In this temperature epoxy sets up fairly quickly, so yielding to my
impatience I did my peel test this morning.   The results were about
what I expected.   I peeled the fiberglass off each piece, and naturally
the piece that had been "flashed" had significantly more resistance, and
pulled more foam with it.   I got some sharper sand, and plan to do
another test, as the media I used was quite large, and very round, as it
was intended to reflect light when sprinkled on freshly pained highway
lines, rather than being an abrasive.    Glass beads are used for an
abrasive also of course, but they are extremely fine.... like dust.  
The abrasive I have now is a very sharp aggressive abrasive, but still
not typical blasting sand.
    What is very clear is that opening up the bubbles on the surface of
the foam allows somewhat more epoxy to penetrate in the irregularities
that are formed, giving it more grip.  It is also fairly clear that the
chemical bond between the epoxy and the polystyrene itself is not
especially strong.  But then many things I use epoxy for, I provide a
roughened surface for more grip.
    Peel strength is not really an issue in a sandwich construction. 
It should not be subjected to that kind of load, but rather to shear
load, and it should be quite strong in that kind of loading.   With any
kind of construction, joint design is important.  I do a lot of metal
fabrication, and of course have built many things in wood as well.   We
design our joints with a knowledge of what works and what does not work,
without even thinking about it with these familiar materials.   The same
is true of adhesives, and of composite construction.   For example,
yesterday I welded a piece of metal vertically to a surface, that was to
support a load.  The weld joint was small, and the load would be
concentrated at that point.  Years of building things.... I didn't have
to be told to install a small gusset.  Likewise, there are acceptable
methods for working with composite construction so you don't place loads
in ways that don't work. It's worth noting that foams like Divinicell
come with "surface features" for exactly this reason.

                                        H.W.

On 05/15/2018 09:24 AM, StoneTool wrote:
> Isn't there an old saying "crazed minds think alike", or something
> like that ;-)
>
>     The lower density of XPS allows for a thicker core for the same
> weight.   The fiberglass laminate on the surface provides the rigidity
> needed to protect the foam from permanent compressive deformation
> caused by sharp impacts.  T his means that the laminate needs to be
> thicker in impact areas, which means more weight, negating some or all
> of the advantage in these areas.... decks and such.   These areas
> normally would have a thicker laminate anyway for this reason, or an
> overlay of something like strips of teak, or ersatz teak, etc, so it's
> probably a non-issue.
>
>     I've had fantasies about laminating two foams together to make a
> progressive core also.   The optimal load bearing sandwich would be
> sort of like the optimal beam structure, which would progress from a
> very high strength outer (fiberglass) that actually carries the
> tension and compression loads, to lighter and lighter structure as one
> approaches the neutral axis, then progressively stronger as one
> approaches the opposite surface.    Charts of beam loading look like
> an hourglass. Obviously this is not practical in the real world.  In a
> sandwich structure we have a load applied that is carried over a wide
> area.  In extreme situations, such as testing to failure, the failure
> mode should be the glass shearing away from the foam, not failure of
> the glue line, but of the foam itself, and the two surfaces moving in
> opposite directions (relatively). In real world use, we simply are not
> going to see this, short of a catastrophic incident such as hitting a
> reef at speed, or a collision between two boats, etc.    In that case
> any structure is likely to fail, and repairability becomes the issue,
> after the ability to make it into port intact.
>
>     In the minimal experimenting I have so far done with XPS, I've
> found that a "flash pass" with a sand blaster using fine media will
> open the surface layer of bubbles, much like proper hot wiring does,
> and is a matter of literally seconds of exposure.   This should
> address any issues of peel strength. Any more than a quick "flash
> pass" will distort the surface.   I don't own a shop sand blaster or
> glass bead cabinet, and have been playing with an ordinary
> undercoating gun.   The only media I have at home is extremely
> course.   The reflective beads used on highway centerlines...... I've
> got a 3 pound coffee can full of it from painting the sides of my
> flatbed trailer.    They are very effective, but do not cut like
> regular media, but I want to do a try with a real sand blasting
> media.  I know where I can grab some.      The undercoating gun is a
> very cheap item, readily available anywhere.  Mine was purchased for
> doing bed liner, which incidentally I suspect would probably make an
> excellent, rugged, grippy overlay for decks.  I wasn't using it on a
> deck, but the stuff I used was extremely impressive.  It was an epoxy
> / urethane hybrid by a company in Texas, called Rattleguard.  No
> longer in business.    Incredibly tenacious and rugged.        I have
> two small pieces of XPS foam with epoxy / glass on them I laid up this
> morning.   I'm not sure the weight of the woven fabric..... 7 oz?   One
> was flashed with media first, the other not.   The epoxy was the only
> epoxy I have in any quantity, a very high quality structural adhesive
> by System Three normally used for wood aircraft and boat building,
> T88. Not a laminating resin, but I see no reason not to use it for
> this.   I have about half a gallon of it on hand.
>
>     Stay tuned for results........ I'll probably let it sit for a day
> or two before testing peel strength, etc.
>
>     H.W.