Planks from carbon tow.

The mechanical properties of a finished product depends on the fibre orientation, fibre quantity and fibre volume fraction. The cost is dependant on the manufacturers profit, resellers profit etc... so no correlation on cost can be made vs home made. There is some variation in fibre strengths depending on the type of carbon and the manufacturer, i know several grades are available - you would need to consult your supplier to get accurate figures for the individual fibre you will use.

I dont know how the mechanical properties are quoted on that website aircraftspruce. Some of their products appear to be quoted in an absolute manner for that product, which means its size, fibre quantity and Vf are calculated for each product and obvisouly depend on its size and shape etc...

The hexcel figures are absolute in a material sense, not a product sense. So when your using the hexcel figures, you do so in an applied manner for engineering. So you might use the figures to calculate a beam from the beams geometric size and shape ie.-its section modulus and inertia using the material properties for carbon fibre they quote provided you achieve a similar Vf in your manufacturing process of this beam. The total beam strength obvisouly depends on its total size and shape, which could be anything...

 

I'm not sure your answer pertains directly to my questions. I'll try to put some better pontification to them:

The stuff I have is Zoltek PX35 - tensile strength 4137 MPa, tensile modulus 221 GPa.
The Hexcel AS4C tow is 4485 MPa, 231 GPa.
Composite compressive strength is 1862 MPa.
T300 is 3530 MPa, 230 GPa.
Composite compressive strength is 1760 MPa.

So, why is the R-G pultruded T300 rod around 450 MPa compressive strength only?
And should I believe what the various Graphlite sellers (CST, Maerske Aircraft) claim, that my wet layup planks are only 300 MPa in compression?
If so, how do I get them as strong as in the Hexcel and Toray datasheets?
How do I test their strength, to within +-50% accuracy or better?

 

You need a materials testing rig to do it right. Graphlite product comes in at least two qualities that I know of. It tests to within a few percent of published figures so you may even be able to trim your safety factor even though most sources suggest a safety factor of 2 for composites.

To get into the prefabricated composite league all parameters are tightly controlled. Fiber orientation, and in the case of tows, even the slight waviness that can only be picked up using a magnifying glass is eliminated. In real world applications microwaviness has been found responsible for hand layed CF spar cap failures in DG300 sailplanes.

All considered it is often advantageous to utilize prefabs to replace hand layed up CF tows.

Dino

 

You would have to ask the manufacturer... this number doesnt seem correct btw... it could be a typo or god knows some other bizzare reason, or simply a poor manufacturing process, we are only speculating...

 

Yea. That's good. I'm wondering if I make a 50 mm tall beam with 3x3mm tow flanges, oval/diamond shaped corecell and some 50gsm bias glass around it.
Make a cantilever fixture, with some extra fabric at the attachments.
Then I might be able to break it with a known weight suspended from the end.
Then would the bottom flange crush, and would there be problems extrapolating the data to bigger beams?
The smaller I made it the less accurate it would be but it doesn't take a big carbon beam to be impractical to break in a controlled manner.

 

I am currently moving toward a similar adventure in carbon experimentation.

I want to build a wing (as discussed in "What is a significance of a wing thickness" and "Understanding Wing Technology") and feel that a spar is the best choice for the structural requirements.

I have done the rough math to build a spar, and think that it will be a truss structure. I am leaning toward something like a 10" square spar 18' long. The loads on the cross members on the cross members & the diagonals are pretty low. The big loads will be the axial loads on the long members at the corners. Using pultruded is probably the best choice if starting from scratch, but I have a spool of tow.

I want to do a small test build using carbon tow for most of it. The cross members are in compression but actually not heavily loaded. I will probably try to see if I can just use wood for for these items. The diagonals would actually be long continuous wraps around the whole thing.

 

That seems a cool construction method. I guess the truss could be lighter and much cheaper than webs/faces of a box beam. Do you need to twist the tow diagonals a bit to keep the fibers together?
What is the skin around the truss area? It is not strong enough to be shear web fore/aft?
Can you tell me how you calculated the spar?

Edit: For the next planks I think I will put nails at the ends of the mold and get some tension in the tow.

 

Tension on Tow seems to be the common suggestion for max strength in compression. One build instruction I read (carbon dragon, see link at the wing technology thread) had each tow tied off with a rubber band on one end to keep all of the fibers straight.


Keeping fibers together (or buckling) is not an issue in "tension only" members.


One of the leading proponents of Graphlite pultruded carbon in spar beams for home built aircraft wings is Jim Marske. I found a copy of part of his guide on the web and it was pretty useful.

However, regardless of the math on stress levels in the main load carrying material (the top and bottom of a beam in typical loading depictions), this material has to be tied to the material that holds it all together, keeps it stable and handles the shear loads. Unless you can find instructions for a beam that is engineered for home building (the only ones I found were for aircraft spars and I did not really like anything I saw), building a test section is probably better than guessing at all of the inputs for the calculations.

Truss design is neat in that you do not need any skin and load transmission between members is more straightforward. There is simplistic truss modelling software out there that you can play with.

For the truss spar I showed, the max cross member load (carried by 2 members) is equal to the overall max beam shear load. For a 45 degree tension diagonal, the max load would be 1.44 times the cross member load. Spreading the cross members farther apart increases the diagonal loading (simple vector geometry). A fraction of the tensile load in the diagonals also add to the compression load on the long corner members (a small fraction, but not negligible). Tow gives great tensile strength even for home build methods. Very little carbon is needed for the diagonals.

A truss may actually not be as optimal when compared to a well engineered I beam with the right selection of core material and the right fabric build up selection. I just see a truss as much easier to work up the right balance of material needed for each member and it should not require any carbon fabric (very high cost per pound of fiber).

For a truss: do a mock up, back out the stresses at failure (or determine the stresses that give max allowed deflection as your limit), scale up to the desired final size, add a safety factor and build away.

Since you are using bulk raw material that is not overpriced, you get easy logistics and less temptation to underbuild.

Again, If you really want better and are willing to buy material specifically for a project, bulk pultruded rod or rectangular sections purchased in coils is probably better for carrying the big loads.

Big concerns for any home build include

Keeping fibers straight if using tow as discussed above

Keeping the high compression loads from causing local buckling (fibers separating) or gross buckling (long thin sections that start to bend in the middle and collapse)

The angle sections I am thinking of were selected specifically to address the gross buckling issue in the most practical way I can think of for a home build with tow. They also work well for picking up the loads from the cross members. If needed, a thin layer of fiberglass cloth could be added on inside or outside surfaces to keep fibers together.

Even with good tension, wet layup tow probably probably needs twice the fiber as pultruded. This does not really hurt you for cost if you shop good (I got my stuff on sale, and I have seen good stuff dirt cheep on E-Bay when bought in full spools). Also, doubling the weight should not be a big deal given the strength to weight ratio of wet layup tow.

 

In a big tension and compression flange, you would normally put a +-45deg fabric of about 20% by weight, interleaved in the flange to keep the main 0 deg fibres in column.

Engineering a beam using the carbon properties is easily done using the section modulus of the intended beam geometry,and material properties of carbon.

Also, most beams are typically designed to handle a maximum deflection rather than ultimate strength. If its stiff enough, its usually plenty strong enough. So you need to first decide on an acceptable deflection, which depends on the application, then go ahead and design your beam to handle it.

A useful resource here -> http://www.engineersedge.com/calculators.htm

I recommend putting the required formulas into an excel spreadsheet so you can use it quickly and easily customized to your particular problems.

 

Marske has also published a small book detailing implementing Graphlite. His methods are novel especially those he uses to transfer loads to hard points (in this case wing pins) and practical methods of incorporating poltrusions in spar caps.

Dino

 

Good stuff, thanks. Do you have something for calculating torsion of fiber beams as well?

I don't think P. Flados has enough flange fibers to make a thick enough skin to resist local buckling. But if he had put his tows in the skins of 5mm corecell flanges, with some bias outside, instead of in the L's, do you think it could have broken by buckling/delaminating of the fibers?

 

The picture was just to show how the various parts would fit together. Actual dimensions and build schedule for the angle pieces are actually the really big complex item for what I would like to do.

First, column buckling is not that hard. It is just a matter of determining the inputs to the equations, and these are not bad. The link below is one of many good references. Note that buckling is a function of length squared. Close cross member spacing really helps.

http://http://www.egr.msu.edu/~harichan/classes/ce405/chap3.pdf

More important is trying to understand actual loading. Free standing rigs are simpler but need lots of strength for the high loads.

A stayed rig is much more complicated. I want the beam near the middle of the front wing element for a couple of reasons, but stays work better attached to the front. There are some choices, but none are clean.

I am also struggling with even more fundamental issues. I really have not settled on an exact final geometry, the max normal righting forces available and have not bounded dynamic loads. Think about sudden load increases due to wind gusts, loads associated with the boat being tossed about due to chop, impact load of rig hitting water, etc.

Max normal static loads are pretty easy, but I do not really have a feel for the rest. Truss design is super for being able to calculate material requirements once you know you loads and your strength values, but since I don't plan to build an instrumented prototype to get real stresses, a lot of the inputs to my calcs will be wild guesses.

Us engineering types really resent having to do business like this, but it is what it is.