Lexan window cracks - help needed

[Wynand N]

I experienced some mayor cracking installing windows on a steel boat under construction and attached photos to show the damages.

But let me start at the beginning. We needed a smoked colour glass to blend in with the colour of all hatches, portlights etc fitted, Since it is side cabin windows and may take a large wave or more when sailing, it was decided to use a Poly Carbonate plastic such as Lexan. Problem being, Lexan only available in 4mm thickness in smoke tinted colour in South Africa - our national dealer and importer in Maiseys plastics, largest in country - and this is to thin for the application and would fit "badly" in its allotted frames.

It was then decided to bond the 4mm Lexan to 8mm clear Perspex to get the desired thickness and tint. The Lexan was placed on the outside of the boat. This was actually a good idea then, the bond between the Perspex and Lexan can act as a trim to hide the actual bond of the window to the frame from the outside - much like the black trim around an automobile windscreen. Secondly, the bond will also keep water finding its way between the two laminates. See first two pictures.

To make this bond I contacted my supplier and he got SikaFlex's technical department phoned me and I explained what we want and he made his recommendations of what to use and we did exactly that. The edges to be bonded has to be treated with a Sika Primer 209N primer and then thin layer of SikaFlex 295-UV applied. This we done straight out of the book.
All holes were drilled 2mm bigger than the bolt diameter and countersunk with countersunk tool for perfect match with bolt c/s head.

Before fitting the window to the frame, a general layer of SikaFlex 295 was applied on the frame and primer again to the Perspex to be mounted against it. Bolts were inserted and only lightly tightened until the window made solid contact with Sikaflex and left overnight for the SikaFlex to cure and form a gasket before tightening things up.
At this stage nothing was amiss and I have to mention that the max curve on these windows are about 6mm over a length of an average 700mm. Some are actually flat. We only fitted the starboard side windows.

Yesterday morning hell came down on me - climbing into the boat to inspect our work, all the windows are cracked badly and I thought it was the Perspex. On closer inspection I found it was the virtually indestructible Lexan that cracked!!.
I immediately went to the plastic supplier and they came in force with their technical guy to check things out. They found the bolts still loose and can be easily turned with allen -key (nylock nuts). The slight curvature they found not to be an issue and suspect that the holes were to small. When informed and shown on windows still to be installed the clearance is about 2.5mm per hole that was eliminated. The suspected the primer and came back later with 18 pages of specs from SikaFlex head office and again it was found we done everything out of the book and used the right stuff. They were flabbergasted just as we were.

Then someone in the know told me of similar problems they had experiences and said these plastics has to be drilled with a blunt drill bit, actually be burned through. We took the smallest window to be still fitted on the port side and slightly enlarged the holes further with a blunt bit (burned through) and we fitted that. And just to make sure, although the windows fit loosely into its bracket, we took off another 2mm off the perimeter of the window. Guess what we saw this morning....same thing happened again.

Can anyone help me or give some advice? This is a serious and expensive disaster and do not want this to repeat itself. Funny, in the past we fitted Perspex only windows to hulls with a lot of curvature and never had anything remotely as this happens.

Finally, we bonded pieces of Lexan and Perspex offcuts and drilled exactly as per windows and put the bolts in and tighten with such force that we strip the allen-key holding the bolt without any crack to the Lexan....

[waikikin]

Countersunk bolts, I think maybe that perfect match to the countersink may be to blame plus the thermal coefficient of expansion steel to plastic? maybe use a screw cap- like a cradle washer for the countersunk head or try some round head bolts with flat washers. Regards & best of luck with it from Jeff

[Ad Hoc]

Firstly...why bolt the windows if you're gluing them as well?..one or the other, not both.

If you have any of the lexan left, take a bit - a half decent size, uncut. Place it on a table, support it at the ends, with small supports, like a simply supported beam.
Then over night, lay a weight, on top of cloth, on to the lexan. Say 2kg..nothing seriously heavy, but enough to slightly stress it.

Look at the result next morning..if all ok, fine

Next repeat, but this time, scratch the surface with an X mark, roughly in the centre and leave over night...if all ok, fine.

Finally, where you have placed the X mark...drill, as you have before, and repeat.

The purpose of this is to ascertain if the lexan has been made prestressed. Just like an ally plate that has locked in stresses when welded, if high restrained.

It should fail under any of the above, if the manufacturing is suspect.

If it does not fail...hmmmmm..interesting!

[Manie B]

Wynand i know this is going against "normal" thinking

but we have had so much sh1t with lexan at the dam - also mostly flat, that simply never again - if its not the one thing - its flippen glue or silicon migration and and and - the list is endless - you farkin didn't do this right - then you didn't do that - so fu3k that sh1t NEVER AGAIN
and after doing it to the book, 3 years later and its got to be replaced because guess what - you flippen didn't do it right - the sh1t doesn't last either.

Now we are going to lamited auto glass - many specs available - and to get the "tint" on the inside we use "film" as what is used on cars for "smash and grab" or to darken the windows, once again MANY types available

Down side is to get curves, it is very costly to make up so basically we are going flat, and glass dont scratch, it lasts forever and you can even have it so thick that it is bullet proof for them pirates :D

And i can assure you that an extra 10 maybe 20 kgs makes no difference to a 43 Steel Dix displacing big tons

sorry about my ranting but it drives me up the wall when the "experts" wanna bullshit me about what works, when the flippen "expert" cant operate a jigsaw :mad:

[Landlubber]

Could it simply be the different expansion rates of the two [lastics, lexan really does grow. It has to be remembered when fitting particularly long wondows in cats for instance. Ask the supplier about the expansion rate, it can be 10mm per 1000!

[daiquiri]

Perspex and Lexan have both quite similar coefficients of thermal expansion, around 7*10-5 m/m/°C.
The strange thing is that some of cracks don't seem to originate from the bores. Could be the glue shrinking/expanding and pulling plates while hardening? Just guessing here.
In any case, maybe the best and most economical thing would be to search for a supplier of a single-layer thick lexan or perspex plate outside of South Africa.

[Luckless]

What was your drilling method? My first thought would be there is a difference in thermal expansion rates of the two materials that is causing breaks along stress points introduced by drilling.

Random thought for if you're going to have to throw the cracked windows out anyway. Use a fine marker to trace the cracks on the windows as they are now, then stick one of the windows in a freezer for a few hours to see if the cracks expand any more. If they expand on the chilled one and not the warm ones then I would chalk the issue up to thermal expansion for sure, if not, then still no idea what the issue is.

You said that Lexan only available in 4mm thickness in smoke tinted colour in South Africa, do you mean that the smoke tinted colour you want is only sold in 4mm sheet? Would it be possible to get full thickness clear sheet, and then apply a tinting film sheet to both sides of the Lexan after? This would at least give you a bit of replaceable scratch protection on your windows.

[Wynand N]

Thanks for you guys thus far. Let me answer some questions raised.

Lexan is only available in 4mm sheet in my part of the woods -regardless of colour.
The different coefficients of thermal expansion between Lexan and Perspex, if any, should not have been a problem as the windows were bonded about 2 weeks ago and no shit happened until fitted to the hull.
Furthermore, the boat sits in a factory where the difference in temperature is minimal and temperatures are still in the mid to high twenties C.

When installed, no visible cracks happened or any sounds heard and the bolts were not tensioned yet - just pulled the window against the SikaFlex and left to cure overnight before tightening. One window has a crack from side to side....

The problem is how to get the windows out again with that Sika cured:?:

[mydauphin]

Getting windows out is easy. I use a serrated wire in between window and frame. Cut it out. You can also use other wire, Also contact Sika they may have softerner.

[Wynand N]

Ad Hoc, your reply makes sense and perhaps the Lexan was not made pre-stressed.
We are setting up the test you described and will see how it goes - the Easter weekend will be a help as we would only be able to check the bend test by Tuesday morning. I all is OK we will continue as prescribed.

However, we are fitting a clear Perspex window - counter sunk with the bolts and primer applied, really tighten up without the Sikaflex and see how that holds up by Tuesday. If OK, we will probably go the thicker Perspex way with a plastic window tint applied - but my client will have to give the OK for that decision.

Why glue and bolt:?: This boat will not be a marina queen and she will spend her time on the big blue. My client is a marine engineer and like things a bit over engineered just to be on the safe side of the parting line...

[waikikin]

Wynand, you can try bolting your windows without the counterunk bolts, the clearance you give the bolt holes is meaningless if the head of the countersunk bolt interfaces the countersink in the plastic, I was taught never to use countersunk bolts/screws with perspex or install them "loose" & as a back up only to the adhesive/sealant, also I never install bolts to the "middle" of the window corner radius but scale the fixings to each side of the pane to even as possible spacings inclusive of the radiused corners. All the best from Jeff.

[Landlubber]

Jeff is very right mate, don't countersink the perspex for the very same reasons you made the holes clearence holes.....asking for trouble

[jehardiman]

Wynand;

Just a question, is there now or recently been any painting or solvent use in the building shed?

I also concur with not using a countersink if you are going to through bolt it, I have used cup washers in the past. If the curve is slight then you can thermoform it a little to get it to lay better.

Here is an old thread on what I have done before.

http://www.boatdesign.net/forums/materials/how-install-acrylic-windows-reliably-15042.html

[rugludallur]

Hey Wynand,

Just thought I would give you a heads up, some stressed acrylics don't react very well to solvents, in particular alcohol.

Here are some videos to illustrate
http://www.youtube.com/watch?v=cU9Ty0L0g7E
http://www.youtube.com/watch?v=SUgNm5VVQMU

Jarl
http://dallur.com

[Kingo in oz]

Wynand,

Just another take on your problem - i actually think it is a combo of the countersunk heads, oversize holes, semi-tightening, and the curing of the Sika, but mostly the latter two.

The Sika will cure at different rates depending on the thickness and the surface area exposed to the atmosphere. By only tightening the bolts until the plastic makes contact with the adhesive and then leaving the assembly to cure, this is allowing the varying forces of the curing Sika to push and pull the plastic as it pleases. So you may have one section of the join curing first and this pulls the plastic sheet in its direction, then as another section cures it pulls the sheet back in that direction. The effect of these forces may be exacerbated by the countersunk heads which could be forcing the sheet up and down (in and out).

My chemistry knowledge of adhesives, sealants in general, makes me think that it would be better to use a very thin bead of Sika and tighten the bolts fully at the beginning. The Sika will take much longer to cure but i doubt this would be an issue for you.
It probably would also help if the holes weren't countersunk and you used pan-head bolts or similar. (as others have already mentioned)

Anyway, good luck with it!

[peter radclyffe]

Wynand,

Just another take on your problem - i actually think it is a combo of the countersunk heads, oversize holes, semi-tightening, and the curing of the Sika, but mostly the latter two.

The Sika will cure at different rates depending on the thickness and the surface area exposed to the atmosphere. By only tightening the bolts until the plastic makes contact with the adhesive and then leaving the assembly to cure, this is allowing the varying forces of the curing Sika to push and pull the plastic as it pleases. So you may have one section of the join curing first and this pulls the plastic sheet in its direction, then as another section cures it pulls the sheet back in that direction. The effect of these forces may be exacerbated by the countersunk heads which could be forcing the sheet up and down (in and out).

My chemistry knowledge of adhesives, sealants in general, makes me think that it would be better to use a very thin bead of Sika and tighten the bolts fully at the beginning. The Sika will take much longer to cure but i doubt this would be an issue for you.
It probably would also help if the holes weren't countersunk and you used pan-head bolts or similar. (as others have already mentioned)

Anyway, good luck with it!
i have used antisilicone & a razor blade for releasing perspex, but i cant say if its ok for your windows, you can make a frame out of brunzeel ply, ali or stainless, if you want it belt & braces bolting thru the plastic is not the way, because of point loading, since you have so many different materials on a boat you have to isolate the most brittle from the solid surround, look closely at why car windows survive,

[usa marine]

why not replace them with glass. you can obtain up to 1/2" -or 12mm/

Thanks
Guill
USA MArine Windows Mfg Inc.

[Ad Hoc]

Wynand

I'd be interested to hear what the results are.
I've used Lexan on many fast ferries before, and all glued too. Never had any problems. So this is indeed interesting and a mystery.

Notes on the bolt and glue...makes sense in the big blue (I think I would too)!

[Boston]

Wynand
I have a lot of experience with glass also various polycarbonate's
what got you was the drilling
you should have burned through with an iron
nd not even touched it with a drill
that will always cause micro fracturing that will eventually lead to a crack
even if you then try and touch it up with an iron after
what you needed to do was hit it with welding solution
also plastics need so much breathing room that we typically never recommend them in applications over a few sq ft
it was even placing them under slight bolt pressure that illicited the cracks to run
funny thing is you might be able to fix it
if you take the window out and put it on a flat surface there is a plastic welding solution that through capillary action might just seal it all back together so well that all you will need to do is flame off the surface and it could look like new
it takes talent to be that good with the weld but Ive seen guys who can do it
if you need some welding solution I could mail you some as its available just down the street from me

as for bolting the windows back in
dont do it
eventually those bolts will get up against the sides of one of the plastic and begin chewing into it
possibly causing another crack
you want to float all glazing all the time

I have another post concerning glazing that I wrote for a newbe
let me go cut and paste it so you can get an idea of what Im on about
feel free to ignore the impertinent parts
but there is a decent description of how to float glass in a opening

[Boston]

tempered glass is best
but there are some important considerations

tempered glass as has been mentioned will scratch easier than annealed
its surface is actually made softer by the tempering process

there are various tempering processed which produce various levels of strength so be carefull what you buy

the edges of tempered glass are its Achilles heal, they should be floated above any hard surfaces on glass blocks ( non expanded neoprene is best )
two blocks on each side placed a few inches in from each corner
you want two inches of blocking per block for each twenty lbs of glass
dont go by sq ft as the UBC states because they wrote it for 1/4 and you should go with 3/8 or better

laminated tempered is best in high risk applications

use a 3m blast resistant film in really really high risk applications on the interior only as it scratches easily but will prevent flying micro chips in case of a blow out

you cant really combine the laminates between temp and not temp because the bending characteristics are so different
which means that if you get hit by a sneaker wave the outer layer of annealed glass would likely crack although the inner temp would remain stable

the glass should be stopped into place not relying on glue to hold it at any point
sealant holds out the water
stops hold in the glass

use a steal frame
it has the closest expansion and contraction characteristics to glass and so when the sun hits it they will move together placing the least resistance on the sealant

use sealant not a gasket
there is basically no gasketing system that has yet to be actually waterproof
the rubber contracts in cold airs and will leave a gap somewhere no mater how well you do the job

double sided sticky tape works ok
but EDPM tape is better
this tape should be applied to the frame in contact with the glass face and the stop
not the glass
its structural and wont compress to much and you can bridge over it easily with the sealant

bridge bead the sealant

Sikaflex is your best bet on sealant
takes forever to dry but remains flexible and sticks to just about anything

sill glaze is a silicone product but also works well
you cant paint it

you have to paint Sika

clean the living crap out of the bare metal surfaces with denatured alcohol before you begin the process and again before you seal it

if you have oil or grease on the metal and you dont clean it off with some kind of solvent and then clean again with the alcohol you going to have a leak

stops
the best stop system is one that allows easy repair
the inside stop should be monolithic to the frame for best strength
the outside stop should be screwed in with sufficient number of screws to counter act any possibility of a blow out by a factor of 3
and Ive seen tornado's that will suck the bark off a tree
so do not underestimate the power of this force

counter sink all the screw's in the stops and use Philips head screw's
why people even consider flat head is a mystery to me
but they sure end up breaking a lot of glass
square drive is ok but no reason to get silly with the screw's
just get good solid large diameter screw's made out of the same stuff as the frame
steal
yes they may rust if you dont treat em right
so treat em right

pay em with silicone an cap em with Sika after cleaning the living crap again out of the counter sunk area so the Sika sticks
Sika doesnt like to stick to silicone or silicone residue so clean it or die

you need the screw's to expand and contract "breath" with the frame or they will work
bad sitch
that will eventually break a window as fast as anything else will
and when you least expect it

stops should be square section
irregular shapes stops can place undue stresses on the glass

screw's need to be paid in
but not with a glue like substance or at least to strong a one
I tend to use any old bar soap or wax when applied in wood but your in metal so you will want something that will prevent water infiltration
for Christ sakes dont use epoxy or Sika you will never get the dam screw out
I tend to use silicone as it will break free easily in this application

http://i354.photobucket.com/albums/r420/maxtanks/conservaotory.jpg

[Wynand N]

Thanks for the input thus far.

I decided to dump the Lexan and got a piece of Perspex 12mm thick to install as a test run on the smallest window with only a curvature of 4mm over about 650mm that is basically a flat surface.

Done everything by the book and burned the holes through with a blunt drill bit. All burr was removed with a sharp Stanley knive on the holes. No counter sunk this time around.
The Perspex was primed and a generous layer of Sika 295 applied and window bedded onto frame. All bolts slipped loosely through holes and we start tightening from the centre out - much the same way one torque a cylinder head. Everything went fine until the very last bolt was tightened. Just when it start to tighten up, there was a snap like a whip cracking and the Prespex cracked from the hole and suddenly its mate also cracked.:(

At the time the last bolt went in the little curvature of the Perspex was already taken up with the first bolts installed, and the mystery is why did the last one screwed up the window?

Perhaps it is best I therma-bend the glass with hot air until it lies snug against the frame and just Sika it into place sans the holes and bolts. Again, that decision has to be taken by my client.
I had fitted many a window as a builder - all Perspex - many of them with plenty of curvature and holes counter sunk with no shit ever, this is a first for me and drives me up the wall.
BTW, a window of the laminated Lexan to Perspex units that cracked badly on the Lexan sides only, was sent to Johannesburg to Lexan people for an analysis of it as they cannot believe it either.

Perhaps someone out there has a little Voodoo doll with my name on it:?: Please Manie, take my name out of your blacklist ....otherwise I have to go and see my witchdoctor down the road.:eek:

[Ad Hoc]

wynand

Ahh...getting interesting now.
The bolts, are these each normal to the lexan?
The frames the window sits in, are these normal to the lexan?
Finally, are all 4 bolt locations in-plane with each other?

Since a triangle is always in-plane with itself.

Take 3 random points in space, join them, you have a triangle, and the triangle is fair and in-plane.

However it is impossible to do this with 4 points, if the 4 points are not in-plane. Therefore the Q's above.

It may well be that one of your bolts, or even all of them, are not in-plane with the others. As such, when you tighten the last bolt (the first 3 could all be out of plane, but in-plane with each other) this creates an in-plane eccentric load. This is stressing up your lexan.

I get the impression this could well be the cause, from what you have written above.

All 4 bolts are not in-plane nor normal to the lexan.

[Manie B]

he he Wynand wasn't me

i aint got staaaaafffoooool to do with your bad karma

Please Manie, take my name out of your blacklist

from my previous post you know that i have a rather "negative disposition" to Lexan (how's that for smart English):D

but the plot thickens even further

as you know i was sailing in Durban this easter weekend because i am working towards my DOC (Durban Operators Cerificate, without which very soon you wont be able to get in and out of the harbour - period)

so anyway while down there i had a good look at


you guessed right


WINDOWS :p :p :p


so well just for the record

laminated safety glass (flat) is IN

lexan is OUT



not that i want to rain on anybodies parade
but struggle on :P
i am sitting here on the side lines taking it all in


NOW GET THAT BLOOMIN BOAT FINISHED
sh1t i want to sail on her before i am a seriously old fart :D

[Boston]

Ive cut mountains of that stuff by scoring and snapping and always made it a point to never bolt through but always clamp with whatever glassing system I used

if you cut the piece on a table saw or some other type cutting tool you have loaded the edge with micro fractures just waiting for that little bit of pressure it takes to snap it just as if you had scored the surface instead of the edge

you must snap all edges and flame them off before you bed the piece
also give up on the bolts and let the piece float in the frame on glass blocks
bolt the stops in parallel to the glassing
if the frame has the bolts to close to the edge to allow you to miss the piece then use a rubber grommet in each hole to protect the edge of the plastic from the metal
it only takes a micro-fracture to start a run like what Im seeing in your pictures


best
and sorry to hear you are having such problems
you need not heat bend non laminated plastics as they remember a deformation just fine
once you place the piece in the frame a simple hair dryer will do the trick just fine if you wish to relieve any stress on the piece immediately

Q
am I seeing that picture correctly
is there no frame or supporting structure surrounding the edge of that glassing

Q
did you flame off all your edges and use welding fluid on all holes to clean up any micro-fractures
a utility knife is not adequate to prepare an edge to finish as it to will leave fractures that could lead to a run

once again
best of luck
B

[Manie B]

oh and even Beneteau with all the smart windows and hatches

couple of years in the Durban sun and you cant see thru

all gone hazy with millions of fine little cracks all over

enjoy :rolleyes:

[Boston]

the binders in all plastics photodegrade rapidly
leading to that milky appearance so often associated with plastic
also plastic scratches so easily that even just salt crystals that form on its surface can scratch it if disturbed

laminated tempered is the way to go
and the thickness is nearly unlimited
the New England aquarium had 9" windows for there big tank
laminated out of 1/4

I prefer filmed tempered in most overhead applications but on shipboard installations of any kind I have consistently gone with triple laminated tempered with out a single blow out
ever

stuff is virtually bullet proof
3/4 tipple lami with blast resistant film will stop a bullet I think its up to a .44 cal without breaking the inner light

dont quote me on that Ill have to look it up but the specs for bullet proof are clearly in favor of multiple layers of glass

you can get bulletproof plastics but they are generally single layers they do not hold blast resistant films well and do not survive multiple strikes nearly as well

getting back on track the most likely reason Wynand is having problems is because of either that the screws are contacting the edges of the holes and scoring them creating the start to a run or because of improper detailing of the cut edges
IE
even if you score and snap that stuff its still necessary to flame it clean before you do anything else with it

I would not recommend any drilling of any kind on plastic
specifically because of micro fracturing
always use a soldering iron or a heated punch to burn through
then clean up the hole buy breaking off the slag with your fingers and then flame it clean with a plumbers torch

best of luck and if Im seeing that correctly there is no compression ring around the edges of that window

bad plan mate
all glassing should be in a compression ring (stopped) and floated then bridge beaded to a adequate sealant surface
IE
never try to bead to the edge of a thin piece of metal
always allow for a decent contact of sealant in setting were the bead can be inspected and repaired with out having to dismantle the hole window

best of luck
B

[marshmat]

Wynand,

Perspex and its counterparts from other brands, ie. poly(methyl-methacrylate) is in fairly common use in the physics lab where I've been spending most of my time lately. We use a fair bit of polycarbonate as well. Strange cracks such as you're getting here are not uncommon with the weird shapes that are sometimes needed for the labs.

The explanation I got from our shop guys was similar to what Mr. Boston suggested above: Any time a sharp-edged cutter (band saw, drill, etc.) touches the stuff cold, you end up with microfractures at the cut edge. Sometimes you can't even see them, other times there are slight, sharp changes in the refractive properties of the clear sheet near the hole. But unless it's re-annealed somehow, the fractures are there, complete with built-in stresses, just waiting to cause problems.

On one project a few years ago, we couldn't see the fractures at all until the acrylic cement was applied to bond the parts together- at which point the edge suddenly went all cracked and crazed in a matter of minutes. Mechanical stress applied once this had happened would sometimes cause the PMMA (or sometimes polycarbonate) sheet to develop long, arcing cracks similar to the ones in your photos.

The obvious solution- avoid stress concentrations- isn't enough, because the micro-cracks lead to enormous stress concentrations no matter how much you spread out the load on the fastener. I've been told that there are special annealing techniques that can be done after cutting, but before fitting, to avoid this. I believe it consists of heating the sheet to near its glass transition temperature for a certain period of time, so that the area near the microcracks can regain its original amorphous structure, but I don't know the details.

Still, I've basically given up on mechanical fasteners for these plastics.... the way commercial windows are now done, with either a pressure plate or a structural adhesive, seems to work better.

Sika 295 is too flexible to cause this kind of stress (and it's not strong or hard enough to damage Perspex- it tears at 160 psi, shore A is only 35), so I don't think that's your problem. Somehow, there is enough of a stress concentration at one or more of those bolt holes to set the microfractures loose.


Edit- the glass transition temperature for the annealing is approx. 124 C and the melting point is about 130, so you have a 6-degree C range in which the annealing will work as intended.

[Boston]

I've been told that there are special annealing techniques that can be done after cutting, but before fitting, to avoid this. I believe it consists of heating the sheet to near its glass transition temperature for a certain period of time, so that the area near the microcracks can regain its original amorphous structure

well said
Im going to have to use that next time I want to charge someone a fortune for flaming there edges

what all that means Mr Wynand is that your going to take a regular old turbo torch ( plumbers torch ) and pass it slowly but not so slowly that you deform the surface of the material over the cut surface
carefully going back and forth untill you have a clear surface at the edge similar to what you have on the surface

a little practice and you will have a perfectly clear edge and not have damaged the ocular quality of the surface in its vicinity
the thicker the piece the easier it is

best of luck
and thanks for seconding me on that Mr Mellow I was beginning to think I was the only one who ever worked with the stuff on any kind of regular basis

cheers
B

[Jimbo1490]

We fit acrylic and polycarbonate glazing in aircraft and face the cracking issue all the time. One helpful hint is to always allow a generous amount of expansion room around the entire window. For a very long pane, this will require that the lengthwise expansion room be proportional to the length. On a square pane of perhaps a 1/2 meter I would allow somewhere between 6 and 10 mm free space between the edges of the pane and other structure. This implies large flanges for the windows and a flexible sealant, perhaps even a non-hardening sealant.

It gets even trickier for through hole fasteners. The through hole must be generously sized. For a 6mm bolt, the hole should be at least 10mm, 12 would be OK. Cutting the edges of the holes out (so that they become slots rather than through-holes) is another good option.

As Boston mentioned, it is really preferable to clamp rather than through-bolt plastic glazing as this solves many problems in one fell swoop.

There are specially ground drill bits (available at Grainger, MSC, J&L here in the States, your local machine shop supply where you are) that reduce the microcracking when drilling plastics.

After drilling, you can anneal with a 'clean' heat source such as a hydrogen torch, or a plastic welder, which is probably more readily available, and less potentially damaging.

Jimbo

[Steve W]

Ive had lots of challenges in over 35yrs of boatbuilding and repair work but fortunatly windows have not been one of them.From what i read on these forums i do just about everything "wrong" and yet in several hundred windows i have never had a cracked window that i know of.I have used glass ,polycarbonate and acrylic and all have their strengths and weakneses.My sons 25ft boat which i built 22yrs ago still has the original windows which are lexan screwed to the outside with countersunk screws on silicone,no cracks,no leaks bot will need replacing sometime as you havnt been able to see clearly out of them for years.This is my only complaint about polycarbonate. I currently have a C&C sailboat which had 3/8" acrylic windows bonded into molded rabbets with no fasteners, there is curvature in the cabinsides and the adhesive had failed at the ends and both the windows were broken top to bottom and came out in several pieces so they didnt need fastener holes to cause them to crack and they had plenty of room around them to allow for thermal expansion so i dont know why they broke but this is very common with C&Cs. I used a combination approach to replace them using 3/8"acrylic cut to allow a generous, even gap between the plexi and the edge of the rabbet. I used a minimum of countersunk machine screw as a belt and suspenders approach,i installed the windows on a thin glazing tape on the inner edge and then pumped in silpruf 2000 silicone which stops at the glazing tape,then tooled it off on the outside with a wet finger.I tend to mask off everything and dont use solvent to clean up if possible.
When i instal a plastic window on the outside of a cabin with no frame or rabbet i drill the holes with a special grind bit,countersink it,mask the cabinside,position it on the cabinside and mark the hole on the boat with a vix self centering bit and run a knife around the window to cut the masking, remove the window and drill the holes thru using a drill guide to keep them square,i then prep all the faying surfaces, then i glue thin plastic flat washers to each hole on the cabinside with superglue,this allows me to tighten the fasteners snug without squeezing out all the silicone,i then install the window and the next day peel the masking.
Steve.

[Boston]

in the process of repair Ive often found windows that survived surprisingly well for how poorly they were installed
catastrophic failure isnt that common its just far more likely to occur when the job is done poorly

[Jimbo1490]

I think a hidden factor is just how rigid (or flexible) the surrounding frame happens to be, or the Cte is close enough. Sometimes you just get lucky that way. Also, surface mounting is a big plus since the panel won't be 'trapped' all around by the frame. Through holes are still a problem, though.

I'm working on a plane right now where one piece of glazing was obviously mis-installed; not enough expansion room all around. But it's a 'clamped' rather than 'through-hole' installation so the whole window just caved in permanently, but no cracks anywhere.

Jimbo