Fiberglass techniques

[G_T]

Kevlar will sand if you wet sand it with fresh open grit high quality sandpaper. I've done it many many times. Don't even think about sanding it when it is dry!

A fiber in an epoxy matrix loses roughly 50% of its strength if it is even just 20 degrees misaligned with the applied stress. When one adds chopped fibers to epoxy, unless the ultimate layup is a thin film, the fibers have a 3D orientation that is fairly random (thin film is closer to 2D). The chopped fibers contribute nearly nothing to the ultimate stiffness of the joint on a fillet, compared to just the epoxy with some other thickening agent.

You can test this by putting down a bead on wax paper, with and without your additives. Wait until cured. Bend and see what happens.

The fibers - particularly if non-brittle ones are chosen - can help prevent or delay the formation of cracks under stress. Slightly. So adding fibers is not worthless, just not worth what people seem to think it is worth.

If you really want strength, and/or stiffness, you want to orient the fibers to the applied load. In the case of a fin fillet, that would be a number of layers of fabric of staggered widths and an appropriate overlap structure. That would make the fillet structural. Otherwise, the fillet is more cosmetic and aerodynamic than structural.

Gerald

 

[G_T]

Post way too long so rocketryforum blocked it, so broken into multiple parts. Please read it through.

I've done some testing to confirm what I wrote - years ago. It should be common knowledge for composite workers and engineers. There isn't anything mysterious going on. Each fiber is a long thin strand that contributes potentially a lot of strength and stiffness along the fiber direction, provided the fiber is adequately supported by the matrix to not collapse, and adequately adhered to the matrix to transmit the required force from the matrix to the fiber and vice-versa.

In a structural composite layup, one engineers the choice of fabrics and uni, along with their orientation, to achieve the desired structural properties based on some criteria of performance and/or cost.

Fibers which are not aligned with the applied force have little to no effect on the FRP being able to withstand the applied force. https://encrypted-tbn0.gstatic.com/..._OyfoAF6iVk94EV_oV5c4ge7K5a4t_UuG0_7vwJRh-IgR

Fibers which are short are not as effective as fibers which are long. A short fiber may not have sufficient bond to the matrix to exchange load to the capacity of the fiber itself to take load. https://www.plasticomp.com/wp-content/uploads/2018/07/Fiber-Length-Benefits.jpg

Some modern manufacturing methods for injected FRP using chopped strands have clever methods to enforce a fair degree of fiber alignment so the fibers are not truly randomly oriented in 3-space. This greatly improves the mechanical properties.

For optimal properties, the correct fiber volume ratio needs to be achieved, and the matrix needs to be degassed or at least have low air content. A typical desirable fiber volume ratio might be 60%. But the number is specific to the fabric and other materials employed. If you do it by weight which you should, then you'll have to convert based on the respective densities.

There are non-fiber reinforcement that are worth considering. There are also nano- and nano-fiber additives which can be added to improve the matrix properties even in an otherwise normal FRP layup.

Gerald

 

[G_T]

Different epoxy systems have different physical properties. Adhesion. Tensile strength. Compressive strength. Ability to wet-out fabric. Viscosity which affects the effort and skill required to achieve the proper fiber volume ratio. The all important Tg, or glass transition temperature. That's where cured epoxy goes from a polycrystalline solid to an amorphous solid, with considerable loss of physical properties. You want your operating conditions to stay under Tg by a sufficient margin.

Most epoxy systems can go through a post-cure heat treatment cycle to improve the mechanical properties by affecting the completeness of the cure and the remaining strain. Of considerable significance is usually the resulting Tg is raised to a few degrees above the max temp achieved in the post-cure cycle.

The max post cure temperature is specified by the manufacturer. DO NOT EXCEED! Doing so doesn't further improve Tg but instead destroys the epoxy or at least damages its properties.

Not all systems can be post-cure heat treated to improve mechanical properties and raise the Tg. West Systems is an example of a system which CANNOT BE IMPROVED by post-cure. Unfortunately. That was as-designed, for wooden boat applications. It is excellent for that. It is poor for most anything else.

Manufacturers of high grade epoxy systems publish technical specs which give details of mix, mechanical properties obtained, required or optional post-cure heat treatment cycles and the resulting properties, etc. If an epoxy system does not have such published data, then IMHO it is not suitable for aerospace applications. Certainly no certifying body would endorse it.

I prefer to stick to systems where such data is available. MGS being one I highly recommend and frequently endorse here. There are others of course. MGS just has some of the best properties for a room temp cure epoxy system and can be improved through optional post curing. It is certified in Germany and likely other countries for use in aircraft manufacturing, and is in fact the epoxy system used by a number of plane manufacturers.

Gerald

 

[G_T]

People mix epoxy wrong, through ignorance. For most systems, measure by weight, accurately. Mix part A and part B until they are clear. In the process, scrape the sides and bottom of the mixing container, and the mixing implement, repeatedly. Expect to take a minute for moderate quantities. Expect to have to use power tools for large quantities. Do not whip the epoxy. That adds lots of air. Fold and stir it...

For a real industrial application, then transfer the epoxy to another container. With a fresh mixing implement, mix again. Now it is thoroughly mixed.

For some applications it is worth vacuuming the epoxy, or subjecting to ultrasonics, for a rather short time to degas.

Spread the epoxy out to a thin layer on a working surface such as a plastic picnic plate. The thin layer allows the epoxy to radiate heat. Epoxy is exothermic. But the cure rate is also temperature dependent. The higher the temp, the faster the cure therefore the shorter the working time. Rough rule of thumb for chemical reactions - each 10 degrees C increase in temp halves the working time. It is not exact, but you get the idea. This applies equally to epoxy cure and R45 cure for propellants!

People wet out fabric wrong through ignorance, and thereby produce sub-standard structures. There are multiple good ways to wet out fabric. Two standard ways are resin infusion and hand layup.

A typical approach for hand layup is to roll down a first coat of epoxy, rather thin (may form lots of little beads like rain mist). Put DRY fabric on top of epoxy. Roll with relatively dry roller to pull epoxy into the fabric from the underside. When completed, roll a little more epoxy on top the fabric. Place next DRY fabric layer on top, repeat... This way epoxy comes up from under the fabric and air is not trapped between the layers. Air has no structurally useful properties in this work!

Fabric is chosen for specific application. The type and quantity of fiber needed needs to be aligned to the expected load. That is part of engineering a composite part. Fabric is a convenient structure for manipulating the fibers to put them where you want them.

Not all fabric is suitable for making a complete layup. Kevlar (A Dupont trademark for two types of Aramid fiber) for instance is strong in tension, has good durability (a variant is used in some bulletproof vests). But it is pathetic in compression. It also suffers from weak inter-layer bonding leading to some failures. So Kevlar is often used in alternating layers with S-glass which addresses these concerns.

S-glass is structural glass. It is typically Silane coated (or other) for greatly improved bond compatibility with epoxy and other matrix systems. Do not bother using uncoated fiberglass in a layup.

E-glass is electrical grade material. It is much weaker than S-glass and ideally should not be used for structural applications. The difference in practice is considerable. Most glass you find in small quantity is E-glass. S-glass is definitely much more expensive.

In a fabric, fiber is (usually) woven. The more fiber crossings per unit length - over one bundle of fibers and under the next - the weaker the fabric. That is because the fibers are oriented more in a direction off the plane of the fabric, and of course are under slightly more strain from being bent. Misaligned fibers do not contribute well to strength.

Many modern fabrics are spread tow. That is where the bundles of fibers are spread out to a layer one fiber thick as a wide ribbon. These ribbons are then woven. The result is few very slight fiber crossings so the properties of the fabric are maximized. Expect to pay for it though! This is nice stuff.

Unidirectional material is also available of course! Either as tow, or spread tow, or as fabric. Note that any fibers perpendicular to the uni that are used to hold it together act like fiber crossings and form weak lines. They are to be avoided.

Fabric comes in different weaves. 2x2 twill for instance is a good choice, as the fiber crossings are not all lined up the way they would be with 1x1 plain weave. The resulting fabric is more flexible, and the resulting layup is also stronger.

Once the wet work is complete, a layup needs to be cured. Typically it will be compressed by some mechanism - vacuum for instance - to compact the layup. This squeezes any remaining air bubbles to a smaller volume improving the matrix properties, and helps extrude any excess epoxy thereby improving the fiber volume ratio.

Once a part is cured enough to de-mold without warping, then it can be moved to post-cure heat treatment.

Gerald

 

[G_T]

Just a pet peeve here, so I'll skip over to talking about the structural layup of a rocket body tube.

What are the stresses a rocket body tube encounters? There is longitudinal bending. The longer and thinner (the higher the aspect ratio) the more bending stresses we expect.

How does a thin walled tube respond to bending stresses? It ovalizes, becoming thinner in the direction corresponding to the bending direction. As it becomes thinner this reduces the stiffness of the tube against that bend, and the foces concentrate. As the bending progresses further, the tube thins further, and at some point the tensile strength along the outside of the tube's bend, or compressive strength along the inside of the bend, or the resistance to ovalization is exceeded. Boom.

So the forces are longitudinal tension and compression along the tube, and ovalization perpendicular to the tube.

So how do we orient fibers to maximally resist the resulting stresses?

For the tube bending compression and tension, we want fibers parallel to the axis of the tube. Since the tube is thin walled, relative placement of those fibers within the thickness of the tube has little effect on the resulting properties. So we have flexability in that decision.

What about ovalization? Mentally zoom into a small patch of the tube. Now it looks like a slightly bent sheet that we are trying to bend a little more or else flatten out, depending on where that part resides on the stressed tube. To resist that bending or flattening, the fibers need to be aligned against that stress - that is perpendicular to the axis of the tube. Fibers in the middle of that thickness are under no stress from this localized bending and therefore are worthless. Fibers at the surface have the maximum ability to resist the localized ovalization forces.

So to resist ovalization, we want the appropriate fibers on the surface - inside surface and outside surface. These fibers are perpendicular to the axis of the tube.

To resist bending, that set of fibers are parallel to the axis of the tube. And since it doesn't matter much in the thin wall where they go, they go between the layers which are used to resist ovalization.

Choose the right type and quantity of fibers for the loads (need to compute) and there you have your optimal tube layup. That's how you do it.

Many places use filament winders to make tubes. What is the fiber orientation produced by a filament winder? It has the fibers at some angle to the axis of the tubes - usually a considerable angle like 45 degrees.

A fiber angle such as that is much more appropriate for making a drive shaft, where torque needs to be transmitted. But for a rocket body tube, it adds a lot more weight than would otherwise be needed as the fibers maximally poorly oriented to the expected stresses.

I've typed enough. I hope those who are not experienced in composite work read this through. There is a lot to learn here. I've probably only scratched the surface. This is a mature engineering field. There are texts on the subject. Feel free to dive in.

Over and out.

Gerald

 

[manixFan]

WOW!! I read it through and found it very worthwhile, thanks for taking all the time and energy to type that. Lots of really great info and explanations. No need to answer now, but I would like to follow up on:

"There are non-fiber reinforcement that are worth considering. There are also nano- and nano-fiber additives which can be added to improve the matrix properties even in an otherwise normal FRP layup."

As an aside, I have appreciated your considerable technical expertise and appreciate your willingness to share it on this forum through your many posts.


Tony

 

[G_T]

Hi Tony,

I'll have to dig up the old research if I find some time. Some years back, three products were available and one you could buy off the web. A bit pricy though. Ceramic based IIRC, though ceramic means lots of things. The little particles replace matrix between the fibers and have better mechanical properties than the matrix, as well as bonding well to the matrix. The result is improved average properties, at the cost of a small increase in density (and of course $).

PS to all the above - I forgot one very important point. NEVER DILUTE EPOXY! Even a couple percent alcohol addition reduces the cured mechanical properties by perhaps 40%. If you want a thinner epoxy, use a different laminating resin. Some such as MGS are quite watery. Others such as System 2000 are a fair bit thicker. There are a range of products available. Choose the appropriate one for your application. Just to note, if one absolutely positively has to thin epoxy, isobutyl alcohol is the preferred alcohol. It does the least damage, going from memory. But better not to go there.

Gerald

 

[boatgeek]

There is an awful lot of really good information up there. The only thing that I'll add in is a quote from my materials professor lo these many years ago: "Engineering is the science of good enough." For the rockets that most of us fly (say Mach 2 or less, not record breakers), a suboptimal epoxy or glass/carbon weave may be just fine. Heck, plywood and wood glue may well be fine. If you are flying high performance, then you definitely want to optimize as much as you can.

I guess it's the difference in optimization between a Camry and top fuel dragster. There are lots of ways a Camry could be tweaked to be faster or lighter, but that's not the only consideration Toyota has when they're designing it.

 

[G_T]

Tip-to-tip...

In most instances, tip-to-top is actually a mistake of sorts. When doing tip-to-top, what one is usually doing is trying to increase how strongly and how rigidly a fin is attached to a body tube, and to increase the speed where flutter might start.

Fin attachment is improved through what is essentially a composite layup fillet that is formed at the base of the fin. The larger the radius achieved, the greater the stiffening of the base of the fin. To a certain point, this also improves the smoothness of the airflow. To get to that point, it takes quite a few layers. Or, one could put down a smooth fillet first, then put on the layers. That takes less overall fabric for nearly the same mechanical result. Also the larger radius of the fillet vs the corner at the base of the fin makes the fabric have less of a stress concentrator so you don't need quite as much of it to get the job done.

Fin flutter onset speed is increased through having a fin that is appropriately rigid, and having the tip of the fin suitably light. The way people often do tip-to-tip adds lots of mass well out on the fin, but is it structurally required? HECK NO!

A fin is a small low aspect ratio wing. Look at this graph for the bending moment of a wing https://static.rcgroups.net/forums/attachments/5/4/6/7/8/a10390510-61-Wing Bending Moment (1).png?d=1506629119 and note that the bending load decreases very rapidly as one looks away from the root. Lots of reinforcing near the root makes sense. Minimal or no reinforcing towards the tips makes sense. You want the stiffness and strength of the resulting fin to match the general shape of the load distribution in that graph, perhaps biased a bit to be even stronger towards the root. After all, if it were to fail, you want to control the failure to occur near the tip rather than ripping the whole fin off!

That's another point about designing in general - design where it will fail if/when it fails. That leads to more favorable outcome than just letting it fail at random.

Gerald

PS - I went back and fixed lots of typos in my earlier posts.

 

[G_T]

Hmmm, I forgot another bit of basic info for composite layup work.

Epoxy or other matrix material can make a bond to a substrate that is mechanical, or it can be chemical, or a combination of the two. Chemical is by far the preferred method. It is much stronger.

Suppose one is doing T-T over a fillet like I mentioned in the previous post. One could make a fillet and let it cure, then sand it out, clean it up, then lay up the T-T fabric. That will achieve a mechanical bond.

One could alternatively put some peel-ply on the surface of the fabric and rip it off when cured.

Did you know that peel ply is considered a no-no in aerospace structural work? The reason is what happens when you rip it off. It causes a whole field of micro fractures that extend down into the matrix from the surface where the peel ply used to reside. The next layup doesn't manage to get epoxy down into these micro cracks so they are not healed up just sealed over. That leaves a weak layer just under the surface of the new layup. That is a field of nucleation sites for failure.

So don't use peel-ply!

Ok, suppose you want to have a chemical bond. Do not allow your fillet to completely cure. If it can still be dented by a fingernail for instance, then it is not fully cured and you can get a chemical bond. Of course don't use a fingernail - you don't want to transfer any hand oils. Oil is quite the enemy of a good bond.

Gerald

 

[G_T]

More on peel-ply and fin layup work...

It seems funny to me all the extra unnecessary work people go through on fins. The end goal is usually a nice looking decently painted smooth fin, right? Do you think composite body panels on cars, or composite wings on commercial airlines, are made with peel-ply surfaces which are ripped off then sanded, filled, primed, sanded some more, then painted, etc? Way too labor intensive therefore way too expensive. Also completely unnecessary.

For years the way parts have been made is to start from a mold (tooling surface for the part) and work from the outside of the part to the inside. When the part is cured it is popped from the mold, already finished.

The mold can be just about anything. For a fin, I'd use semi-rigid sheets of plastic such as Mylar. Use un-scratched sheets and wax them so they don't stick permanently to the fin. Paint the wax. Thin misted coats, letting it get a ways towards dry (but not dry - we want chemical rather than physical bonds, right?) before spraying the next misted coat. Then if you like spray a light mist of primer, but you probably don't need it, because... you don't wait for the paint to completely dry. Chemical, remember? Then add a little epoxy, some dry fabric, roll it out, a little more epoxy, another layer of fabric, roll it out...

Repeat until all the fabric has been put on the way you want it. Now put the mylars and fabric over the fin. Rig up a vacuum bag and vacuum in place. When cured, peel off the mylar and there you have painted in place fins needing only a little edge cleanup.

I didn't mention the fillets. You'll likely want to have their layup curing but not yet fully cured (chemical bond, remember?) when applying the tapered skins on the Mylars.

I threw all this out to get you thinking about how to use composites in a more efficient fashion. I gave an example of a simplified version of part of this in my thread https://www.rocketryforum.com/threads/sprite-6-and-a-baby-o.37382/ Since I documented it I have referenced it several times so hopefully I'm not boring you with it! But in there you'll see a simplified setup for bagging fins that come out painted. They were attached with metal angle and bolts rather than composite fillets. So it is of course simpler than what I discussed above. Once you are used to doing this sort of stuff, bagging is easy. You wouldn't even consider using peel-ply!

Gerald

 

[G_T]

I think I've dragged this thread way too far OT. I apologize to the Original Author, and I'll try to shut up now on this sort of stuff.

Gerald

 

[Nytrunner]

On the contrary, this has been perhaps the most concise presentation of composite layup mechanics I've seen on this forum.

I, for one, thank you

 

[rocketcharlie]

OP here. This really caught fire like I didnt expect it to. I have learned that some commonly held beliefs about lay up, fillets and tip to tip are up for review. I am still scratching my head over trying to align 1/16" milled fiberglass. When I include it in a mix it is all jumbled up into every possible orientation.

 

[G_T]

https://www.grm-systems.cz/en/epoxy
https://www.cotronics.com/vo/cotr/ea_1.htm

If interested in various epoxy systems, find sites like these... Then dig into the technical specs. Ones that interest you, investigate further.

Good sources for fabric and supplies:
https://www.r-g.de/en/home.html
https://www.sweetcomposites.com/
https://www.sollercomposites.com/
https://www.uscomposites.com/
https://www.fibreglast.com/
https://www.aircraftspruce.com/
https://store.acpsales.com/
etc. I've used all the above at one time or another, and probably many more. That was just a short list from memory and quick googling to get the addresses right. I'm sure I've forgotten some. Note, even though R&G is in Germany, often times it is the cheapest source for carbon fiber, unless things have changed. I've bought as much fabric from them as everyone else combined. But then again, I probably still have more scraps than most of you have ever used! I did not end up doing some of the things I had planned. Work gets in the way...

Gerald

PS - Also look for convenient references such as https://www.r-g.de/w/images/6/69/R&G_Handbuch.pdf and various references online for aircraft composites repair.

 

[Steve Shannon]

I think I've dragged this thread way too far OT. I apologize to the Original Author, and I'll try to shut up now on this sort of stuff.

Gerald

Not at all. This has been excellent.

 

[manixFan]

Not at all. This has been excellent.

I agree, super useful post and again, thanks for taking the time to pass it along. I went back and read the build thread you referenced and found it very instructive as well.


Tony

 

[REK]

This is very good information indeed. This should be a sticky thread.

As a composite maker I approve of this.

 

[tim cubbedge]

This should be a sticky thread.

It is!

 

[Senior Space Cadet]

Kevlar will sand if you wet sand it with fresh open grit high quality sandpaper. I've done it many many times. Don't even think about sanding it when it is dry!

A fiber in an epoxy matrix loses roughly 50% of its strength if it is even just 20 degrees misaligned with the applied stress. When one adds chopped fibers to epoxy, unless the ultimate layup is a thin film, the fibers have a 3D orientation that is fairly random (thin film is closer to 2D). The chopped fibers contribute nearly nothing to the ultimate stiffness of the joint on a fillet, compared to just the epoxy with some other thickening agent.

You can test this by putting down a bead on wax paper, with and without your additives. Wait until cured. Bend and see what happens.

The fibers - particularly if non-brittle ones are chosen - can help prevent or delay the formation of cracks under stress. Slightly. So adding fibers is not worthless, just not worth what people seem to think it is worth.

If you really want strength, and/or stiffness, you want to orient the fibers to the applied load. In the case of a fin fillet, that would be a number of layers of fabric of staggered widths and an appropriate overlap structure. That would make the fillet structural. Otherwise, the fillet is more cosmetic and aerodynamic than structural.

Gerald

Well, that's disturbing information. I used chopped fiberglass for the fillets on my stich and glue boat. It is a commonly held belief that chopped glass adds strength to epoxy. All the books on boat building say so. Luckily that's not the main thing holding my boat together.

 

[G_T]

Mix up some epoxy. Split it into two batches. Add fiber like you are used to doing to one of the batches. Lay a bead down of the neat epoxy and a bead down of the reinforced epoxy on a sheet of wax paper or something similar. Let them thoroughly cure. Bend them, and see how they behave.

If you are using the fiber filled epoxy as a glue, then generally it is squished into a thin layer. Three things might be happening there. First, the addition of the fiber may prevent the epoxy layer from being squished out forming dry joints. Second, in a thin layer, the fibers are forced to an orientation closer to 2D than 3D, if they are longer than the joint thickness. Third, mixing in the fibers forces you to more thoroughly mix the epoxy! Many do a poor job of mixing without knowing it.

Chopped fibers can help keep epoxy from running. That is often a benefit! But if the fibers are short they do not impart the strength one might hope, even ignoring the misalignment problem. The glass is much stronger in tension and compression than the epoxy matrix material. It is also much stronger than the adhesive bond. A fiber which is too short is not anchored well enough for it to break before it would simply break free of the matrix. That's sort of like a steel screw with too few threads engaged in aluminum - it rips out the threads.

Anyway I recommend testing epoxy with whatever additives one might use. See if it makes an improvement, and how much of an improvement. You might find that some other additive may work better, or a different brand of additive, or a different quantity, or a blend of additives, etc. I found that for many of my purposes, cotton flox worked as a better additive than chopped fiberglass. It was less brittle so it would hold threads better. YMMV - the specifics matter. Beware the dust though - I believe they call it brown lung disease.

Gerald

 

[G_T]

Just thought I'd bounce this thread back, since the sort of issues mentioned here come back around commonly enough.

Gerald