Unidirectional Carbon Tube as Control System Rod

[wantobe]

Some long control system rod sizes are limited by Euler column stability which is directly related to the stiffness of the rod. Unidirectional carbon tube is a good fit for this application because of its anisotropy and high stiffness. This gives it the potential to reduce control system weight and complexity. Have anyone seen such an application? What are the you-need-to-know caveats in design and construction?

[hogheadv2]

Diameter is a factor over thickness of the tubing. I would not recommend "Rod". The carbon rod I have seen, (extruded) was stiff bit also flexed with compression forces. A carbon arrow tube 1/4 the weight was far more stiff and rigid. I understand this is not control actuator size but my feeling is the same would be true. *** I have considered making and testing custom carbon tubing using a pipe or bar as a mandrel tape wax paper the length needed wrap it and the wrap with the carbon / epoxy (pre-wet) squeeze out wrap the pea-lply and vac. [this may be much easier with a few helping hands. /// Hoping compressed air would release the tube and non-destructive and destructive testing with flex and strength would begin. [ work = money = weight ] Is it worth it,,, or just go without the French fries .... :eek:

[nrpetersen]

For short carbon fiber control rods there might be x-country ski poles. I don't know if they are pure pultrusions though.

[Victor Bravo]

Composite Structures Technology, and Aerospace Composite Products are two companies selling pre-made carbon tubes of both the extruded and wrapped types. They have diameters all the way up to a few inches if memory serves. They serve the large R/C aircraft and UAV markets primarily, with experimental aircraft in themix to a smaller degree.

The problem with using a carbon rod is not stiffness or strength, it shines pretty well in both. But once you exceed the limits the carbon tends to shatter or collapse all at once. This may not be a desirable characteristic for a control pushrod only because of what's at stake if it does get over-stressed.

A carbon and Kevlar, or carbon and fiberglass tube might offer a significant safety factor. for very little extra weight.

Another consideration is the end fittings. They are traditionally put in and riveted in two places. This works great in the metal tubes... But composite structures do not like rivets or fastener loads. A rivet will wear a composite material. You may wind up with a brilliantly light and stiff pushrod, and a weak or short-lived attachment to the end fiting.

[wantobe]

Thanks for the replies. Very good information.

I think all carbon structures have the same problem of brittleness. So it is a matter of margin and proper usage. The end fittings are a bit tricky. In an ideal situation we would like to have the metal fittings built in a composite tube as it is manufactured. This is impractical if we need to modify a length of a tube from a catalog. Anyone with good experience in attaching a metal end fitting to a (like 1.5" OD, 0.065" wall thickness) carbon tube? What are the tricks?

Is this a common practice to use composite tube in a primary control system of an experimental or certified airplane?

[hogheadv2]

Bicycle have been most successful while taking the harshest beating, that's were I would research attachments.

[Victor Bravo]

This response assumes you have not already built with composites. This response assumes you are not an engineer, and are asking for tips and explanations on a "beginner" or "intermediate" level. I'm not an engineer myself but I do have a fair understanding of this subject matter.

Epoxy will adhere to steel reasonably well, very poorly to bare aluminum. Steel threads (for turnbuckles, spherical joints, control horn or bellcrank attachment, etc.) are of course stronger than aluminum, so these two factors make a steel end fitting the clear choice. IMHO do not try to save weight using Aluminum or Titanium fittings.

Make sure the end fittings have several deep grooves turned into the shank that slides into the composite tube. I would guess the shank length needs to be 2 times the tube diameter. If the shank on your fitting is a thick wall tube, make sure your wall thickness is enough so that the grooves do not over-weaken the wall. Sandblast or "wire wheel" the grooved area immediately before applying the epoxy to it.

Some carbon tubes may have a mold release or lubricant on them, to make the material not stick to the manufacturing tools. So you should go in there with a small Dremel tool sanding drum or sandpaper wrapped around a dowel, and rough up the inside of the tube where the glue will be. There should be a couple of layers of bidirectional cloth, preferably Kevlar or fiberglass. wrapped around the last 3 inches of the carbon tube where the rivets will go. This is "fail-safe" protection against lengthwise "splitting" of a pulltruded unidirectional fiber tube.

Use a slow setting epoxy like you would use to laminate fiberglass structure... meaning do not use 5 minute or 15 minute epoxy. The slower setting glues are usually the stronger ones.

DO NOT DO NOT DO NOT use "micro balloons" or "glass bead" as the thickening agent for this application. A control rod fitting glue bond is loaded 100% in shear, and you need to INCREASE the shear strength of the joint. "Micro" will drastically reduce the shear strength of the epoxy, and create a very dangerous situation. You MUST use "cotton flock" or "chopped fiber", which greatly increase the strength of the bond.

Paint a thin layer of your un-thickened laminating epoxy onto the inside roughened surface of the tube first. Then mix a little cotton flock into the epoxy until you get an applesauce or "wet cottage cheese" consistency. Make sure it is mixed thoroughly of course. Prep the fitting as above, then apply the thickened epoxy to the fitting and grooves with no air trapped in the grooves, then insert your fitting into the tube, then remove all excess drips and glops.

After the epoxy has cured fully, then drill two #30 holes through the tube and the fitting, at perpendicular locations. Make sure the "edge distance" to the edge of the fitting is 2 times the rivet diameter. Put AN470AD4 rivets through these holes and squeeze the rivets per the instructions in AC 43.13.

Paint a stripe of "torque seal" or bright lacquer paint across the rivets, and the edge of the steel/carbon contact point, so that 20 years from now you can go in there with a camera or inspection mirror, and see whether there has been any relative movement between the fitting and the rod.

Pull-test the control rod assembly to a load of 200 pounds, to "prove out" the integrity of the joint. If the lacquer is cracked, you have to find out why, because with the epoxy as the primary bond and the rivets as a fail-safe, you should have had a completely solid and movement-free joint.

If the lacquer has not cracked, you now have a control rod assembly that you can bet your life on. Until ten minutes after the first flight, and every 200 hours or annual inspection thereafter.

I am fairly sure this admittedly complex method will keep you at a high level of safety, and is well worthwhile for a primary control circuit.

[Bob H]

Just came across thread and I'd like to add a few points regarding composite tubes and fitting assemblies.
If you are considering an all uni fiber tube, it would be smart to add at least one ply of woven matrial at +-45 deg to provide some shear/torsional strength. Pure uni layup has little torsional stiffness and if rod fittings were to bind up in service and put a torsional moment in tube, these plies would help withstand it. It's mostly a safety item.

You don't want to use aluminum in any dierct contact with carbon as alum will galvanically react and corrode from exposure to atmospheric moisture. The easiest way to avoid this is to apply a thin layer of glass between carbon and alum as an insulator. Even steel will react with carbon given sufficient time.

Structural epoxy adhesives have fillers added to make them thixotropic so they won't slump and will stay in a joint without flowing out. This maintains the bondline thickness which is generally best at .010". Too thick a bondline at ~.050" or too thin at ~.002" reduces shear strength.

Don't use alum rivets in a carbon structure because of the galvanic issue; rivets will corrode away. Better to use SS (A-286) or titanium if available/affordable.

The loading enemy of an adhesive bond is peel. Bonds are intended to react in shear. Any applicaton of bending on the bondline that produces a peeling moment may fail the adhesive premayurely. Failure of a bondline usually occurs as a combination of adhesive and cohesive rupture. Adhesive is the separation of the adhesive from the bond surface. Cohesive failure is the adhesive breaking within itself. You would like failure modes to be cohesive which tells you that the surface prep was very good. Speaking about surface prep, it's the most important factor in a strong bond. The best all around prep is a light sandblasting of mating surfaces to provide a roughened surface. When done on carbon parts, you have to be careful not to blast away the surface plies.

Sorry if I got carried away with comments; spent many years in the business.

Bob H.

[wantobe]

I am an engineer, but not a composite engineer. Your reply is exactly what I need. Many many thanks!



Pull-test the control rod assembly to a load of 200 pounds, to "prove out" the integrity of the joint.

I have two comments on the above statement. First, control rod usually is limited by buckling, so push-test is more critical than pull-test. Second, using a fixed 200 pounds for all control rod assembly may not be appropriate. The maximum pilot force for each control (rudder, aileron, elevator) is different and specified in Part 23.397. Then you need to multiply the lever ratios along the chain of cotrol system to figure out the maximum force on the specific rod. This gives you the limit load on that rod. Now multiply it with safety factor 1.5 to get the ultimate load. Push your rod with this ultimate load to see whether it buckles. Then pull it with the same ultimate load.

[Bob H]

Composite parts are usually better in tension than compression because fibers naturally can resist loads in tension, like a rope. In compression, the resin plays a major role in keeping the fibers from buckling. When evaluating a resin, you use a compression test as a primary screening criteria. The degree of efficiency of the resin is dependent on the morphology of the fiber surface and on the sizing applied by fiber mfgr to maximize the adherence of the resin to fiber. This is suposed to be optimized for a given resin chemistry, with epoxies being the most popular. There are other resin systems used for structural parts normally based on operating temps with polyimides maxing out around 600F.

If you were building rods to only see compression buckling, you would pick a higher (intermediate, 40 MSI) modulus fiber than the standard modulus (30 MSI). This gives inherent incease in buckling capacity for a given tube geometry and the fibers cost more. For special structures like space telescopes, where high stiffness is really needed, you can use fibers with 60 MSI or even higher. With the increased modulus comes higher price and lower strain to failure so the parts can't tolerate abuse or much impact loading.

As someone mentioned earlier, the limiting factor in composites is the absence of plasitcity which translates into damage tolerance, the ability of a part to withstand impact and still carry a high percentage of load.
So designers use the linear ultimate strength/strain with knockdown factors to keep the part stresses well below the ultimate.

Yeah, you really didn't care about all that.
Bob H

[wantobe]

Bob, I do care about all of these! Thanks!

I do not think that a contol system has significant impact loads. Anyone has specific number on this?

I think some considerations are similar to those in determining design allowables for composite structures, namely, the effect of temperature, moisture and impact dammage. It would be wise to check that neither maximum tension nor compression strains of the tube exceeds the design allowables. I expect these tubes to be buckling limited, not strain limited.

[Frank Giger]

I don't know anything about composites, but if you want to know about controls and impact forces you should come watch me do touch and goes in a Champ.

;)

[tonycondon]

It looks to me like Dick Butler used carbon rods for the controls on Concordia. One of the pictures at least showing the final control setup appears to have a metal rod end bonded and riveted to the carbon push rod, similar to as described earlier. It also shows how they set up a load test on the system to ensure the strength of the system.

http://soaringcafe.com/2011/04/building-concordia-the-wing-control-system/

[58boner]

OK just jumping in here with a bucket of cold water. Every composite structural or control surface part on certified aircraft I inspect on a regular basis (Gulfstream) has a mandatory periodic inspection for DE-lamination. Would you include such an inspection program for your push pull tubes?

[Bob H]

You would do a visual check periodically on composite tubes, much like you do on metal parts looking for corrosion or deformation. Delaminations on large area parts like a control surface are initiated by interply separation from resin matrix breakdown usually due to an external impact force or very high cycle fatigue forces. Normal commercial aircraft inspection techniques use ultrasonic interrogation to map delams or disbonds, followed by a judgement as to repairs/replacement. On a homebuilt plane with composite tubes, the primary loading is axial tension/compression with buckling as critical mode of potential failure. If tube is designed correctly, I would not expect normal operations to produce delams and a visual check during annual inspections should be ok.

[Thomas Stute]

I just dropped into this thread and would like to add some comments based on my composite engineering experience from the last 30 years in aerospace structures.

In general a major consideration should be given to the diameter of the push/pull tube. This defines stiffness and buckling resistence. The larger the diameter the better the tube can withstand buckling. However, pushrods in an aircraft control system usualy are not exposed to ery high loads, so this problem should easy be solved by a simple calculation.

In order to stay on the safe side the amateur or homebuilder with no experience in composites should stay with standard carbon fibers like the Toray T300 or similar since the higher the Youngs-Modulus of the fiber the lower is the allowable elongation until rupture and the more critical it is to process.

Commercially pultruded tubes/rods have a mainly unidirectional fibre orientation and hence are critical when torsional loads would be applied, e.g. during screw tightening at integration steps. To be safe you should design your tubes with some +/- 45° layers. Braided carbon sleeves come in handy when opting for angles different from the tube central axis.

When you want to manufacture the tubes by yourself use an aluminium mandrel and cure the tube at elevated temperature (e.g. 60°C or whatever temperature is best suited for your resin system). After curing and cool down to room temperature the aluminium mandrel will shrink and will enable de-moulding. The fibre volume content and the consolidation of the laminate can easily be achieved by finally wrapping the lay-up with a band of peelply under tension. Surplus resin will be forcd out and with it air and voids.

Tube end fittings is a theme: This are the points wher the loads are introduced and they need to be designed with sufficient care. The easiest way is to design the fittings with the inner diameter of the tube and to foresee a circumferential glue pocket for bonding layer thickness of 0.1mm to 0.2mm. A major problem can be the material of end fitting due to the CTE mismatch between CFRP and the end fitting. Avoid aluminium (CTE = 23.5 E-6 1/K) and take stainless steel (CTE = 13 E-6 1/K).

When selecting the adhesive for bonding the fitting into the tube don't use a mixture of epoxy resin with some additives for making it tixotropic but take a certified and proven adhesive like HYSOL EA 9321 or Scotchweld 9323 or something similar. These adhesives are designed especially for structural bonding and have some additives to improve their toughness. Epoxy resins are too brittel for critical bondings like in the aircraft control system. For calculating the necessary bonding area don't use the nominal lap shear value given by adhesive manufacturer (usually in the range of 20 MPa to 30 MPa) but use a reduced value to account for irregularities like poor bonding surface treatment, etc.

Don't rely on a single load pass (bonding) but use an additional one, rivetting. Rivets should be used only after the adhesive has completely cured. In order to avoid corrosion don't use aluminium rivets or steel, but titanium rivets. Apply some glue when inserting the rivets.

Once ready you should expose your tubes to proof loading for being sure your workmanship is fine. A good load is the design load. Don't overtest, since you want to have the inbuild margin for emergency cases.

The are just a few aspects to avoid the major mistakes.

[wantobe]

The depth of this forum always surprises me!


In order to stay on the safe side the amateur or homebuilder with no experience in composites should stay with standard carbon fibers like the Toray T300 or similar since the higher the Youngs-Modulus of the fiber the lower is the allowable elongation until rupture and the more critical it is to process.

This is kind of a compromise between process and performance. Young's Modulus is what really matters in buckling resistance, yet we can not use higher ones due to process sensitivity.



Don't rely on a single load pass (bonding) but use an additional one, rivetting. Rivets should be used only after the adhesive has completely cured. In order to avoid corrosion don't use aluminium rivets or steel, but titanium rivets. Apply some glue when inserting the rivets.


Where can titanium rivets be found? I can not find it on aircraft spruce website by searching it on that website.

[hogheadv2]

Diameter is a factor over thickness of the tubing. I would not recommend "Rod". The carbon rod I have seen, (extruded) was stiff bit also flexed with compression forces. A carbon arrow tube 1/4 the weight was far more stiff and rigid. I understand this is not control actuator size but my feeling is the same would be true. *** I have considered making and testing custom carbon tubing using a pipe or bar as a mandrel tape wax paper the length needed wrap it and the wrap with the carbon / epoxy (pre-wet) squeeze out wrap the pea-lply and vac. [this may be much easier with a few helping hands. /// Hoping compressed air would release the tube and non-destructive and destructive testing with flex and strength would begin. [ work = money = weight ] Is it worth it,,, or just go without the French fries .... :eek:

I stand to correct myself, Looking further I have found CF tubing of a far greater strength and quality than any I've seen before. Not that I can endorse the use of these. http://www.cstsales.com/Carbon_Fiber_Tubes-ss3.html I could see these used as control rods.... With proper testing for application.

[Bob H]

There are appications where composites make sense and some that don't, especially if you are homebuilding and have limited equipment resources. If you are playing around, have money to spend and want to practice some composite fabrication techniques, then making tubes is a learning exercise and the first few may not turn out just right. When measuring the overall weight savings, you need to include the end fittings which usually makes up over half of the strut weight. Assuming rod end bearings that have to be attached via an insert bonded into the tube end, the tube dia may have to be increased just to accommodate the fitting. And you would want the fittings to be removeable to change rod ends should they wear. By the time all is done, it's easier to use an aluminum tube off the shelf, slide it into a female rod end and fasten.

Back in the late 1970s, I made parts for the Gossamer Condor, the human powered vehicle that went across the English Channel. These guys were extremely weight consicious where saving a fraction of an ounce was a big deal. I redesigned the pedaling hub out of composites to reduce the weight by 2 oz and thought it was not worth the expense but McReady's guys thought it was a major savings. For most engineering weight reduction choices, the cost/lb has to enter the decision process and sometimes forcing a composite solution just isn't worth the effort.

[Thomas Stute]

I just don't know where in the US you can find Ti rivets. I think you have to check in the internet. Manufacturers are Hi-Lock, Shure-lock, Huck and some others. For my needs I would get a few of them directly from my company. A good source could also be suppliers for aerospace industry. There should be many in the US.

[Bob H]

Use monel rivets, not titanium. You can't easily upset Ti fasteners.
Monel is nickel based and does not corrode in carbon. It is also hard to upset; work hardens.
What are you trying to do that calls for such rivets?
Bob H

[Bob H]

Just one more point on riveting-
Don't use solid rivets in any carbon fiber part because the radial expansion of shank from upsetting will damage hole surface and reduce bearing strength. And don't use alum in direct contact with carbon as it will corrode; must be galvanically isolated. Fasteners in carbon must be clearance fit and non-reactive.

[David J. Gall]

....now multiply it with safety factor 1.5 to get the ultimate load....

I read somewhere that composite aircraft engineering generally calls for a minimum safety factor of 2.0.

[Bob H]

Usual factor is 1.5 applied to loads, whether it's composite or metal, and the allowables on the material properties dictates part thickness/geometry. Composites are designed to a strain level rather than a stress level.
If margin is too high, plane won't fly because of increased weight. Some spacecraft are designed to lower factors, 1.1, to reduce weight for launch if loads after launch are minimal. Expendable rockets can use 1.25 as factor to decrease their weight. But the lower the factor, the less tolerance for operational errors.

[wantobe]

Icon A5 uses carbon composite control rod. This is what they displayed at AirVenture.3174