What does spelunking have to do with ceilings? Spel – what?
Spelunking—the exploration of caves. Not much, usually, but if
you read on, you’ll find that not all ceilings are Plain Jane horizontals,
as some contractors stated when asked whether they had
been challenged by any ceiling projects. Luckily, some architects
explore outside “the box” concept and give contractors something
to scratch their heads over.
“Anything that needs to be framed in a radius with some veneer
board over it and then plastered is probably the hardest system
to install,” states a Californian, a view echoed by a fellow Californian
and probably held by many others. Thomas Engel of
Shepherd in California sent in pictures of tricky ceiling work his
company has completed. The Dublin Library [Pix 3a] shows
framing of a rotunda that was 30 feet up, with a segmented dome
lid and radius soffits surrounding, which were chamfered. The
challenge was to lay it out so that all areas were backed for the
gypsum board installation, and the soffit and light coves were
perfectly round as they tied into the dome ceiling.
A Colorado contractor has some words of wisdom for this kind
of project: “Drywall soffits or radiuses built out in an area to go
around with ceiling grid can be difficult. You just need to pull
enough string lines off the existing area that you can cut around
to maintain the straightness. Many people try to hurry by guessing.
Experienced guys can just put an eye on it and tell if it’s
straight for a little ways, and then they’ll pull a string on an area
and make sure that it’s right. But many piecework-type companies
cut corners, resulting in sight-unappealing work without
true lines.
STRAIGHT LINES ARE NOT SO EASY EITHER
“Linear ceilings,” he adds, “al ways have a degree of difficulty,
although I haven’t seen too many linear recently-probably
because they are very expensive systems that are expensive to
install because they are so complex. You have to put up the carriers,
and the panels snap to them. You have filler strips in the
middle pans and have to make sure these strips don’t cover the
areas for the ventilation.
“One challenge with wood ceilings in remodeling projects is trying to match the color of an existing wood ceiling. Most own ers or GCs don’t understand that W light does hit inside surfaces and lightens or darkens them. Then there’s dust and dirt
in the air. And in the old days, cigarette smoke would change
the color, too. You have to take a piece of the old wood and
match it as closely as possible by tinting the stains.”
Another Coloradoan also found a challenge in wooden ceilings,
but in this case, a particular system. “On the Pentax building
project, we used a Conwood Natural Line linear wood ceiling
of maple, which ran wall to wall with a 3/4 inch reveal all the
way around. In some places, a single piece of wood ran wall to
wall—the walls being radiused to boot.
“Trying to snap the wood under the clips that hold it in place
was really difficult. The clips were spring steel that snapped into
a groove in the back of the wood. But they weren’t strong
enough: half the time, they bent instead of popping down into
the groove when you pushed them. We learned to push up at an
angle. But once snapped in, you couldn’t move the wood at all.
“To ensure we had the wood correctly positioned before snapping
– into place, therefore, we cut the wood longer than needed, and
– then made a scribe out of a piece of the wood and a razor knife
blade, scribing both radiused ends, and
then cut it. We then used that 3/4 inch
spacer on one end to try to hold the gap
that we needed while we snapped it into
place. The job took us about five times
longer than anticipated. I was not familiar
with the linear product and had bid it
in terms of production time like a metal
linear ceiling, so I was way off base. But
it was a beautiful ceiling when done.”
SSSHH!
For an Arkansan, the challenge with one
ceiling was not in complexity of design
but stringent sound-dampening requirements: “We installed sound ceilings for
a church group that the project manager
lost a lot of hair over trying to seal
everything against noise penetration
into the rooms. They were using the
rooms as recording studios, and when
you consider that ceiling wires transmit
noise from one floor level to the next,
you can see how tricky the job was.
“We isolated the ceiling wires with
brackets that mounted to the deck
above. The mounting brackets had a coil
spring that separated them from the
deck and a rubber damper between the
coil spring and the metal framing that
held the spring in place just above the
ceiling.
We used foam because the wallboard
could not be attached to the wall. We
hung it like a grid ceiling, but butted up
against a piece of rubber so that the
sound couldn’t transmit into the space
above the ceiling line.
The usual rating in decibels for sound
levels permitted to penetrate from one
room to another is 45.0, and we took it
all the way down to 1.0 for that job.
When inside the rooms, you could not
hear anything right outside, even if
someone were knocking on the door or
a window. We had installed special door-frames,
and triple-insulated doors that
weighed 450 pounds (with angled-cuff
hinges to make opening them almost
effortless). The windows, likewise, did
not transmit sound because they were
made of three layers of glass with air
between them.
“Part of the challenge, of course, was
finding qualified guys for acoustical. We
used someone from the office who had
several decades of experience in the field.
You lose your butt over these complex
ceilings projects, sometimes even paying
to do them. We try to learn from our
mistakes and hopefully make a little bit
on the next one. That church ceiling we
made nothing on, but we’ll know how
to price it better next time!”
THE CAVE CONNECTION
“The most challenging ceiling I have
ever worked on is the Founder’s Room
at the Disney Concert Hall, where the
high-rolling patrons gather prior to a
function for drinks.” So states Mark
Enquist, project manager at Raymond
Interior Systems, Orange, Calif. “Fundamentally,
it’s a suspended plaster ceiling,
but after that, all the rules changed.
The ceiling isn’t flat or horizontal; it goes
vertically and then inverts, or rolls over
onto itself. The overall effect resembles
the inside of a cave. It was designed by
Frank Gehry who some say crumpled a FROM CAVES TO COMPUTERS
piece of paper and said, ‘That’s it!’ I personally
believe in the other claim: that
he was inspired while caving.
“The ceiling is made up of 22 three-dimensional
“surfaces” that stretch from
hard edge to hard edge, undulating.
They intertwine and climb over each
other and do some fun things. They
have also been likened to the petals of a
partially opened rose as you look
upward, which then opens up at about
12 feet above the floor, undulating back
to the vertical walls at the perimeter of
the room. As you walk into the room,
the ceiling is flat for approximately 15
feet, and then starts undulating and rising
to a hard edge. Then it vaults up
another 30 feet above that. Total height
is about 43 feet.”
FROM CAVES TO COMPUTERS
So how does one design and build such
a complex ceiling? A computer is a good
start. On this job, a computer was used
from start to finish.
“The drawings were minimal” Enquist
explains, “just for permit purposes. The
true model of the ceiling was electronic,
and we built the entire ceiling from an
electronic file. Gehry was using the
CATIA format, common in companies
that use numerically controlled equipment
to manufacture models for auto
industry designs.
“So the first challenge was how to extract
the information without purchasing a
$30,000 program for this one job. We
use Mechanical Desktop extensively for
three-dimensional layouts of steel and
other components of buildings. So we
were able to convert Gehry’s files into a
format that Mechanical Desktop could
read.
“Originally, the ceiling was drawn conceptually,
using plywood gussets etc., cut
on a numerically controlled router-the
router tip is in essence the pen, and it
cuts lines in a big sheet of plywood
instead of drawing them on paper.
“We had a 1-inch tolerance within the
curves of the compound curved ceiling
and were suffering on how to create the
structure behind it and then end up
with our finished surface that was within
the tolerance.
“One concept that had been presented
to the owners before we came in was a
metal stud format, where they would
identify all the studs in the ceiling and
come up with a profile for each stud,
bend it and then tighten everything
together somehow. Discussion went
back and forth from lath and plaster to
drywall. But once we had our final
design, we said the ceiling had to be
plaster, and so it was.
“At that point, we studied the model
and investigated some of the methods
that people had proposed for manufacture.
Five of us in the office brain-stormed
and came up with the idea of
using tube steel, because we could roll-form
tube steel and it would be much
easier than crimping and bending studs
or stretch-forming studs.
“So, we settled on a design that incorporated
two-dimensional bending of pipes
welded into frames that, when they were
all assembled, created the foundation and
structure for this ceiling. We were able to
create two-dimensional pipe by taking
the electronic file and slicing it. Each slice
through the surface of the model then
created a two-dimensional line.
“Because we had limited access and door
height, we sliced the model horizontally
at 8-foot segments to create panels
that would fit through the doors. So we
assembled 4-by-8 panels of l-inch diameter
pipe in our shop and, with the assistance
of the computer, were able to
number them and coordinate the way
they fit into the structure of the steel, so
we could then put them all together in
the field.
“After slicing the building horizontally
in 8-foot increments of elevation, we
could define where each panel was going
to be created in the horizontal band, so
we could then extract each panel from
the model. We created a jig in our shop
that had two walls and a floor. The jig
represented a point in space so we were
able to measure x, y and z coordinates
within this jig. Because the jig mirrored
a file that we had set up electronically,
we could extract the information from
the model and orient it into our electronic
jig the way we wanted it built into
the physical jig. From there, we labeled
them surface 1, level A, panel #3, for
instance. That’s how we identified where
we were in the ceiling.
“We also created coordinates for every
corner of each panel in the computer
and labeled these coordinates in the
shop, providing the coordinate values in
a spreadsheet format. Our carpenters
then laid out an x-y grid on the floor of
the Founders’ Room and had a z-axis
from the floor up into space to obtain
elevations at 8-foot intervals. With the
x-y on the floor, they had a horizontal
laser and another laser shooting the z,
and where they crossed was exactly
where the corner would go. So, we
would have a laser shooting at 16 feet,
for instance, and then we’d lay out our
x/y coordinates all over the floor and
start raising them to the 16-foot mark.
“The panels had to be suspended some-how,
so we used cable. We created a sub-structure
from the existing large steel
members that had been bent to create
the arc of the vaulted ceiling. In between
those ribs, we welded tube steel horizontally.
We had pieces of stud coming
off the tube steel into the center of the
room to carry the lateral load of the panel,
transferring that load back to the
structural steel. Then the vertical load
was carried by cable—galvanized steel-braided
rope like aircraft cable.
“Using the computer again, we were
able to establish the profile and coordinate
plans for strong backs—steel bands
around the room every 4 feet in elevation
at 4-foot interval—to reinforce this
structure we had created in space. All we
had to do, then, was find the elevation
of a panel and where it was supposed to
orient in space, and attach the vertical
pipe in the frame of each panel to the
strong-backs with a U-bracket.”
BUILDING ITSELF
Enquist continues: “Because the ceiling
was three-dimensional and a compound
curve, once we had set the first row and
started working the second row, we
found that the panels almost set themselves.
If one corner hit the other corner,
the arc across that would match, and the
upper points always hit where they were
supposed to hit—plus or minus an inch,
and we’d push it or bend it, do whatever
was required. That was one of the
interesting results of the installation that
we were hoping for but didn’t expect.
We were expecting to have to measure
every single corner. As it was, we didn’t
have to; we were able to throw these
things up and screw them together. We
attached them to each other with plates
and straps.
“This success bore testament to our
three guys who had a very high priority
on accuracy with the computer model.
They took a couple of months to break
it down and create these plans. I am
looking right now at three 5-inch
binders full of cut tickets for each pipe.
We had to bend 2,600 pipes each to a
different profile and then assemble them
into three-dimensional frames. And
then all of those sheets were built down
into an architectural set that was about
20 sheets thick of various elevations.
“Because some of the corners were so
acute that the l-inch diameter pipe was
interfering, we used 1/4-inch pencil rods to create the corners. We had all
those corners tied off and everything attached to the structure we lathed
and then corner-beaded with a plastic-nosed radius down each edge, and
took out all the bumps and irregularities to create a nice, smooth ceiling.
After that, we scratched it and used a USG gypsum plaster finish.
“The computer was our primary tool. We broke the entire three-dimensional
ceiling into a two-dimensional drawing. To help my field understand
that drawing, I would take the computer out every once in a while
and spin the model around until we could see exactly where a hole, overlap
or offset might be. It was a first for me—using a computer as a tool to
define what we were building, to determine what studs we needed and
where they went and how they were configured in space. We had to know
before we built this ceiling that the cables were feeding correctly, because
we had very strict structural criteria. The cables were carrying a certain
amount of weight and we couldn’t overload the tube steel or the panel,
and we had to make sure we had just enough cables etc., etc. That’s why
we created a solid model and went to the extent of drawing every stud,
strut, cable and connection.
“For me, it was a sculpture. Our plasterers freehanded most of the surface
without using screens, so it really was sculpting for them; even though they
had a base form to follow, the final surface is all their work,” he says. Reading
from an anonymous newspaper article called Taking Pride, Enquist
continues: “‘With Disney Hall almost completed, [name] visited not long
ago, climbing scaffolding inside the still unfinished Founder’s Room. The
plasterwork on the curved ceiling is so fine that the work of the individual
plasterer’s hands is visible. [Name] is thrilled to see the beauty and complexity
of the ceiling, that human mark, like a brushstroke on canvas, a
project that he was hired to work on 14 years ago, so near to completion.”
Enquist says, “We had expected a certain amount of failure in the panels,
but, fortunately, we didn’t experience any that were so far out of dimension
that they were unusable.”
THE USUAL CHALLENGES
Raymond Interior Systems did run into challenges other than the technical
one of building an awe-inspiring ceiling: “We had shift issues as we
were running 24/7 for three months, but we were able to keep the information
flowing from shift to shift. The 24/7 made coordination among
the MEP (Mechanical, Electrical and Plumbing) trades pretty tense. We
would make progress at night and they would arrive in the morning and
say, ‘Oh no! What have you done!’ Tempers flared sometimes, but we
always managed to keep things rolling with communication. Everyone
working recognized the value and significance of the project.
“This was a hard job and with the right
amount of money, it could be done. But
we had to compress it down into an
approved budget after extremely competitive
bidding. I am not sure we came
out ahead, but we didn’t lose money. We
were hurt on the overtime, as the project
was compressed because HVAC design
lost about 1.5 months that cost us in the
long run. It was just extremely complex
to coordinate all these issues, because the
HVAC was running around behind our
structure and we had to make sure that
it fit within all of this strut work we were
installing. We had to have every single
strut cut just right-we couldn’t say,
‘Oh, just go ahead, just put it there.’
“We also had restrictions on weight,
because this project is built on a parking
garage; so we weren’t allowed to have lifts
closer than 5 feet to each other, and the
room was only large enough to fit three
lifts, adding difficulty. Coordinating
with the other trades wasn’t so difficult
after we became used to modeling in
three dimensions.
“We had to scaffold for the plastering,
and that was unique because the scaffolding
was not a simple straight up with
the undulating ceiling. Sometimes it was
overhead and sometimes underneath
our feet. So we created outriggers for the
scaffolding.
“The project ended up not being as difficult
as one might think looking at the
model: Once we had broken it down
into its measurable components, we
found everything just flowed. When the
model said there was supposed to be a
duct there, by golly there was! The wires
fed the way they were supposed to.”
The design team made the job easy, but
the folks in the field had “a lot left up to
them and they were very creative in
developing brackets and connection
methods, feeding and fishing things
through, and I was really pleased with
the way they executed the drawings,”
Enquist concludes.
Which goes to show, what lies above
that glass ceiling is where the real ceiling
challenges lie.
About the Author
Steven Ferry is a free-lance writer based
in Clearwater, Fla.
