Basinranger wrote:I've been following this site & the discussion of using cheap insulation foam like you'd get @ home depot as a structural core.(I assume we are talking about blue dow foam insulation board - although there are MANY dow foams that are blue... only three are listed as structural foam - panel core 20 30 & 40 - and you don't find it at home depot...)
My first response is are you kidding? I'm looking forwards to seeing the results of using cheap foam structurally because if it works it would be great! It's a big leap from using thick ply skins with extensive solid wood infills and cheap insulation board glued in to an actual "composite sandwich structural panel" - meaning a panel where the skins take the load, giving you the weight savings & stiffness which make the trouble & expense worth it.
Never having made anything out of it I can't say it wouldn't work - but I have made many panels & structures out of structural cores - mainly balsa with both ply & glass skins bonded with epoxy.
Here's the argument for why it's a terrible idea - Without getting into too much detail the core material needs to have two main properties beside holding the two skins apart. It has to have sufficient compressive strength to keep the "top" skin from buckling - so it transfers the load efficiently to the other skin. Also it has to have enough shear strength to keep the skins from de-bonding & sliding off the core. It's also useful to have enough compressive strength to resist spot loadings from impacts.
Dow doesn't list the structural properties of it's cheap insulation foam - because it's not structural...
But here's how it's structural foam compares to the structural core materials I'm familiar with. Their properties get stronger as the density goes up in weight.
Dow Panel Core 20 - 20psi in compression 25 psi in sheer
Dow Panel Core 30 - 30psi in compression 35 psi in sheer
Baltek BalsaCore - 917psi to 3813psi in compression, 262psi to 712psi in sheer
Nida-core's H8PP 10 to 25 - 188psi to 348 psi in compression, 72psi to 87psi in sheer
Core Cell - 80psi to 336 psi in compression, 98psi to 287 psi in sheer
It's a good idea to know the basic properties of the materials you are using - often that explains their attractive price... I consider using 13.00 a sheet foam structurally to be a big risk - if anybody has experience with it or better yet Dows Panel Core product I'd love to hear about it, building lighter is better! Peter
You tell ‘em Glassice! (BTW, Glassice ain’t afraid to give you his opinion!)
Well, there’s structural foam and then there’s ‘structural’ foam HEhe
Dow doesn’t want to accept liability for the lighter-weight mass-produced foam, but the properties aren’t much different than the lower density foam that you listed (If I’m not mistaken, the densities are the same). If you scrounge around the internet, you’ll find that there are a lot of people using it to build all kinds of stuff.
My buddy is building and airplane out of it right now (from scratch, and not his first aircraft build!). He’s used it for the ribs, parts of the empennage, and parts of the fuselage. Interestingly enough, it wasn’t lighter to use it for the whole fuselage (he could have) as a wood (fir) stick frame (truss) turned out to be lighter. The whole plane including 3-axis control system and engine will be under the 254 pounds required to be categorized as an ultra light aircraft. He has already completed a successful static test of the wing (that’s pretty much the biggest and most important test as it’s generally considered bad form to lose the wings while flying!).
The process for the ‘structural’ foam might be a little more tightly controlled so that Dow can ‘guarantee’ the mechanical properties (or probably not), but pretty much the strength is dictated by the core density. If the core is denser, it can carry more loads (both tension and compression). Generally, and especially in the case of the Home Depot foam (also available at other bog box stores like Lowes) I’m guessing that it’s pretty much identical – it’s just mass produced. My airplane buddy is using reduced allowables (I think he is using 15 psi for the shear allowable), but when another airplane buddy of his did some testing, the strength was actually nearly 3 times that (around 70 psi if I remember correctly).
The problem with actual structural foams is that they are VERY expensive [as you may have noticed]; so, the key is getting a non-structural foam to do the task if possible (it isn’t always possible). For local load introduction we can use a puck (round insert which is generally potted into a composite sandwich and is either a metallic part, or made from potting compound, but we can use plywood, and it will do the same thing – it’s just not as fancy (we’re not about fancy, we’re about cost-effective (OK cheap, but don’t tell anyone)). I’d point out that there is a difference between the extruded foam (blue, pink, and sometimes white) and the white ‘beaded’ foam. The later is crap (and even it can be used if done carefully, but I personally wouldn’t use it).
Also, Balsa is generally pretty heavy core (relatively speaking), so that’s why the strength numbers look so good. Nidacore, is polypropylene, which is reportedly tough as nails, but it’s a bit expensive and flexible for my liking. Structural foams other advantage is that they come in many graduated strengths so that you can tailor the strength requirements very closely by splicing different densities in the same panel, only using heavier core where it is required (using this type of material is all about weight-saving in the end).
I could discuss composite failure modes with you until your eyes rolled back into your sockets and you slipped into a coma from which you likely would not return, but I hate to do that to anybody. Suffice it to say that generally light-weight thin-skinned sandwiches fail in either core shear, or facesheet wrinkling. Core shear is core shear, so that one is easy to figure, but facesheet wrinkling is a function of not only facesheet thickness, but core stiffness (denser foam is stiffer). If you provide more stability (stiffness) to support the facesheet, the wrinkling allowable goes up; when you make the facesheet thicker, the wrinkling allowable goes up. So, you can basically use thicker skins, because we don’t want to use denser foam, since it is more expensive (it may not be perfectly optimized, but it’s a wee bit of a compromise to save a lot of additional cost). Also, you can reinforce the core with plywood locally if you need to (Like where the axles attach, for instance). Usually this reinforcement won’t have to be very large as the load, and thus stress, spreads out pretty quickly.
One of the biggest areas of concern (pointed out by Glassice) is that core can peel away from the foam (I think this isn’t necessarily how the failure starts, but it can easily be mistaken for this failure mode). Regardless of how the initial sheet-to-skin failure occurs, the delamination can spread. This isn’t unique to the Home Depot foam, but all foams. I read a recommendation by one foam manufacturer that said you should put some holes 2 mm dia. X 3 mm deep (5/64 X 7/64 in, or thereabouts) spaced from 50-100 mm (2-4 inches). It’s also probably a good idea to rough up the surface with 60-80 grit and of course remove any films or coating that might be on the sheet. I think it likely that the initial failure is a facesheet wrinkling, followed up by facesheet peel. The hole should help arrest that (time will tell).
One other source that I m checking out is there are apparently people working in the insulation business that can custom slice whatever thickness you want out of a big block (the standard block dimensions are 5’X5’X10’) of the stuff and can vary the density as well; so, theoretically, you could get a bit denser (stronger) foam if you need it and get uniform and specific thickness sheets as required. I’m still checking that one out – there are evidently a couple of those places near Ft. Worth (where my permanent home is located).
I will be doing some verification test of mechanical properties before my build begins. All of the tests are pretty standardized, so that part isn’t particularly hard to do.
BTW, I’m CERTAINLY not the first to do this – there have been several in the past. Check the Uber ultra thread – I made a finite element model of a tear body and put in all of the properties for the sandwich when I ran the model (I hear I wasn’t the first to do that either, but I did document it so others could take a look at it).
The only problem with pre-made sandwiches is that you can’t do a curved surface, which limits the profile a bit.
by eamarquardt » Wed Jul 28, 2010 11:36 am 
