The stair's concrete treads are threaded on a steel column and locked in place by steel balusters. Each of the 300-lb. steps was cast individually in a mold lined with plastic laminate.

This circular staircase was inspired by the stone stairways built in Europe during the Middle Ages. Designer Paul Tuttle wanted to create the sense of timeless solidity that massive stone steps evoke, and the two-story greenhouse in the Douglas residence overlooking Santa Barbara, Calif., gave him his chance. The room needed both a stairway and a sculptural focus, so Tuttle captured the medieval aura with the l0-ft. high, 7-ft. dia. concrete spiral stair shown in the photo at right.

Even before I had seen the design, Tuttle asked me if I would be interested in building this staircase. My first impulse was to say no. A poured-concrete spiral stairway seemed impossibly difficult. But once I saw his drawings, I got excited by the challenges its construction presented. Along with two of my associates, I agreed to build it. But I still wasn't really sure I could.

Layout - The first thing we had to do was to determine the number of steps, the rise of each step and where the beginning and end of the spiral would fall. We eventually decided on a landing and 15 steps of 7 ½ -in. rise. Each one is a 25° segment of a circle.

Next we made a full-scale mechanical drawing of one of the steps, including sleeves for holding the balusters, a cutout for toe space, indentations for carpeting and a sleeve for the center column. This drawing (left) proved to be an indispensable reference in all stages of the project.

Making the concrete form - After we figured out what one of our concrete treads had to look like, we had to build a form that would duplicate it 15 times. The form had to be durable, yet flexible enough to be taken apart after each pour and reassembled for the next one. Carpenter John Bunyan and I thought about our options and eventually came up with the one shown below.

The form's main components were the base, the sides, the pivot end and the outer end. The base was a 3-ft. by 4-ft. piece of 1/2-in. plywood that we covered with plastic laminate (for more about working with plastic laminates, see FHB #9, pp. 39-41). We made the form's sides out of fir 2x8s, with 2x2 blocks screwed and glued to their ends. We made the tight-radius pivot end from five 2x6 blocks laminated together with yellow glue. We cut out its 4-in. radius arc on the bandsaw. The outer end was made up of five 2x4s. It had a 21-in. arc of 3 1/2 -ft. radius cut out of it. We let the tails run long on the end pieces so that they could be bolted to the end blocks on the sides of the form.

Once we had the two sides and the tight- radius end piece bolted together, we lined its inside face with plastic laminate. This part of the form stayed together as a unit throughout the casting process: We lined the outer end, and then bolted it to the side pieces. We put the entire assembly on the base, and screwed the two together with 26½-in. screws.

Next we screwed plywood pieces to the bottom and lead edge of the form to create an indentation for carpeting. The shape of these pieces had to be carefully worked out so that the carpeting would flow from one step to the next without any offset. We planned for the bottom of the form to be the mold for the top of each step, so we could place the largest of the ½ in. let-in pieces for carpeting on the bottom of the form. Pouring the steps upside down also let us trowel the bottom of each tread-the largest and most visible expanse of concrete on each step. Finally, a 2½-in. square piece of wood 28 in. long was covered with laminate and screwed to the top edge of the form. This would form the indented toe space on the bottom lead edge of each step.

The last parts of the form were the registration pins for the steelwork-two 1 1/8 in. tubing sleeves to support the rebar, and the 4-in. pipe sleeve that would fit over the central column of the stairway. We screwed two 7 ½ -in. lengths of I-in. wood dowel to the base near the corners at the wide end of the form (drawing, above right). At the other end, we attached a 7½-in. piece of 4x4 with chamfered corners. These pegs, carpet inserts and toe board were screwed in from the outside of the form so they could be easily released when we stripped the form from the cured concrete.

Structural steel work and balusters - We decided to cast a short length of 3-in. pipe into the slab-floor footing to act as an anchor for the 3½-in. central-column pipe. Each step would have cast within it a sleeve of 4-in. pipe. These sleeves would slip over the center column and become an integral part of the re- bar assembly in each concrete tread. The re- bar grid would also include the two 1 1/8-in. tubing sleeves into which the railing balusters would slide and be secured, as shown in the drawing above.

The sleeves in each step are attached to one another by a matrix of 3/8-in. rebar. Fifteen grid assemblies were required, one for each stair, and each one had to match all the others exactly in order for the balusters and central column to fit properly.

To achieve this kind of accuracy, our steel expert, Dean Upton, welded each assembly on a jig. In a 3/8-in. steel plate, Upton drilled holes at the sleeve centers and tapped them for ½-in. bolts. Then he tack-welded posts to the plate. These posts had been turned to fit the inside diameter of the sleeves, and were bored to accept the ½-in. bolts. The sleeves for each tread were cut to length and squared. Then Upton bolted them to the jig and welded the rebar in place.

The posts that support the handrail are 1-in. OD steel tubing with a .083-in. wall. The sleeves into which they fit are 1 1/8-in. OD tubing with a .049 in. wall, allowing a clearance of .027 in. That seemed a bit loose to us, but proved to be a necessary margin during the final stair assembly.

Upton silver-soldered a ¼-in. ring (from the sleeve material) to each baluster to make sure they ended up at the right height. Their tops were cut at the angle of the stairway, and as they were installed, the balusters were rotated to match the direction of the handrail. Once the treads were in place, we plugged the underside of the baluster holes with Bondo (a brand of auto-body putty).

Picking out the pipes - Material selection required some careful sleuthing. Our basic plan called for several pipe sizes, each to fit snugly over the next. The central column is 3½-in. pipe, schedule 40. The sleeves for the steps are 4-in. pipe, also schedule 40. Pipe goes by the inside diameter (ID) while tubing goes by the outside diameter (OD). With pipe, oddly enough, the OD is constant while the ID changes with the schedule or wall thickness. A 3½-in. pipe (schedule 40) is actually 4-in. OD by 3.548-in. ID, with a wall thickness of .226 in. So depending on the schedule, we had a choice of clearances between sleeve and column. With 3½-in. pipe and a 4-in. sleeve, there is a theoretical clearance of .026 in. But this clearance did not prevail on all of the pieces so some sleeves had to be turned down on the lathe.

Concrete technology - The concrete mix was critical for two reasons: weight and finish texture. In order to decrease the weight of each tread from more than 400 Ib. using concrete with standard aggregate to about 300 Ib., we used !/2-in. Rocklite (The Lightweight Processing Co., 715 N. Central Ave., Suite 321, Glendale, Calif. 91203), a lightweight aggregate. But we also wanted a dense, pure white finish on each tread. This led to our using two different batches for each one. The outer inch or so is made up of 1 part white portland cement, 2 ½ parts 60-grit silicon sand and 2 ½ parts Cal-White marble sand (used mainly for swimming pools), made by Partin Limestone Products Inc. (PO Box 637, Lucerne Valley, Calif. 92356). Once we got this outer layer of white concrete in place, we filled the core of each tread with the lightweight mix.

We carefully measured all the ingredients because slump was important-too much slump would cause the two mixes to flow together in the form. The lightweight mix for the core was 1 part cement, 2 1/2 parts sand and 2 parts aggregate. To speed setting time, we added a little calcium chloride to each batch.

Originally we'd hoped to pour two steps per day, but found that producing one a day was quite an accomplishment. Placing the mixes in the form required two of us-one to tamp the outer mix and the other to keep the core mix from migrating to the edge of the form. We placed the concrete in layers, agitating it thoroughly after each layer to eliminate voids. Between each pour we cleaned the form, coated all dowels and wood insets with floor wax and sprayed the plastic laminate with silicone.

Surprise and delight filled us when we stripped the form from the first tread. The result was magnificent, but not what we'd expected. The plastic laminate made the surface smooth as glass, and a swarm of tiny, irregular air pockets made it look something like travertine. We were elated with this first success.

Assembly - Once the 15 treads were cast and carted to the site, our next hurdle presented itself-slipping the 300-lb. steps onto the steel column. Obviously we needed some type of device to lift the treads. It would have to be sturdy enough to carry the heavy loads, yet also adjustable so we could fine-tune the position of each tread over the center shaft.

Our solution was the homemade chain hoist shown in the drawing at left. It consisted of a 12-ft. 4x8 I-beam and a chain hoist mounted to a freewheeling trolley. We centered the beam over the column, and held it up by the stair landing on one side and a sturdy frame- work of 2x6s on the open end.

We fabricated a special metal carriage and harness to carry the treads as close as possible to their center of gravity. Each tread was then lifted above the l0-ft. high column, and its sleeve was centered over the shaft. Then the tread was slowly lowered into position. Once a step was in place, we would brace its outer end with a 2x4 and then one of us would tap a baluster through the aligned sleeves to secure the new tread to the one below it.

Disaster nearly befell us midway in the assembly. As we rolled the tenth tread along the I-beam, a lurch in our movements caused a sudden shift in the tread's center of gravity. Instantly the harness slipped off the carriage and the step plummeted, bouncing against several of the steps already in position and demolishing the bottom picket on its way to the floor. We were stunned. Fortunately no- body had been standing in the way of the tread when it fell. We surveyed the damage and it appeared enormous. Chunks of concrete were knocked off the treads in half a dozen places. We were so badly shaken that we packed up and went home for the day, believing the project ruined.

The next day we reassessed the damage and concluded that it wasn't as severe as it had seemed. We decided to patch the damaged edges and corners with Bondo. In some places we had to build up numerous layers of the stuff, but it worked far better than we had dared to hope. Because the Bondo was a different color from the pristine white concrete, we knew we'd have to paint the final product.

At the landing - The top step had a different shape from the others because it needed to flow into the cantilevered landing. To link the stairway to the landing, we built a triangular rebar grid to lock the top of the central steel column rigidly to the 4x12 landing girders. We welded sleeves for the balusters and the central column to this grid. The grid in turn was welded to a 4-in. by 30-in. by 3/8-in. steel plate, which was bolted to the landing framing. Then we erected a form around the grid with supports down to the floor. We were able to use several curved components from our breakdown form, but most of the pieces were new and had to be covered with plastic laminate.

We poured this last step with the same two mixes and care that we used with all the other treads, and when we took off the forms it flowed perfectly into the landing.

Handrail - Probably the most challenging part of this project was bending a 1 7/8 -in. dia. thin-wall tube (.063 in.) into a helical handrail. We chose this size tubing because there are stock fittings for 1½-in. pipe that fit closely enough to be used with the tubing (1.875-in. OD vs. 1.900-in. OD). At the top where the stairs meet the landing, we needed a tight re- turn bend to blend the rising stair rail into the horizontal landing rail. We made this transition with two wide-radius elbows and a little cutting and fitting. We used another stock fit- ting-the half-sphere cap-to finish the bottom end of the handrail, and we used floor flanges to attach the landing rails to the wall.

The radius of the stair circle was 42 in., but the radius of the line of balusters was 40 in. The inside radius of the handrail was 40 in. less 15/16-in. (half the diameter of the tubing) or 39 1/16 in. Taking his cue from a tubing bender, Upton designed a press that used a hydraulic jack to generate the bending force needed to arc the straight lengths of tubing. He used an oak form block with a radius of 38Va in., a little tighter than the required radius to allow for some springback. As it turned out, the springback was almost nil.

The forming tool we tried out first had two spools about 12 in. apart. It bent the tubing, but it also left slight dimples at each point of contact between spool and tube. A handrail with a dimple every 6 in. was totally unacceptable (it looked like a segmented worm), so Upton made a pair of bending shoes out of channel steel and Bondo as an alternative. They needed periodic greasing to allow the tubing to slip through as it was bent. This jig, shown in the drawing above left, worked fine. A 5-ton jack supplied the pressure.

Upton first tried to form the helix as the tube was being bent by rotating the tube a little at each bend. But it was difficult to keep track of the rotation. We could calculate how much rotation was required, but to control it was tough in a small shop. Even though it wasn't the right shape for our railing, the sculpture resulting from the first try could be mounted on a stone block and placed in front of a library.

We learned two things from this attempt. One, the press could put a wrinkle-free radius in our tubing, and two, trying to form both the radius and the helix into the full-length railing was too ambitious. Instead, Upton cut the 24-ft. tube into three 8-ft. pieces. Then he made a plywood jig that had a radius of 39 1\6 in. (drawing, above). This jig represented about 1/2 of a turn of the staircase, and was tall enough to allow the rise of the handrail to be marked diagonally on it. As he shaped each 8-ft. section, Upton checked its bend against the jig. This worked well, and the three pieces closely approximated the required helix plus the radius.

To make final adjustments in the helical twist, each 8-ft. section was cut into three equal pieces. After tack-welding the first section to the bottom balusters, Upton rotated the second section slightly to create the helix. Section two was then tack-welded in place, and the third piece rotated slightly more than the second and so on until all nine parts were tack-welded in place. Each piece was aligned with its neighbor by using a short offcut of 1% in. tube as a dowel. Before he welded the balusters to the railing (photo above), Upton torch-welded the whole unit into one continuous piece. Then all the welds were ground down, and any little pits were filled with Bondo and sanded smooth. The finished rail is painted brick red, and appears to flow as one piece from top to bottom.

Dean Upton torch-welds the handrail to a steel baluster. Although It appears continuous, the railing is composed of short segments of steel tubing that were bent on a homemade press, and then assembled on site to create the necessary helical shape.

Our final job was whitewashing the treads. We wanted to preserve the texture of the concrete and to have it not look painted, so we experimented with several finishes. We finally settled on white latex paint mixed with a small amount of white portland cement. This gave the surface a little roughness to the eye, but did not destroy the glass-smooth texture to the touch. One coat completely covered the grey-green Bondo, and we were done.

The project took six weeks of concentrated effort, and it kept our attention with a series of snags and surprises. But everybody is happy with the way it turned out. The stairway cost almost $10,000-a lot for one flight of stairs, but not for a sculpture that anchors a special room.

Dennis Allen is a general contractor living in Santa Barbara, Calif.