impact resistance of thin glass on structural backing layer 2

[damienmorton]

I am building a device comprised of a 0.5mm-1mm thin glass layer bonded to a rigid structural backing layer, and this device is required to be impact resistant.

The impact resistance test for the glass is the ball-drop test, and I am wondering if there is a way of estimating what adhesive properties would maximise the impact resistance in this relatively low-velocity environment (<100m/s and <100g impactors).

For example, would a thicker softer layer act as an energy absorbing layer, or would it merely facilitate local deformation and failure of the glass.

 

[RPstress]

Four days and no takers...well, here's my five eggs:

Assuming you want to minimise mass and/or the substrate thickness, then it might help if the substrate can be made stiff in bending and flexible in through-thickness compression. It does depend quite a bit on how the substrate is being supported. This approach assumes the substrate is placed in bending by the impact.

So, you might be able to minimise substrate thickness with biaxial or QI carbon fibre in an elastomer, or maybe in polyethylene, polypropylene or Nylon...if there's no real constaint on thickness then using a thick substrate with a lower fibre volume might be effective.

If you could introduce carbon at about 30° or 45° through the thickness it might also help, though it would be tricky (essentially a 3D weave). You don't want the substrate to be all that flexible in shear, as well as making it stiff in bending.

Essentially you want the impact force to spread out sideways.

Another poster, PanelGuy in thread thread327-246750, reckoned that a thick honeycomb substate could support a sheet of glass with impacts. If you're looking to minimise mass and not thickness then a sandwich substrate might work well.

If it's important to avoid too much testing I'd advise doing some FE.

However, practically it would help to make some examples and test them. You'd need a gun of some sort to fire the projectile. 100 m/s is slow by ballistic standards but still a bit too high for a falling weight.

If you can do a bit of testing with a test machine then you can estimate the impact force using the formulas in Roark (article 15.3 in Ed. 6), and then approximate things with a static test. You'd need to estimate the flexibility of the assembly being impacted to use Roark.

 

[damienmorton]

Hi RP,

Thanks for your comments.

Right now, the substrate is going to be metal - either 2mm titanium or 3mm aluminum.

The constraints on the substrate are that it should be thin and have a high yield and microyield strength. Basically, a human being should only be able to permanently deform this device with great difficulty - i.e. bending over the knee, or putting it in a back pocket and sitting on it shouldn't induce any permanent deformation - it should remain flat under those circumstances.

I saw PanelGuys post earlier, unfortunately a thick honeycomb is out of the question, and its not entirely clear what the mechanism in play is.

Ive also talked with several other people, a solar guy whose thoughts are that the glass should be as directly bonded to the backing material as possible, and a ceramics dude who spoke about the failure modes of ceramic tiles - any voids in the cement underneath the tile are weak points.

100 m/s is the top end, and I suspect that none of my designs will survive even 30 m/s.

I am going to resort to testing - got myself a bunch of samples of 3M VHB tapes of various thicknesses and softnesses on order and, well, drop ball bearings on my glass mockups until they break.

I was hoping that there might be a model out there ready for use, but looks like there isnt.

 

[madvlad]

google says that if you want to glue glass to metal, you use this:

http://www.dymax.com/products/adhesive_products/429-series.php

They say it's some sort of acrylate which is cured by UV. Acrylics I think are pretty stiff (high E)...among the highest modulus non-thermoset polymers anyway. I figure that's what you need for good load transfer to the substrate.

Once I got to the glue webpage, found out they use the same glue for bonding layers in manufacture of bulletproof glass...Sounds promising, doesn't it?

Also, maybe if you're not having luck when it comes down to it...try pyrex instead of glass? FWIW, pyrex things don't break as easily when you drop them.

...I'm assuming that using polycarbonate or acrylic instead of glass isn't appropriate...

 

[RPstress]

Hmm.

For 100 m/s (>220 mph) I get a drop height of 500 m...for 30 m/s it's 45 m. s=v^2/2a?

How would the glass and substrate be supported when struck?

For what it's worth, in bending 3 mm of Al with 1 mm of glass is about twice as stiff as 2 mm of Ti with 1 mm of glass.

 

[damienmorton]

Ugh - you're right my numbers are confusing.

There are two situations I am trying to model:

First is daily use - dropping the thing or throwing it around. Ideally, as a demonstration of its toughness, I'd like to make a video of someone banging nails into wood with the glass face of the device. How would the supports be modelled here? In the first case, I guess its probably better to model it as dropping the device onto a ball bearing rather than the other way around. In the second case, the same but with a soft weight added to the back of it. This is all happening in the 1-10 m/s region, I guess.

Second is something more like an airplane crash, but in this case there will be a protective cover of some kind (rubber ?). How will it be supported in this case - I guess there are two cases - the first is if the device is flying around on its own and hits something hard, which is covered above, and the second is if its got a mass behind it (i.e. a person), in which case it would be modelled as a soft mass behind the device as it hits a hard object. This can all be happening in the faster regions, like >30 m/s.

The MIL-STD test for these devices is to drop a 11 pound, 1.5 inch tubular shotbag on it from a 8 feet with all edges supported.

Im inclined towards titanium because it is closely matched in CTE to the glass I am using, its corrosion resistance, and because of its microyield strength (which is quite a bit higher than that of aluminum), but also because it has a certain cachet as a miracle material.

The glass I have selected is pretty amazing stuff, and you can see a ~10cm segment of the stuff bending to around 1cm displacement here:

http://www.youtube.com/watch?v=Ope6uViLcEY

I think that a thinner more flexible device is preferable to a thicker stiffer device, assuming that the thinner device doesnt yield on bending, and assuming that the glass isnt bent beyond its limits. Thats in the static case, I dont know about the dynamic case at all.

 

[damienmorton]

RPStress, what did you use to compute the Glass/Ti and Glass/Al stiffnesses? I have been using various online beam bending calculators, but havent found one that works for laminates.

 

[RPstress]

I actually used LAP, but any classical laminate theory program will do—and is overkill, really.

For two layers like you have you can do it by hand—see thread327-250463 for a few bits of advice about compound sections.

http://home.comcast.net/~brandihampson/cd/index.html

is also quite helpful, especially if you have access to Mathcad.

 

[allcomposites]

Hi damienmorton,
We had a similar application in the past, in which localized low-velocity impacts had to be damped. The solution was to a add a damping layer between laminate and rigid structure.

There are plently of viscoelastic sheets on the market that apply. I think you have already the right solution, a 0.25mm VHB tape will be appropriate. You may have a limitation if the part is too wide. There are similar visoelastic sheets for larger parts at Roush Damping Products and SMAC.