holes in the joists, by the way. We wouldn’t be able to just fit it in, but it would be a lot less invasive than chopping them out completely.

Somehow attach steel cables to the side of the beam, stick them out the front parapet, anchor them at the back, pull them tight, wedge it off. In this way, you’ll be able to actually reverse the stresses by picking the building up a little bit here. If you can feel yourself going (grunting sound), you can just feel yourself pulling the end of that beam up.

Imagine you’re standing up there and the cable is a little bit higher. You can actually pull it up. That’s what we do. You’ll see by the explanation exactly how that works.

This is the plan view now. The one bolster we’re not worried about is this one that’s cut out like a keyhole with the stair that goes down. The other 3 – 1, 2, 3 – are the 3 main beams. Here’s the fireplace boulder.

We’re going to put red cabled on each side of this beam, each side of this beam. We can only put them on one side of this beam because, here, we have the stair going down, and there’s no way to get it in. Also, it turns out the stress was a little bit lower here.

In addition, because this is really a double cantilever and this terrace’s cantilever is the other way, we’re also going to post-tension it across the other way with much smaller-sized cables, but to hold up the terrace as well and try to limit the deflection out there.

We’re going to put very high strength bundles of steel wire called post-tensioning cables on the sides of these girders. We’re going to anchor them at the back here in blocks of concrete. We’re going to have a middle block of concrete that’s a diverter block that causes this thing to raise up and gives it some drape to it. Then we’re going to have, at the very end, at the south parapet here, something called the anchor ridge block, where the jacking end is. The cables are temporarily going to stick through the front of this thing so we can grab them with a jack and pull them.

In this section, the main cables are here. This is the south parapet. They’ll temporarily stick the cables through here. They’re going to run along the side of the beam. They’re anchored in the back to a block of concrete. The diverter block is here. The jacking block is here.

Here’s the stream bed with the water, the edge of the bolster. Here’s where the crack was. We also were going to put some fiber glass rods up there, just so it never cracked again.

These are the steel Ts that hold up the edge and come down the south parapet. This green is the temporary shoring that will go away when we’re all finished. That’s the house.

In an isometric, the sketch of the inverted T-beam where the reinforcing bar that are there that are not enough. This big, red bundle is 13 strands of 1/2-inch diameter twisted wire. Each strand is good for 40,000 pounds of tension pull. 13 of them. You can do the multiplication. A lot of force.

Then the very fine red line’s going the other way, or the cables that go across to the 2 terraces. You can see the stone floor is on top of that. There’s some sheathing and some sleepers. We’ll see what that does here.

This is the jacking end here, where there’s a piece of machined steel here through which these cables come. They put a jack on there and they pull the cables tight. They drive wedges in there so the cables can’t slip. They take the jack away, cut the cables off, and patch the hole. Now the cables are in a huge tension force and it has the effect of picking up that drooping in.

How does it work? They had to rebuild the scaffold to get at it. This time, instead of having all that hammered-together wood, we have very elegant … This is actually aluminum scaffolding that’s very lightweight and easy to install. Again, done some of it in the river bottom, at the base of the falls, some of it up on top of the falls to give them access.

In addition, it’s not really to trek up the whole house, all of the access is going to be done from outside, so they built a little bridge. This is the main entrance bridge that you go across to get into the house because the door is in the back. Typical of Frank Lloyd Wright. You can never find it. The access for the workmen was going to be across this bridge.

They also had to take out all of the built-in furniture, all the stone on the floor to get that out of the way so we could get at the concrete. Again, it was the easiest way to get it out across this temporary bridge.

Every stone was numbered and photographed so the orientation was understood. As I say, the workmen from Fallingwater had actually taken up outside terrace stone before to put waterproofing in. They knew how to do that. They took the stone up, they took up the diagonal floor sheathing, which was actually redwood. You can see then the sleepers underneath that. They took all of that up.

Now you’re down to the bare concrete, which looks like that drawing that I showed you right in the beginning, where you can see the concrete beams and the joists. Then there are these little bricks on top which were just leveling devices for the sleepers.

Now we’re down to the bare raw material of the house that we are working on structurally. We brought the radar guys back because now they had full reign to go over the whole thing. They checked the reinforcing. We just did one last check of that. We were pretty sure of what we had.

There were a lot of cracks that we found in the house. We decided that if we’re going to put this huge post-tensioning force in there, we don’t want to waste any of it by closing up cracks. Let’s fill the cracks first so that every ounce of post-tensioning we put in there will be used to reverse the high stresses that are in there now.

That’s what they’re doing here. They’re pumping epoxy into the cracks. You see these little ports. They’re plastic ports. There’s a little tiny jack here with a pump. They’re pumping the cracks full of concrete. This is in the ribs, but they did it in the tops of the beams here, too, as well. All of the cracks were filled first.

Now we have to attach these new blocks onto the existing beams. We have the end anchor ridge block, the middle diverter block, and the forward jacking block. 3 blocks on each side of the beam.

We’re going to do that by drilling through the beams. You can see the core drill here drilling holes. There’s one drilled here, one drilled here. They’re going to out very ultra high-strength rods in there, cast concrete, and turn the nuts up with a jack and turn the nuts.

They’re going to post-tension those against the sides of the existing beams so that when you pull on that, it’s going to lock on to the beam in an action called shear friction, as well as these dowels working for you. It’s sufficient that we can resist the pull and put that force into the existing beams.

These are going to be new concrete blocks. They’re starting out to drill them here. You can see they have 4 bars in here. These are the trumpets that will let us get the jack in.

This black thing is a hose. It’s a hollow duct. We’re going to thread the 13 cables through that and splay them out and anchor them in the concrete block here.

The spiral is because the stress is so high that it wants to burst the concrete apart. We hold it together with a spiral there. You can see these things are the tubes that go in here so the post-tensioning rods can slide when they pull them.

They’re getting ready. They’re going to eventually grout all the post-tensioning rods. These tubes will get grouted in there.

You can see, sticking out in the back here, each of these wires. There are 13 of them that are splayed out in here. I see 1, 2, 3, 4, 5. There’s 13. Trust me. They come out the end of this black hose.

We have all these jacking-ins here. You have the nuts that are going to be run up tight. You have the jacking blocks here. Everything is ready to go to pour that.

In case the blocks with wood formwork, you can see the duct here, the black hose alongside here. The jack’s going to go on the outside of this. This is the diverter block. They haven’t formed this one in wood yet. They formed it on this side, but not the other side.

Here they are with the concrete hose. It’s a very dark slide, but they’re pouring the concrete into these blocks now. After that, they’re going to post-tension them by tightening them up against the side of the beam.

Here you can see it’s done. Here are the post-tensioning rods. 1, 2. There’s 2 on the other side. You can see another one here. That’s what squeezes these blocks on there. You can see the black hose still there, the duct hose here. Notice there’s a shoring tower here and here.

We were nervous. We were putting so much load in this that we were worried that we might overload these window mullions, these little 2 1/2-inch Ts. If they broke or failed, the house could come down.

We decided the better part of discretion would be to build these little shoring towers. If they broke, nothing would happen. We could always replace them. They didn’t break.

Here they are with a very small jack, tightening the monostrand cables and jacking them that hold up the terraces. These are the small cables.

This is the real story. The 5 bundles of 13 cables each that stick thought the south parapet and come out on to the scaffolding, they’re protected in plastic here, but you can count the 5 bundles. Each one has 13 1/2-inch diameter strands of twisted wire. There they are.

You’re looking at a machined end of the jacking fitting, which is a permanent part of the installation. That’s going to stay in place. You are looking at a split wedge here. This is actually after they’ve been jacked, but it explains how it works.

This wedge is just a piece of high-strength metal that, when they pull these wires tight, there’s a ram that drives the wedge in tight. When you let go of the wire, the wire is nicked by that wedge and it’s pulled tight against the face of the steel here and the strand can’t slip. It’s a friction fit that causes this thing to work.

It’s the best sign of the house. We did this in April 2002. I was pretty well-stressed out, I have to tell you. We just had 4 months of working at the World Trade Center, every structural engineer in New York, of trying to recover the bodies of the firemen. It was the worst job I ever did. I was emotionally wrung out that winter.

Then we come out here and working on the number one structure in the United States and worried that if we didn’t get this one right, that would be the end of my career. Stressing in progress. There was plenty of stress going on.

This is what we did. Here is the hydraulic jack that applies 390 tons of force to each girder. Think of the amount of load that’s in this thing.

This handle here is the ram end that moves. That’s the jacking end. It moves because there’s hydraulic fluid in these 2 black hoses.

Safety precaution says that nobody’s supposed to be standing out there because if you post-tensioning this thing and some of the anchor ridge fails, it’s going to whip one of those cables out and it’s going to be like literally a horse whip and anybody in its path will get killed. Of course, everybody’s out there. We all wanted to know what was going to happen, what was going to go on.

That’s me here in the red jacket and the white hat. This is mostly the construction crew, but you can see there’s people filming with a microphone here. There’s all kinds of people. This guy is the VSL superintendent. Then the guy’s measuring on the jack.

I want to show you something here. This is a film clip. I have to get out of this for a moment.

Here they are. They’re putting an extender on the jack because they actually can’t fit into that little square hole we’ve cut in the end of the parapet, but they’ve threaded the 13 strands through this piece of extender. They’re going to get ready now to mount the jack on here. There’s some loose wires at the end that they’re twisting just so it doesn’t get in anybody’s way.

Here’s the jack. This big guy in blue is, unfortunately, huge and I couldn’t film through him. You’re going to see these strands come out where his right hand is, where that gloved hand is.

The guy in the red shirt is the measurer. They measure the amount of force that’s put into these tendons in 2 ways. One is the amount of hydraulic pressure that you put on it. The second way is how much the strand elongates. They know how long it is. This thing was about, I don’t know, 30-something feet long. They know, given the amount of force we’re expecting, how much it should elongate. He’s got a ruler, and he measures that, too.

They’re putting this thing on. You can see in a minute that the cables are going to stick through the back end of it. There they are. They’re stuck through the back.

He’s going to grab that handle every once in a while and shake it. That’s the ram end of what moves. This handle here is actually the jack part that pulls the cables. Everything is done by friction here.

He’s hooking up the hydraulic cables now. That’s what’s going to give us the force. It’s a relatively small pump.

This was the measurer. He just wants to make sure with his tape measure. He’s just checking every last thing.

We stress these cables, 1, 3, 5, 2, 4. First 10% of the total load, just to set it. Then another 40%, so that’s up to 50.

All right, here he goes. He’s opening the valve. We leave that overnight at 50%. Then the second day, we do the last 50%. Overnight, you see if there’s any losses.

Watch that needle go here. There it goes up. Watch the hose straighten up over here. You can see that, up. There it goes. There it goes. There it goes. He’s waiting to get to a certain pressure before he shuts it off. Done. 20 seconds, it’s all over. 7 years of work and it’s over in 20 seconds. It was really quite amazing.

We left it overnight for the first night and come back the next day just praying that you had the same amount of force in there. One way we were able to do it was to measure with stress gauges.

This is a Penn State student doing his master’s thesis, by the way, with remote strain gauge technology. A very interesting project that he was doing. We tried to get as many people into this as possible.

Somebody said to me, “Did you try to pick the house back up and get rid of the 7-inch deflection?” The answer is no way. The house had adjusted over the years. If we tried to ever lift it back up again, it would crack every single finish, every piece of glass, all the plaster, all the doors that had readjusted and been reset. The answer was no. We were going to leave it in position.

It did pick up 3/4 of an inch, which is what we calculated, but that was all. There was no way we were going to try to set it back.

He’s measuring now how much the strain was in it. We had all kinds of other strain gauges on it to measure how much deflection.

This is the temporary shoring beam. In fact, you can’t see it easily, but it did pick up off the shoring here by about 3/4 of an inch. That part was perfect. Here, it’s a little better.

The shadow line that you can see here wasn’t there before we did the post-tensioning. It was tight down there, so we knew it had picked up here. It was just about what we calculated. We all breathed a sigh of relief.

They came in, they put the sleepers back – new sleepers – new floor – it’s plywood this time, not [vinyl 52:39]. The stone goes back. They brought the furniture in after they conserved it.

This was before they repainted it. You see some of the marks on the bottom after they took the shoring out. Basically, the house looks exactly the same as it did when we started.

All right. Interesting. When we get all done, or even in the middle of it, we say, “How did somebody make a mistake like this?” This is a simple calculation, as I said before. 2 really good engineers. How could they screw this up? How did they make a mistake?

My good friend, Edgar Tafel, he’s very fond of telling the story how Frank Lloyd Wright only saw the site once. Before the house was built, Mr. Kaufmann brought him out there. He saw it and he said to him, “EJ, get me a topo of the site.” Of course, they ordered a site topo. It was delivered to Taliesin.

Several months go by and EJ Kaufmann hears nothing from Frank Lloyd Wright. One day he’s got to go on a business trip. He calls up from his office in Pittsburgh. He says, “Frank, I’m coming to Milwaukee on a business trip. I thought I’d drive over to Taliesin and have a look at the drawings.” Frank Lloyd Wright says, “That’d be splendid, EJ. You just do that.” He said, “I’m leaving tomorrow.” “Okay.”

Frank Lloyd Wright gets out the site plan and he puts it on his drawing board and he walks past it. All the apprentices are aware of this phone call and of what’s going to happen in 3 days.

The next day, he gets a call from Chicago. Mr. Kaufmann says, “Frank, I’m in Chicago. I’m on my way to Milwaukee. I’ll see you in a couple of days.” Frank Lloyd Wright said, “That will be splendid, EJ. We’re looking forward to seeing you.” He walks past the drawing board with the site plan on it.

The next day, he gets the call from Milwaukee. He says, “I’m going to leave tomorrow morning. I’ll be over for about lunch time.” “Splendid. See you then.” He walks past the drawing board.

The next morning, about 7:30, Mr. Kaufmann calls from Milwaukee and he said, “Okay, Frank. I’m living at about 8 or 8:30. I’ll see you at lunch time.” “We’re looking forward to seeing you, EJ. We’re going to show you the drawings. It will be wonderful.”

This time, Mr. Wright sits down at the table. All the apprentices gather around him. They were all standing there with their knives and their sharpened pencils for him.

Frank Lloyd Wright is this whiz-bang draftsman. He is so fast. There are films of him drawing and you can’t believe it. It’s like a computer. He just do so fast.

He takes a big piece of tracing paper, overlays the site plan, and the apprentices are astounded to see that he’s placed the house on the top of the waterfall, not on the bottom looking up at it, which everybody assumed would be its location.

He draws this 4 bolsters that we see, lays out the foundation, takes another sheet of tracing paper, puts it down, draws the first floor plan, puts another sheet on it, draws the second floor plan. There is a small third floor and a roof. This takes him the better part of the morning.

They’re sitting there astounded because what’s coming out of this is just this whole house is pouring out. A butler comes in and says, “Mr. Wright, Mr. Kaufmann’s here.” Frank Lloyd Wright turns to the apprentices and says, “Okay, boys. While we’re at lunch, I want you to draw the elevations. Here’s what it’s going to look like.”

They go to lunch. Edgar Tafel tells this story, said, “I could draw about 1/10 as fast as Mr. Wright. I took one elevation and Bob Mosher took the other one. We struggled to get this thing done in an hour and a half. Wright had done the whole house in a morning.”

They got the elevations done. They brought Mr. Kaufmann in. He was absolutely dazzled by this. He said, “How fast can we get started?” The answer was, “We can start right away.” They didn’t have any working drawings. They had this set of stuff.

Frank Lloyd Wright, he was like Mozart. Mozart carried symphonies around in his head and would sit down and write them. Apparently, Frank Lloyd Wright did the same thing. He must gave carried this building around.

People say that he had made sketches on the fly in his bedroom before. None of them has ever been found. I don’t know if that’s been verified. For sure, there were no working drawings or big drawings.

At any rate, they started this house very quickly after that concept set was done. There was a complete set of drawings, but we have the feeling that it must have been done so fast that something got lost in translation.

We’ve all done work, those of us who were professionals in the design business. We say, “Oh, I haven’t got time to do that. I’ll come back to it.” I have a feeling that’s what happened, and they never came back to it. It got built wrong.

I hate to blame anybody because they were first class engineers, these guys. Anybody can make a mistake.

We hope we fixed it. We hope we fixed it for all time. We hope it will still be the Best All-Time Work of American Architecture. Thank you.

National Center for Preservation Technology and Training
645 University Parkway
Natchitoches, LA 71457

Email: ncptt[at]
Phone: (318) 356-7444
Fax: (318) 356-9119