Welcome to the

White Paper: Window Considerations

About:

Contributor:	Gary Boucher

Last update:	11/17/97

Synopsis:	This page contains a discusion on the topic of widows or
		view ports for a submersible, how they are made, tested
		and cared for.

Disclaimer:	You are resposible for your own safety. Even though
		this information may work or seem resonable in some 
		cases you need to approach the subject carefully and 
		even aquire the help of a Marine Architect before applying
		into one of your designs.

Body:

WINDOW CONSIDERATIONS

    One of the most critical areas of submarine design is window
fabrication.  Windows can be one of the weakest links in the system.  If a
window fails at depth it can easily be fatal to the occupant as the water
can fill the sub in a couple of seconds or less causing the pressure to
increase far faster than the human system can tolerate.  For this reason
safety dictates that window design be thought out most carefully.  
    When I first started thinking about windows, my first uneducated
thought was to use tempered glass such as what industry uses for sight
gages.  I have no idea how well this would have worked in the long run and
I do not recommend using any type of unproven material.
    I contacted a local plastics company in Shreveport and asked about
super thick acrylic sheets that I could cut a window from.  I was thinking
of plastics that were maybe in the order of two inches thick.  The salesman
told me about a man that lived in San Diego California that was a world
class authorities on submersible windows.  I got his name and found his
home number and called him.  Jerry D. Stachiw worked, and possibly still
works for the Naval Underwater Research Center in San Diego.  Stachiw told
me that he would answer any questions that I had that night but he wanted
me to buy his book.  I said that I would buy it and he proceeded to answer
some simple questions that I had.
    I ordered a copy of "Acrylic Plastic Viewports", from Marcel Dekker,
Inc.  Their address is as follows:

     Marcel Dekker, Inc.
     270 Madison Avenue
     New York, NY  10016 

This has to be one of the most complete studies of the acrylic viewport
ever published.  It is worth the money even though it cost $100.  I was
amazed at the complexity of considerations that go into window design and
at the same time the simplicity of use of the final design information that
the book provides.
    In my discussion with Stachiw I brought up the subject of a rectangular
window.  I had already constructed the seat for the window as my primary
forward viewport.  He said that he could not recommend it because there is
no data on rectangular windows.  He did not discourage me from the pursuit
of the rectangular design but he said that the only way to be sure that
such a design would work would be to build a test jig and pressure the
window through many cycles as they had done in the labs where he worked. 
Then I would have had to destructive test the window for maximum depth etc.
 This was too much extra labor and expense for me so I opted for a round
flange in a square hole fix and put a large flat plate window in instead.
    The one thing that amazed me in regard to safety was the detail to
which they tested a design, making what seemed to be no assumptions as to
it's reliability.  This was his life's work and he has a Doctorate in
mechanical engineering and knows his stuff.      As part of the testing
procedure his team will pressure test a given geometry to the breaking
point over and over.  Destroying many windows of a given geometry until an
average failure point is reached.  This is then repeated with a slightly
different thickness or diameter.  Curves of predicted failure are
established for a given type of window.  The user finds on this curve the
thickness desired, the diameter, and also the unsupported flange diameter
and enters the chart to find the failure depth.
 For my sub I actually used 1.25 inch thick windows but using the curves
I assumed it to be 1.00 inches thick.  I took all the other information and
located the estimated failure point which for my sub was about 1600 feet. 
He then tells you to derate the window by a factor.  I used the factor "8"
from a table of factors provided.  This allows me to go to a 200 foot
depth.  He states that the factors used take into consideration arctic
conditions and normal wear and tear on the windows.  He also places a 10
year maximum life after which the window should be replaced.
    I also learned that you can not just cut a piece of plexiglass and
insert it into the window mount etc.  Acrylic plastic is cast into sheets
at the factory under heat and pressure.  This is the way that acrylic for
windows is sold.  If the designer simply cuts the plastic into the desired
shape there are stresses set up at curing time that will cause crazing and
possible failure of the window.  For this reason the plastic must be heat
treated.
    Two cycles of heat treating are required.  First comes the shrink
cycle.  If you cut a piece of plastic for a window you must cut it maybe 10
percent larger than you want to use.  During the shrink cycle the diameter
of the rough pieces will be reduced and the actual thickness will be
increased by as much as 4 or more percent.  This is the reason that the
initial cut should be larger than the resulting window to be machined.
    This heating for the shrink cycle has to be slow.  It takes several
hours to take the window to a temperature around 300 degrees F.  It is held
there for several hours and then cooled for several more hours.  It takes
about 10 hours to run one window.  I built a computer controlled oven to do
this.
   After the window is shrunk it is machined into the shape desired.  But
the machining causes more problems.  Stresses are set up that need to be
relieved.  So another anealing cycle is in order.  This time the
temperature can be somewhat cooler but still must be slow in rise and fall.
 When this is done no more machining is allowed.  The window is polished
before this last heat cycle.  In all, this is a lengthy process with
different choices of temperatures and cycle length.
    There are several shapes that can be used for windows.  Spherical
windows are used but they are not easy to make.  Conical sections
(Frustrums) are used for very deep diving subs.  Machining here is a
premium.  Not only does the window have to be machined in a conical section
but also machining is required for the seat where the window sits.  The
easiest is the flat plate window.  These can be used to moderate depth
usually deeper than most personal subs are rated to dive.  The above
mentioned book shows all the types and gives what I would think is all
necessary information and more.
    From reading Stachiw's work it becomes obvious that window design is
number one critical, and number two something that must be researched in
the most technical ways.  What this means to me is if he is skeptical about
a rectangular window, what would a true professional think about some of
the off-the-wall designs being used on some of the subs today.  You can not
just take a piece of acrylic plastic and heat-form it to fit your sub and
trim it with a saw and grinder and have any true confidence in what it will
do on a certain dive or after fatigue factors set in.  It is most dangerous
to assume that because a window looks great that it will perform at
desired.  I personally think more of me than to subject myself to many
unproven window designs that people dream up.
    I recommend that every designer stick with proven designs where failure
can be predicted and known limitations can be set.
I have seen Lexan used for windows.  It seems like it would be a great
material as in general it is tougher than acrylic.  However, once again
unless there is proven data on Lexan I plan to use acrylic.