Saturday, May 31, 2008

Carbon Carbon Everywhere

Look around, carbon fiber is everywhere. For only $269 you can get a decorative CF panel to stick on the pillar between the front and rear doors of your Scion! Yep it seems to be ubiquitous.

Only a year ago, prognosticators in the composites industry were predicting a major CF shortage. Looking around at suppliers, not all of them have all products in stock. That said, it's easy to find the materials I use in frame building. So that's a good thing.

A little known fact outside of the industry is that carbon fiber tubes cost about the same as high end steel tubes, such as from Columbus or Reynolds. True that. Oh and yeah, I said tubes. In fact I source my tubes from the same place as Trek. Did you know that Trek and many other bike manufacturers assemble their frames from tubes? For the consumer, that's not an important issue. But marketers have done a good job of selling the idea that CF frames are built of a a single carbon fiber monocoque - not assembled from tubes. So consumers don't like the idea of joined tubes and manufacturers don't talk about using CF tubes.

Well I'm not afraid to admit to working with tubes. The fact is, filament wound tubes can be manufactured to tighter specifications than a complex molded part. Which means that tubes offer the opportunity to build stronger and lighter! Yipee!

How these tubes are joined together is the real heart of the building process. And, for many manufacturers and frame builders, joining offers the potential for product differentiation. And I too have been working on my proprietary methods - with some success.

Most builders & manufactures use some form of wrapping the joint in CF and epoxy. Within this method, there are two primary approaches: a) wet wrapped CF vacuum-bagged until cured; b) pre-preg CF wrapped, heated under pressure in an autoclave until cured. The second approach is heavily used by manufacturers. It's an easier method to control the amount of epoxy in the CF (because it comes pre-impregnated), and pre-preg is relatively easy to handle while setting it up to cure. Two problems exist for this method in small volume production: a) Autoclaves are expensive; b) the product is molded - molds are expensive and limit dimensional flexibility. So, pre-preg is ideal for volume production.

Wet wrapping involves several steps. First, the layers of CF need to be cut out with the fibers oriented to plans. CF has little compression strength, so it requires fibers to be aligned in a variety of directions so that any force on the joint will be compensated for by fibers working in tension. In fact, if CF had the same strength in compression as in tension, we could make much lighter frames - using much less CF.

Anyhow, CF we use (except for cosmetic out layers) is unidirectional. That is, it isn't woven, all the fibers run in one direction. Various methods are used to hold the fibers together prior to being laid up with epoxy. None of these methods are perfect. So just cutting out the patterns on the dry CF can be difficult and requires a sharp scissor.

After the layers are cut out, we prep the tubes. This means lightly sanding the surfaces and then cleaning them with rubbing alcohol or acetone. The goal is to have clean bare CF on the tubes for bonding.

Then we mix up some epoxy. There are various approaches to mixing including: a) Electronic scales; Graduated cups; Calibrated pumps. Any of these approaches will work if care is taken. Once the epoxy base and hardener and dispensed, we have to mix them together thoroughly. This usually has the result of infusing oxygen bubbles into the CF - which we will address later.

The mixed epoxy has a limited pot life. By choosing different hardeners, and being sensitive to the ambient temperatures, it's possible to adapt the pot life for the task at hand. Note that there is a general rule that the longer the pot life, the longer the cure time. So we want to limit pot life to what we really need to assemble a joint and get it ready for curing.

Next we need a flat surface, which can be covered with saran wrap or wax paper (to keep the surface clean). We take our CF pieces and lay them down one at a time. Pour some epoxy on top and use a scraper or squeegee to spread the epoxy between the fibers. We want to avoid having the CF be soaked in and dripping with epoxy, but we want it to be full of epoxy. Depending on the setup, we may do this to all the CF pieces first, and them layer them on the joint. Or, we may apply each piece of CF as it gets wetted out. In either case, we end up with our tubes wrapped in layers of epoxied CF.

Over this we put a layer of material that won't stick to the epoxy. A mylar film can be used to get a very smooth finish, or a teflon coated polyester fabric can be used. The later gives a rough surface, but is better at allowing excess epoxy to flow through. And we want it to flow through to the next layer - which is a synthetic cotton batting. This batting performs two tasks. We will put this whole contraption in a sealed plastic bag, and use a vacuum pump to suck out all the air. The external air pressure will act as a giant clamp holding things together during curing, but more importantly, it will compress the layers of carbon fiber in the joint. In so doing, excess epoxy will be squeezed to the surface, and the batting will catch and hold this excess so that it doesn't enter the pump (which would be a disaster). Also, the batting provides a channel through which the pump can continue to suck air even as the bag collapes. Otherwise, the bag opposite the vacuum fitting would get sucked into the fitting and stop it from evacuating the rest of the bag - which would do us no good.

In the process of sucking epoxy through the layers, we hope to make sure that any voids in the CF are filled with epoxy and any air bubbles are pumped out. The reality is that this will never occur perfectly, but with good vacuum pressure we can eliminate enough voids and bubbles to ensure a strong, quality joint.

A key to making all of this work is holding the tubes together, in the proper position, as the CF is wrapped on, and until the epoxy cures. A number of approaches work, from fixturing the tubes to gluing them together.

I like the later approach, as it's possible to assemble a full front triangle and then vacuum the joints one at a time. But the bonds are fairly delicate and this got me thinking of a better way to join tubes. I've developed a proprietary method that I call full surface bonding. Without giving away too much, bond a solid surface, not a hollow tube to the adjoining tube.

How strong is this? Well, I wouldn't ride a bike so built without CF wraps around the joints. But, the point of failure is delamination of a tube surface. Think of it like this. We have a plain tube and one that is mitered. The mitered tube is bonded to the plain tube. When this joint fails, it is the surface of the plain tube that is failing - not the adhesive and not the mitered tube. In other words, this joint is as strong as it can be given lamination strength of the tube to which it is bonded.

One notable feature of this method is that it adds negligible weight to the joint. As implied above, I've been doing destructive testing of my joints. So far, with full surface bonding, I'm still using the same schedule of CF laminations on the joint. This ultimately produces a stronger joint, with a weight difference that is hard to measure. The goal, is to establish a joint that is as strong as a normal wet wrapped joint, but which has fewer laminations of CF and epoxy to save weight. When testing indicates that this is ready for market, I'll be sure to let you know.

In the mean time, I can miter and jig assemble my frames similar to steel frames, when these are set (and naturally in super alignment), I come back and vacuum one joint at a time - allowing for perfectly laminated joints.

Well, that's it for tonight. Gotta run, so we'll see you soon.

No comments: