Carbon Fiber Composites

This page will cover the application of composites in model airplanes and ways to utilize carbon fiber and fiberglass in model airplanes. This page consists of my personal experience with composites and is by no means definitive.

Common Mistakes

One of the most common mistakes that I have noticed when a carbon fiber or fiber glass part is designed is the tendency to fabricate it in the same shape as it's metal or previous counterpart. This is especially evident in landing gear where the gear is molded to the same dimensions as its metal counterpart. By making a composite landing gear flat one can speed up the production process but you loose out on rigidity and weight savings. When a composite landing gear is constructed with carbon or glass fabric bagged into a female mold resembling a gutter all of the fiber that is running the chord of the gear is adding weight and contributing no rigidity. Since the gear is flat, rigidity is also compromised. The solution to this problem is to make a carbon fiber gear that has a teardrop cross section with unidirectional tow running the length of the gear. By creating a teardrop cross section you increase the strength and rigidity of the gear while lowering it's weight. When the gear is molded with a teardrop cross section only one can be made at a time due to the mold configuration thereby compromising production speed and simplicity. Carbon cloth or Kevlar can be added to the leg of the gear where the axle is mounted and where it is mounted to the fuse to offer splitting resistance.

Ever see that iron on carbon tow? That spider web glue that melts when heated is only there to keep the fibers oriented and together while making a preform. A preform is made to fill a closed mold while the resin is injected or pulled through the mold by vacuum thereby wetting out the preform. The tow is not made to be ironed on without any other resin added.

Some thoughts on core materials. The use of core materials can be very advantageous. When core material is used to double the thickness of a structure, the relative stiffness increases 7 times. The strength increases 3.5 times while the weight only increases 1.03 times. When core material quadruples the thickness, the relative stiffness increases an incredible 37 times, the strength increases 9.2 times, but the weight only increases a mere 1.06 times.

When glassing the center section of a wing orient the fiberglass cloth so that both the warp and fill cross over the center section seam. Warp is the tow that is running lengthwise in the woven fabric and fill is the tow that runs across the roll of cloth. By orienting the cloth so that both the warp and fill cross the center section you are doubling the strength for the same weight. By using a low viscosity resin for glassing the center section of a wing and extending the resin a little past the fiber glass will create an edge to sand against while blending the glass into the wing.

Resin Overview

All epoxy is not created equal. There are literally thousands of combinations of resins and curing agents available to the fabricator. The fabricator is not only responsible for making the part but also creating the material that the part is made from. There are many resin manufactures that have proven resin and curing agent combinations available. A good understanding of the cured and uncured properties of these resins and curing agents will allow the fabricator to make the right choice when choosing the resin. The choice of resin is dictated by the parameters that you want the resin to operate in.

Bellcrank and mold.

Polyester is not a structural resin

Polyester is ok for car repair and non structural uses (tanks, bins, bathtubs) Lowest cost, highest production rates (very fast cure rates) Won't work on polystyrene foam (it softens the foam) Heavy, and hard to obtain proper fiber resin ratios due to wet out properties. Good secondary bonds are hard to obtain. Poor bond to Kevlar. Bondo and Featherfill are polyester-based.

Vinylester is similar to polyester but has better structural properties than polyester

Better bond to Kevlar and Carbon fibers than polyester. Gel time can be extended to 10 hours. Mid-cost, still less than epoxy.

Epoxy is the material of choice for structure

Cure cycles pace production work. Health hazards are present. Costly materials. Oven cure varieties are available to allow large lay-ups. Many varieties to match specific structural requirements.

General considerations for all resin systems

Health hazards. Workability. Cost. Chemical resistance (fuel proof). Compatibility with fibers and core materials. Required service temperature and moisture environment. Fabricator must understand the materials they are working with. Clean up issues and hazard waste.

Resin Matrices

Polyester

Cures by polymerization (long parallel molecular chains). Lowest cost resin. Unsuitable for structural lay-ups, low properties. Limited to low temperature applications. Insensitive to mix ratio (amount of catalyst affects cure rate not material strength) High shrinkage (unstable parts/tools, cloth print through) Polyester part will not bond to a epoxy part. Contains styrene; therefore cannot apply over polystyrene foam.

Vinylester

Improved version of polyester resin, better properties, higher cost. Less health risk than epoxy. properties are between polyester and epoxy. Extended pot life to allow larger lay-ups

Epoxy

Most common structural resin, many different varieties available. Cures by cross linking (three dimensional process). Very sensitive to proper mix ratio of resin to hardener. In the small batches that are used in modeling a Triple beam gram scale is needed. Highest cost. Room temperature or oven cure variants available. Absorbs moisture (hydroscopic). Oven cure variants have higher Tg and Heat Distortion Temperatures. Will bond to a polyester part. Multiple health issues. Lowest shrinkage (highest stability). Excellent adhesive properties (good secondary bonds). Face coats available for tooling surface finishing. Laminating and tooling resins available.

Epoxy Resin Specifications

	Mix Ratio: by weight (most accurate) or by volume. Typical ranges 100:44 to as low as 100:5 by weight. When mixing small quantities for model airplanes a triple beam gram scale that measures in 1/10 of a gram is required.
	Mixed viscosity, low viscosities required for laminating resins to ensure proper wet out, high viscosity for tools. A good viscosity for a hand lay-up is around 500 to 800 centipoise.
	Pot Life, 100 gram mass; larger masses cure faster, pot life characteristics limit the maximum laminate thickness possible per cure cycle. Oven cure resins have extended pot lives, allowing larger lay-ups to be accomplished. Most resins for a propeller or similar size part would require only 20 to 25 grams of resin. Mixing more than you need decreases the pot life and wastes resin.
	Pot Life for thin film: More representative of time available to wet out laminate and bag if required. 100 gram mass pot life is representative of a small batch mix according to the resin manufactures. Actual pot lives are even less then specified when thixotropics are used. Thixotropics include micro balloons, flox, Cab-O-Sil and chopped carbon fibers.
	Gel Time: Similar to pot life; resin is too thick to wet out fibers once gelled.
	Cured Hardness, Shore D: Cured resin can be hardness tested to assure full cure. When making a part it is important to save the left over epoxy in order to test the cure.
	Glass Transition Temperature (Tg): Maximum temperature at which resin properties diminish appreciably, sometimes referred to the resins "red line" temperature. When a cured polymer is heated, vast changes in thermal and mechanical properties occur. These changes are particularly large near the glass transition temperature, Tg. Below the Tg, the polymer is hard and glassy, and above the Tg it has a rubbery state. At this temperature, tensile strength, hardness, electrical properties and chemical resistance depreciate rapidly, while tensile elongation and flexibility increase markedly. Tg usually occurs over a range of temperature, but for simplicity a single temperature is selected as Tg.
	Heat Deflection Temperature (HDT): Temperature at which the resin begins to soften but still has good structural properties. The deflection temperature is commonly used as approximation of Tg. The method for measuring DT has been standardized by ASTM. The DT is determined on a casting which has been permanently stressed at (264 psi) by flexural loading and then heated at a constant rate until the casting deforms a specified amount. The DT method usually requires a larger sample than Tg methods.
	DT's and Tg's provide a measure of crosslink density of the polymer. Those polymers with higher DT's have higher crosslink densities, better performance at elevated temperatures and generally better solvent and chemical resistance. The choice of curing agent and the cure cycle (degree of cure of the polymer) are the largest factors affecting DT. You would want a higher Tg resin on a tuned pipe than on a wheel pant.
	Notch Sensitivity (Izod Impact): A measure of the resin's brittleness. A water ski would require a resin that is a little more flexible than a model airplane propeller blade. A plasticizer can be added to make the resin tougher and less prone to fracture.
	Post Cure: The manufacture's recommended elevated temperature cure cycle to be used to attain the best material properties. Post cures either follow a room temperature cure or an intermediate temperature oven cure for oven cure materials. Free standing post cures are typically successful if a gradual ramp up in temperature is used. High-temp assembly fixtures are required if a free-standing post cure cannot be accomplished.
	Peak exotherm, Fahrenheit: An indication of a resin's likelihood to exotherm uncontrollably. The chance of exotherm can be reduced by limiting mix batches to small quantities, proper disposal of leftover resin, and knowing your resins properties (testing). Exotherm is a term used to describe the internal heat generated by the cross linking of the resin to the hardener. On some resin hardener combinations a 50 gram mass is great enough to melt a plastic cup and become hot enough to burn your skin. Larger quantities create a fire hazard.
	Resin "Physicals" Include: Density, Hardness, Viscosity, Elongation, CTE or coefficient of thermal expansion, Tg, HDT, Pot life, Mix Ratio, Color, Peak Exotherm, Shrinkage, Izod Impact and others.

Back to Top

Materials

Carbon Fibers

Carbon fibers, Though known since Thomas Edison's development of the incandescent light in the 1870s, were not made in large quantities until the late 1960s. At that time it was found that carbonizing several fibrous materials resulted in a continuous fiber with relatively low density and high Young's modulus of elasticity. Modulus of elasticity is a parameter indicating a material's stiffness. Young must have been the one who came up with a mathematical way to measure this. High modulus materials are stiffer than low modulus materials.

Fiber Sizing: Sizing are added to fiberglass and carbon fiber to aid in processing and to allow the resin a better bond to the fiber. Silane coupling agents are a used as adhesion promoters. While the sizing helps in the processing of fibers they can hinder wet out of the fabric or tow. Carbon fiber sizing must be applied to the fiber tow (which may consist of 12,000 filaments of more) to prevent the individual filaments from contact damage between themselves or with eyelets or guides during weaving or prepreging. When using tow to manufacture propellers and bellcranks it was necessary to massage the tow in order to loosen the sizing for better wet out. The thing to remember when ordering carbon fiber and fiber glass is that the sizing is compatible with the type of resin you are using. Some sizings in the fiber glass industry are for polyester only but most sizings are for both epoxy and polyester resins. One can obtain a discount on fiberglass with expired sizing when ordering from the mills in large quantities.

If you have ever handled carbon tow in the raw state without sizing it is very soft. Fabricators rely on its tensile strength for the parts rigidity. It is of great importance to keep the fibers in line with the loads being applied to it for the best strength properties. Even the crimp in a carbon tow that allows the tow to be woven into a cloth weakens the material. This is why unidirectional carbon lay-ups tend to be the strongest. While talking to Hiroshi Kiyomoto who fabricated Kaz Minato's carbon wing of his latest Blue Max at last years Nationals he stated that he used unidirectional carbon in the wing. One layer was from root to tip and the other was from the trailing edge to the leading edge. By doing this he could obtain a stronger lighter wing than would have been possible with a carbon cloth. Carbon cloth holds excess resin in between the tows if not compacted with a vacuum bag. Carbon cloth greatly speeds up the production process though.

While carbon fiber is considered a light weight material in the full size world, it still remains a great challenge to save weight on a model airplane with this material. If you are trying to replace 4 to 6 pound per cubic foot balsa with carbon fiber the results are usually disappointing. The tools to achieve the proper compaction are out of the hobbyist's range. These tools are autoclaves and composite ovens. It's still hard to beat a good piece of 4 1/2 pound balsa for formers, ribs and the like. Carbon fiber excels on the materials normally constructed from hardwoods, metal, plastic etc. or high stress areas such as spars and the like.

One of the biggest obstacles to overcome when designing a part from composite materials is the natural tendency to copy a part exactly as if it were made from its previous material. The part should be redesigned to take advantage of the formability and strengths of the composite material. A composite part should not be limited to the shape of a metal or plastic part.

There are three types of carbon fiber, Rayon, Polyacrylonitrile (PAN) and Pitch.

Rayon precursors, which are derived from cellulosic materials, were one of the earliest precursors used to make carbon fibers. The processing disadvantage was a high weight loss, or low conversion yield to carbon fiber. Typically only 25% of the initial fiber mass remains after carbonization, which means that carbon fiber made from these materials is comparatively more expensive than carbon fibers made from other materials.

Polyacrylonitrile (PAN) precursors are the basis for the majority of carbon fibers commercially available today. They provide a carbon fiber conversion yield that is 50 to 55%. These precursors can be thermally rearranged before thermal decomposition, which allows them to be oxidized and stabilized before the carbon fiber conversion process, while maintaining the same filamentary configuration. The chemical composition of PAN precursors defines the thermal characteristics that the material displays throughout the oxidation/stabilization portion of the conversion process. These thermal characteristics influence the processing sequences that are used to convert PAN precursors to carbon fiber. Carbon fiber based on a PAN precursor generally has a higher tensile strength than a fiber based on any other precursor. This is due to a lack of surface defects, which act as stress concentrators and, hence, reduce tensile strength. Carbon propellers and bellcranks that are made by Winship Models utilize PAN based carbon fiber.

Pitch precursors based on petroleum asphalt, coal tar, and polyvinyl chloride can also be used to produce carbon fiber. Pitches are relatively low in cost and high in carbon yield. Their most significant drawback is nonuniformity from batch to batch.

Glass Fibers

There are literally thousands of fiber glass fabrics and tows available to the fabricator. For modeling use the varieties of fabrics under 2 ounces per square yard are the most common.

E- Glass is the most common (and least expensive) grade of glass fiber.

S -Glass is a special grade of glass that is much stiffer than E-glass and somewhat stronger.

The use of fiber glass on a stunt ship is usually limited to the center section of a foam wing or around the nose section, either on the outside or for reinforcement around the nose formers. If using fiber glass to strengthen the center section of a foam wing keep the glass fibers running at a 45 degree angle to the center wing joint. This will allow both the warp and the fill fibers of the cloth to span the wing joint giving double the strength with no weight penalty. The warp direction is the direction of the long fibers as the cloth is pulled off of the roll. The fill fibers are the fibers that run side to side as the cloth is pulled off of the roll.

Almost all glass cloth has sizing applied to it to aid in resin wet out and adhesion. Some sizings are for polyester resins and some are for epoxy resins. If the glass cloth or tow has the wrong sizing for the type of resin used, then the bond will be weak between the resin and cloth or tow. While this might not be of great concern on a pair of wheel pants it might prove to a problem if a speed prop was constructed without the proper sizing applied to the tow. Usually a fiber glass that is not properly wet out will have silver of white streaks or spots in the laminate.

I decided to make a carbon fiber landing gear for the Midwest Extra 300 to save a little weight. The carbon fiber landing gear should cut the gear weight in half over the aluminum component.

I am using the original gear for the mold. This gear is a one off so if more than one gear was to be made a matching face mold would be constructed. The landing gear is mounted to foam blocks for easy layup. Along the edge of the aluminum there is a balsa dam that is 1/4" above the gear. The height of the balsa edge above the gear will determine the thickness of the gear. This balsa edge will hold the carbon fiber on top of the gear while the layup is in progress also. I will attempt to sand in some teardrop shape to make the gear more aerodynamic after demolding. Clear packing tape was placed over the predrilled holes to keep the resin and carbon from running through them. A Freekote 700 mold release was also used to assure the carbon gear would release from the aluminum gear.

The carbon tow has been cut to length and is now waiting for the resin. I am not sure how many filaments there are since it is not important due to the fact that every gear will require a different amount of carbon fiber. I guessed on the amount by placing the carbon tow in the mold to see how much it would take to fill it. I guessed pretty close since didn't need to cut additional carbon. The tow was cut a few inches oversize and was trimmed where it hangs over the end of the gear. The carbon tow trimming was them placed in the center of the gear

The carbon tow was wet out on wax paper before placing in the mold. With this much tow and resin it is a good idea to work fairly fast since the pot life of the resin is dependant not only on room temperature but the mass of resin. If the resin is spread out onto the tow you will avoid an exotherm and increase the working time. The mold was filled to the top or a little over and will be sanded down to the proper thickness after curing. The balsa sides will be sanded off after the gear is sanded or ground down to the height of the balsa. A mid temperature resin was used that has a heat deflection temperature of 260 deg. F. after post cure. A high temp or mid temp resin is mandatory for a landing gear in my opinion since the heat radiating off of blacktop in the summer can reach as high as 120 deg F. in Indiana. The high temp resins also have greater structural properties. There are many landing gears that are constructed by laminating layers of carbon cloth over or in a mold but the carbon fiber that is running chord wise in the mold is not providing any rigidity and adding unnecessary weight. If you desire greater strength at the wheel/axle attachment point you can bond on a piece of carbon fiber woven plate for splitting resistance. Unidirectional carbon fiber tow is the strongest and lightest way to make a landing gear and since the original landing gear can be used as the mold it makes it affordable for the person who would like a starter carbon fiber project. If for some reason you still want to use the original landing gear you will be able to after removing the carbon part provided you do not forget the mold release.

Also in the pipeline is a plug tutorial for a pattern plane.

The carbon gear has finally been released from the aluminum landing gear. Some of the balsa side structure is still stuck to the aluminum gear but can be removed by sanding.

The carbon fiber gear in the picture has been sanded to a teardrop shape and is constructed entirely from carbon tow. I feel that a secondary bond of some carbon fiber woven plates at the axle attachment would be beneficial.

The aluminum landing gear weighs 347g or 12.23 ounces and the carbon gear weighs 229g or 8.07 ounces for a total weight savings of 4.16 ounces. The carbon fiber gear at this weight is stiffer than the aluminum gear and could be made lighter. Since the carbon gear was formed over the aluminum gear it is slightly taller and wider which I feel is beneficial. If these composite landing gear were going to be mass produced a matched mold would have to be constructed to form the cross section of the gear.

Since I was only going to make one of these landing gears I used the TLAR (That Looks About Right) Engineering process. After making numerous parts from carbon tow I had a good idea of the amount of tow it would take to fill the mold and was able to pre cut the right amount. If making these landing gears from a matched mold one would need to keep track of the amount of resin and tow in order to completely fill the mold. There are production processes available such as VARTM (Vacuum Assisted Resin Transfer Molding) that would be suitable to producing parts with this geometry economically. VARTM processes will be covered in the future although the process isn't cost effective for only one or two parts.

Thumbnail of the bottom of the Extra 300 carbon fiber landing gear.

Thumbnail of a carbon fiber control line landing gear mold. The mold is a matched die type mold and is the type of mold used when you want to produce a carbon gear that has a streamlined cross section.

When I started to mount the kill switch on the Spacewalker I realized that I needed a right angle bracket to mount the switch. The bracket material was fabricated from six layers of carbon cloth with four layers of glass acting as a core. Since the bracket wasn't going to carry a load the fiber orientation wasn't critical. The bracket could have been made of wood but the light weight and compact nature of the carbon fiber bracket was appealing.

The picture shows the aluminum angle that was used as a mold. I sanded the aluminum to remove the dirt and lightly polished the surface before applying the Freekote mold release. The laminate in the picture has been trimmed and the actual switch mount has been cut away from the part shown. Since making these brackets are relativity easy I made a piece large enough to cut multiple brackets from. After the part was laid up an oven cure of 150 degrees F for four hours the part was demolded and trimmed. A 250 degree F resin was used but since the part wasn't going to see that high of a temperature a 150 degree F was adequate. C clamps provided the pressure to debulk the laminate.

Pictured here is a vacuum pump sold by Aerospace Composites. The pump will pull about 18Hg. This pump is an excellent vacuum source for bagging balsa to foam since the vacuum can be adjusted. When bagging expanded bead polystyrene you will need to stay under 7Hg to keep from crushing or distorting the core. The other advantage of this pump is the vacuum

THE BIG SUCKER

This is a composite de-bulking vacuum system that was built similar to the one shown above except that it is an industrial version. The Gast pump is rated a 1/3 hp and will pull 28Hg. This composite vacuum pump is well suited for vacuum bag mlding. The two PVC schedule 40 pipes provide the vacuum tanks and are inserted into a wooden box that acts as a motor mount and tank holder for the Gast pump which sits on four Lord Mounts. The black project box was purchased from Radio Shack to hold the relay, switch and fuse. The vacuum switch, relay, vacuum gauge, check valve and miscellaneous fittings were purchased form Aerospace Composite Products. George Spar sells a new vacuum switch that is adjustable over the full range of vacuum from 5Hg to 28Hg. The vacuum switch and 10 amp relay are a critical components for this 1/3 horse setup. The switch purchased from Aerospace Composites provides a smaller amount of deadband than other switches that I have tried and is highly recommended. To mount the vacuum switch, gauge and fittings, the PVC pipe was drilled and tapped with NPT taps to match the fittings while a liberal amount of RTV silicone seals the threads. A special thanks to Steve Ragsdale for his input and help on this pump (he is and electrician by trade). To view the schematic of this pump click here. The schematic shows relief holes between the pump and check valve which are not required on the Gast Rotary Vane Pump since the vacuum isn't present in this section of hose when the pump isn't running. The capacity of this pump will debulk large composite lay-ups. Future projects for this pump are carbon honeycomb laminates for fuselages and oven cured composite lay-ups.