Historic Preservation - Technical Procedures
- Preservation Tech Notes: Windows 12 Aluminum Replacements For Steel Industrial Sash
- Procedure code:
- Preservation Tech Notes, National Park Service, Pad
- Doors And Windows
- Aluminum Windows
- Last Modified:
- Preservation Tech Notes: Windows 12 Aluminum Replacements For Steel Industrial Sash
- Last Modified:
PRESERVATION TECH NOTES: WINDOWS Number 12
ALUMINUM REPLACEMENTS FOR STEEL INDUSTRIAL SASH
This standard includes the bulk of information contained in the
original Preservation Tech Notes developed by the National Park
Service and the Center for Architectural Conservation at Georgia
Tech. The Preservation Tech Notes are case studies of exemplary
projects designed to provide specific examples of sound
preservation techniques. To obtain a complete copy of The Window
publications including figures and illustrations, please contact:
Historic Preservation Education Foundation
P.O. Box 77160
Washington, DC 20013-7160
The Window Handbook, jointly prepared by the National Park Service,
Preservation Assistance Division and the Center for Architectural
Conservation at Georgia Tech, also contains all of the Tech Notes
on Windows and is available for purchase from the Historic
Preservation Education Foundation for $32.00. The Window Workbook
is available for $49.00. The two publications together can be
purchased for $72.00.
(CHARLESTOWN NAVY YARD)
Building 149 is a 10-story 700,000 square foot reinforced concrete
structure built during 1917-1919 for use as a naval warehouse and
offices. It is located in the National Historic Landmark Boston
Navy Yard, which was established in 1800 and which comprises
approximately 130 acres and nearly 90 buildings associated with the
naval shipyard operations. Portions of the installation now owned
by the Boston Redevelopment Authority consist of sheltered
shipways, warehouses, offices and residences. Vacant since the
decommissioning of the shipyard in 1974, Building 149 recently has
been renovated for use as offices and retail space by a private
development firm under a long-term lease.
The building's fenestration - nearly 2000 steel window units set
within 500 openings - was considered a very distinctive feature of
the building. Through careful planning and attention to detail, an
innovative aluminum replacement window system was developed by the
project team that successfully maintained most of the
distinguishing features of the original windows.
The inside-glazed, historic green-painted windows had narrow 7/8"
wide muntins with an exterior cove bead shape profile to the
muntin. Most of the openings consisted of a bank of 4 side-by-side
window units. Each of the middle two units consisted of 20 divided
lights, including a 6-light center hopper; the two end units were
fixed and contained only 16 lights. Typical of pre-World War II
steel windows, the glass panes had a narrower width than height.
The vertical mullion connecting each unit was approximately 3"
wide, noticeably dividing each opening into 4 window units.
The contractor's survey of the historic windows in the spring of
1984 revealed that extensive rusting of the frames had occurred and
that many were racked. The severe rusting had also contributed to
the spalling of sections of the concrete sills, jambs, and
spandrels. Repair and upgrading options to maintain the historic
windows were quite limited due to the size of the glazing bars.
The shallow depth of the metal glazing bars (muntins) seemingly
precluded the installation of sealed insulating glass within the
existing lights, even if the windows could structurally support the
additional weight. The only practical way of double-glazing would
have involved the use of interior storms with units that were
either operable or were removable for ease of cleaning. Even then,
however, the severe deterioration of the steel windows still would
have needed to be addressed. Considering the size of the bay
openings, the decision was made to replace the windows.
Four replacement options were considered:
1. Replacement with matching steel units in combination with an
operable interior storm window system.
2. Installation of large sheets of insulating glass, maintaining
the principal 4-part division of each bay while eliminating
the small multi-light pattern which existed.
3. Installation of large insulating glass units, maintaining the
principal 4-part division, and applying an exterior aluminum
grid in an attempt to recapture the appearance of the historic
multi-light steel windows.
4. Development of an aluminum window system with true divided
lights with insulating glass, maintaining as close as possible
the profiles of the historic glazing bars and overall historic
The use of steel replacement windows was considered only briefly
because a double-glazed system in such large openings would be high
in cost, and would not be able to retain the narrow sight lines and
profile of this particular type of steel window. The existing
profile was available in a replacement steel window but could
accommodate only single glazing, which was not considered adequate
by the developer for energy purposes. Thus an interior storm
window would have been necessary; however, the large size of the
window openings would have required an expensive commercial storm
A mock-up of the second alternative was installed, consisting of a
fixed aluminum window with large sheets of insulating glass. Each
opening had three vertical mullions, dividing the opening into 4
parts; this matched the principal division of the historic windows.
Since the glass was not divided further into smaller lights, there
was a dramatic change in the appearance of the building and this
alternative was quickly dismissed.
The third alternative, however, was seriously considered since it
provided for an addition of an exterior aluminum grid applied to
the face of the fixed aluminum window described in the second
alternative. The grid was intended to simulate the appearance of
the historic windows. The extruded aluminum grid would duplicate
the cove-bead profile of the exterior portion of the historic
glazing bars and would be attached directly to the glass using a
special epoxy glazing tape. This system had been used recently by
at least one developer on a similar project. The estimated
fabrication and installation cost of this window solution was $1.1
million for the 500 openings.
The project director, Richard Graf of The Congress Group, Inc.
(developer), and the Boston Redevelopment Authority (holders of the
ground lease) both had reservations concerning the long-term
performance of the exterior aluminum grid. In the late 1970s there
had been a number of projects where wooden muntin grids had been
glued directly to the glass and where subsequent failure had
occurred. Besides the question of the performance of glued-on
aluminum grids, there were some visual changes that would result
from the exterior applied grid compared to the original glazing
bars. The Boston Redevelopment Authority was also concerned over
the growing use of false muntins in the rehabilitation of large
industrial buildings within the historic navy yard and the negative
impact it was having on the overall architectural character of the
These collective concerns and the need for rapid approval of the
rehabilitation plans led to the decision by the developer in May,
1985 to choose a fourth alternative: an entirely new aluminum
True divided lights with insulating glass would be used as part of
the new system with muntin profiles and framing members that
closely matched the historic design.
The project architect and construction manager were responsible for
preparing preliminary design guidelines for the new window system.
Two local window contractors submitted bid proposals. One company
proposed that the glass be exterior-glazed using integral muntins
that were close to 1-1/4" in width. The other company showed an
interior-glazed window and claimed that the integral muntin could
be made as narrow as 1-1/16". Since inside glazing would
facilitate both installation and maintenance, the decision was made
to work with this company in the design of the windows to be used
in Building 149. The contractor's bid for this window system was
$1.4 million, which was approximately $300,000 more than the
Further development of the window system was required and the
window needed to be performance tested - all requiring fast track
scheduling. A development and construction team for the window
work was assembled consisting of the following parties: the
developer, the project architect, the window contractor, the window
fabricator in Denver working with the window contractor, a testing
laboratory in Boston that would assist with the performance needs
and design of the window, the general contractor, a preservation
consultant and an independent testing laboratory in Dallas
responsible for final testing.
The engineering and design of the new window systems required close
and frequent coordination between the various team members because
of the number of important issues which needed to be resolved, all
within a very short time frame.
One of the first major design issues to be resolved concerned the
need to match as closely as possible the shape and dimensions of
the original 7/8" wide glazing bar (muntin) with its decorative
cove-bead exterior profile to simulate the profile of the original
steel window muntins. The project team concluded that in order to
keep the muntin on the aluminum window as narrow as possible, the
traditional cast thermal break (cast plastic) feature of most
modern windows could not be used. Instead, a series of spacers and
gaskets principally would be utilized to achieve a thermal break
for energy conservation. By using this approach, the window
fabricator would be able to use a 1-1/16" wide integral cove-bead
muntin. The only short circuit in the thermal break would be at
the point where screws were used to connect the inner and outer
portions of the muntin.
Besides the final detailing of the nonconventional thermal break,
the representative from the local testing laboratory, was
particularly concerned about water infiltration. A system was
designed to ensure that moisture buildup behind the glazing tape
would seep outside, rather than inside the building. A twenty-foot
mockup was eventually constructed and successfully tested according
to accepted industry standards.
A third important design consideration centered on how to keep the
framing members and muntins narrow enough to maintain the thin
profiles of the steel windows. The need for a thermal break in
addition to the use of aluminum, which is structurally weaker than
steel, necessitated some increases in sections and profiles. A
technique more commonly found in skylight construction was used to
hold the glass in place. This consisted of screwing members
together rather than using snap-on aluminum sections to secure the
glass. Snap-on sections would have required more metal and wider
A fourth design and engineering issue arose with the construction
detailing of the muntin joints. The decision was reached to face
glue the joint on the front and spot weld behind. The fifth issue
concerned the visual impact of the spacer used in the insulating
glass. The original plans called for an aluminum spacer that
turned out to be too shallow in width to properly glaze the sealed
insulating unit. Since the acceptable width required a slight
encroachment beyond the edge of the muntin, there was a concern
over the potential visual impact. By selecting a bronze spacer,
the metallic reflection that would have occurred from the typical
aluminum mill finish was avoided and the visible portion of the
dark bronze spacer was not noticeable from the street below.
The sixth design issue, which ultimately was not resolved,
concerned operability of the windows for ventilation. While there
were some advantages to having operable windows, they were not
paramount considering the building's new use as offices. With
aluminum frames, a 6-light hopper or projecting section as existed
in the historic windows was not considered practical at that time.
The primary reason was the need to keep the aluminum sections as
narrow as possible to match that of the original steel. Given the
structural requirements of an aluminum window, it was considered
possible to fabricate only smaller operable units (1-3 lights).
With the tight construction schedule,, the additional development
time that would be required, and the higher construction costs, the
decision was made to proceed with a fixed window. This meant that
there would be a noticeable change in one feature of the historic
windows as a result of deleting the hopper section in the middle of
two window units. The overall appearance of the new window and the
building itself was judged to be sufficiently close to that of the
historic appearance, however, that a marked change in character
would not result.
The seventh and last major design decision concerned the number of
pane divisions to be provided in each of the four sections of the
window openings. The relationship of solids to voids (frame to
glass) was important to retain. Since the muntins were to be
increased in width from 7/8" to 1-1/16", discussions arose
concerning possibly reducing the number of lights. Besides cost
savings, changing the number of lights would help solve another
problem stemming from plans to lower the sills due to the high sill
height within the building. The light pattern that was developed
while reducing the number of lights, maintained the vertical
orientation of the glass panes, and the proportion of solids to
voids, further reducing any visual impact of the slightly wider
The basic aluminum window unit consisted of 9 different aluminum
extrusions, including the decorative cove-bead muntin. The muntin
assembly actually consisted of 3 extruded aluminum sections. The
principal muntin section was the cove bead portion that had a long
glazing channel with a receptor at the end. Attached to the
interior-facing side of the muntin was a U-shaped glazing stop
secured by self-tapping screws to the receptor on the cove-bead
section. This stop secured the glass in place. For aesthetic
purposes, the stop had a snap-on cover to hide the screws and
create clean lines on the interior. Through the use of neoprene
gaskets and plastic and neoprene spacers, a thermal break was
achieved, broken only by the screws.
While the horizontal muntins were continuous across the window
unit, the vertical muntins had mitre-joints where they intersected
the horizontal muntins. The vertical muntins were secured through
a combined use of epoxy glue and spot welding. A system of weep
holes and channels was provided to ensure that any water trapped
between the glazing tape and the glass muntins would be diverted to
the outside of the windows.
The overall window unit was not set into reglets as were the
original steel windows but rather were bolted to the masonry
because of the greater depth of the aluminum jambs. To keep the
width of the frames sufficiently narrow to match the historic
appearance, a 3/4" wide jamb was designed, narrower than standard
window jambs. Due to high wind loading requirements for Boston,
steel reinforcing bars were needed at certain corner windows, but
otherwise, the aluminum window system was designed and successfully
tested with the narrow jambs.
***WINDOW FABRICATION AND DELIVERY***
Through weekly meetings among the window project team, it was
possible to provide for a rather complex manufacturing process for
the overall windows that yielded cost savings and also met a very
tight production schedule.
The 9 extrusions required for the aluminum windows were
manufactured in Portland, Oregon, and painted the historic green
color in Salt Lake City; both companies had worked before with the
fabrication plant. Fabrication took place in a window plant in
Denver that previously had done work for the Boston-based window
contractor. The fabrication work was complicated by the fact that
there were a number of size variations for each of the 9 different
types of windows in the building, although approximately 500 of the
2000 window units were the same size. The greatest variation
occurred in the height rather than the width of the windows. A
maximum of 3/4" tolerance was allowed around the sides of the
overall window units in each opening; such tolerance was necessary
because many of the openings were skewed.
While the windows were being manufactured, the tempered glass,
required by the Fire Department, was cut in a plant in Tennessee
and shipped to Easton, Massachusetts, where the glass was made into
insulating units. The window contractor helped to coordinate all
this work and was responsible for ensuring that the glass was
properly sized and that the spacers in the insulating glass did not
encroach more onto the visible glass area than was specified. A
number of the units had to be sent back to the glass assembly shop
in Easton due to inaccurate sizing or misalignment of the bronze
spacer. This work involved the greatest problem and biggest
expense, since the limited tolerance for encroachment onto the
glass area required very careful work.
***INSTALLATION AND SCHEDULING***
While the windows were being assembled, the existing openings were
being prepared. The work included the installation of all new cast
concrete sills due to the lowering of the sill height. The windows
were shipped to the site and installed unglazed.
The scheduling of the work reflected the fast track of the project
as a whole. The decision to go with true muntins was made in May
1985; by June the general design of the window had been made and by
July the final extrusion drawings were approved by the architect
and consultant. By mid-August, the extrusion work was underway in
Oregon and in September, the final testing by an independent
laboratory in Dallas, Texas, was complete and the go-ahead for
production was given. Fabrication started in September and the
last of the windows were shipped from Denver in late December 1985.
Installation of the windows began in January 1986 and the final
glazing was complete by June 1986, well in time to coordinate with
the scheduled completion date.
The local window contractor was responsible for coordinating the
extrusion and painting work, the window assembly, glass
manufacturing and installation. Vital to the success of such
complicated work was the close coordination and series of weekly
meetings between the architect, developer, facade consultant,
construction manager, and window contractor. During installation
the facade contractor - responsible for the rest of the exterior
work - was also a participant.
The total cost of the window work was $1.4 million. It was hard to
estimate the total development cost of the new window system,
although design and testing cost somewhat in excess of $50,000.
Despite the special work required and the complexity of the
development and manufacturing work, the window system was only
$300,000 more expensive than the grid system initially proposed and
subsequently abandoned due to performance and aesthetics
considerations. The resulting windows cost approximately $25 per
square foot installed. Except for several changes at the building
expansion joint, there were no cost overruns due to the window
design. The window contractor, however, absorbed some unforeseen
labor costs in this initial project.
The window work at 149 Constitution Park was noteworthy in several
ways. First, it represented a significant improvement over past
attempts to recapture the distinctive qualities of a steel
industrial window with narrow cove-bead glazing bars, using an
aluminum replacement system with insulating glass. Equally
important was the manner in which the new window system was
developed for the project.
The risks that were inherent in developing a totally new window
system for a large rehabilitation were minimized by the team of
highly qualified people; who coordinated closely and who kept to a
tight schedule. The additional costs incurred in the development
of the new window was not excessive considering the massive size of
the project; the manufacturing and installation of the new windows
with true divided lights did, however, appreciably increase the
cost of the window work. The results, however, are quite
impressive and this innovative window system is commercially
available for use in other projects.
This project shows just one way that significant improvements can
be made on the quality of aluminum replacement windows used in
historic buildings. The planning team involved in this project
also identified further improvements that might be possible with
this particular window system. While the new windows lack the
hopper detail and altered the size and number of the muntins, many
of the characteristics of the large steel industrial windows have
The project team were concerned not just with appearance but also
with quality, engineering and high performance. This is important
since poorly built windows, whether old or new, can lead to
excessive maintenance and high energy costs. The assembled team
brought together the different professions and perspectives needed
to produce an energy-efficient, cost-effective, and aesthetically
While the window work was on a fast track from planning to
completion, the decision to address the window issues early in the
overall planning of the project provided the necessary lead time.
Too often, window issues are addressed late in the planning of a
project, providing little time to fully explore available treatment
options. Where an innovative solution is necessary, as with 149
Constitution Park, extensive planning is crucial to the successful
execution of the work.
- 149 Constitution Park
Charlestown Navy Yard
- The Congress Group, Inc.
PROJECT DATES: 1985-86
- Richard Graf
The Congress Group, Inc.
- Amir Man
Huygens and DiMella
CONSULTING TESTING LABORATORY:
- Thompson and Lichtner
- The Dallas Laboratories
- William MacRostie
Heritage Consulting Group
- Custom Windows
- L. Rubin Glass and Aluminum, Inc.
- The total construction cost of the window work was $1.4
million or $25 per square foot of window. There were
additional development costs for the design and testing of the
window which were approximately $50,000.
END OF SECTION