In composite aircraft, molded parts can be broadly classified into three types: (1) external shells, such as wing skins and cowls; (2) internal support bulkheads, such as wing ribs and fuselage formers; and (3) other parts, including spars, brackets, fairings, etc.
In Part 1 (September 2017) of this series, we explored how to make simple flat composite panels, and then in Parts 2 and 3 (November 2017 and January 2018), we looked at slightly more complex molded panels. In Parts 4 and 5, we’ll discuss a method to fabricate bulkheads such as ribs or formers using male molds. Note that I’ll use the words ribs, formers and bulkheads somewhat interchangeably since, for purposes of this article, they are essentially the same. I’ll also assume you have read Parts 1, 2, and 3 and are thus already familiar with the concepts we’ll be discussing.
Two common approaches for fabricating composite ribs: Flap rib has been CNC cut from a flat carbon-foam sandwich panel and is then bonded in place. A 1- to 2-inch wide strip of BID tape (2-3 layers of bias carbon or glass) can subsequently be bonded along the joint line to reinforce the bond.
Two common approaches for fabricating composite ribs: Molded ribs have the advantage of providing a bonding flange.
Parts from Flat Panel Stock
Traditionally, items such as ribs and bulkheads have either (1) been cut from flat panel stock (either by hand or with a CNC machine) then taped into place, or (2) molded in female molds then bonded in place. Both of these methods have their own particular advantages and disadvantages.
Parts made from flat stock have the advantage of ease of fabrication, and when cut by CNC are dimensionally accurate to thousandths of an inch. Flat stock ribs pose a challenge though in structures like wings, which are closed out by bonding on the top skin as a final step. Before closeout, the builder has access to the lower surface/rib bond line, so this bond is easy to deal with, and BID tape can be easily applied as reinforcement if necessary.
However, once the upper skin is bonded in place, the builder mostly likely no longer has access to much of the wing internal structure, and so filleting/taping the ribs to the top skin is impossible. This means that a bond flange must be fabricated on top of the rib before closeout, which can be quite time consuming and potentially introduce dimensional inaccuracies. In some cases the designer is OK with an untaped joint, in which case the builder simply creates a trough in the foam core with a troughing tool, fills the trough with sufficient adhesive, and performs the closeout.
Bonding flat-stock ribs. The trough is approximately 1/8- to 1/4-inch deep and is pushed into the foam core with the trough tool. The trough is then filled with structural adhesive prior to closeout. Depending on the foam core, you may need to re-roll the trough prior to filling with adhesive, since some foams spring back after a few hours or days.
As an aside, top/bottom skin here refers to the orientation during building, not the actual upper/lower surface of the wing. The actual upper wing skin bond is generally considered more critical than the lower since it is more likely to separate during positive G-loading. (Positive G-loads tend to push the lower skin against the spar and rib surfaces, but push the upper skin away from the spar and rib surfaces.) Therefore, some manufacturers assemble their composite wings upside down so that they have access to the critical upper surface bonds, leaving the less critical lower surface bonds for the blind closeout. That said, other considerations such as sealing wet wings, landing gear installation, etc. can make building the wing right-side-up more advantageous.
Another disadvantage of flat stock ribs is that a filleted and taped joint is slightly heavier than a joint with a molded-in bond flange by approximately 0.5 to 1.0 ounce per linear foot (see sidebar). That may not sound like much, but given that most aircraft have several hundred linear feet of joint line, the weight difference can amount to several pounds. Nevertheless, for ease, speed of construction, and parts accuracy, CNC’d panel ribs/formers are hard to beat.
Parts fabricated in female molds address some of the problems with flat stock formers. The bonding flange is molded into the part, so filleting, taping, and closeout flange construction is avoided. If the mold is CNC’d, it will be also be very dimensionally accurate. Different layup schedules can be used without affecting the OML (outside mold line).
On the other hand, making the molds themselves can be time consuming. And once the mold is made, you are stuck with that shape (whereas it’s pretty easy to CNC a new rib from flat stock if you need to modify the dimensions). For multiple pulls (for example, if you are manufacturing a kit, or if all the ribs in your wing are the same dimension), the time required to make the molds is probably worthwhile, but for rapid prototyping or one-off projects, it’s a lot of work.
The method described below is something of a hybrid approach with a number of nice features. The male molds are relatively fast and easy to make, and also easy to modify. It is much easier to vacuum bag convex shapes over male molds than concave shapes in female molds. The bonding flange can be easily trimmed to whatever width the designer wishes. Parts made this way are accurate to about .005-.010 inch, vs. .001-.005 inch for parts CNC’d or made with a female mold, but for most applications this should not be an issue given that it is significantly less than the bond gap itself. Finally, the flange itself can be easily wrapped with unidirectional carbon fiber during the layup to provide cap strips tailored to design loads. Given the accuracy, relative ease of fabrication, and reduced weight, I decided to fabricate the SR-1 wing ribs using this method.
Making a Cross Section
The first step is to generate a cross section of your part. If the cross section is an OML (outside mold line), you must then offset the thickness of the skin to which you will bond, the bond gap, and the predicted bulkhead flange thickness from the OML (see Figure 1). In the example, the wing skin is .290 inch thick, the bond gap is set at .060 inch, and the flange will be .040 inch. The total offset from the OML is thus .390 inch. Near the leading edge, there is no foam core, so the offset decreases .250 inch to .140 inch. I suggest making one or two practice ribs to master the technique and also work out your material thicknesses. For example, I find one ply of vacuum-bagged 6-ounce 282 plain-weave carbon fiber consistently measures .009 .0005 inch. It’s also easy to incorporate alignment holes or other assembly aids into the template, as seen in the opening picture.
A long-reach thickness gauge is convenient when measuring wing- or fuse-skin thickness for determining rib offsets.
The offset profile is then stored as a dxf file for the CNC, and the template is machined or laser cut from a suitable template material (I recommend 1/8-inch aluminum). While this might sound expensive, I used a local CNC laser cutting shop, and the turnaround time was less than an hour and about $15 per template for a full set of ribs.
While bond gap thickness is outside the scope of this article (we plan to deal with it in a future article on adhesives), a good rule of thumb would be to plan for a bond gap thickness of .050 to .100 inch. Aerospace industry standards are closer to .005 to .015 inch, but that is using very precise alignment jigs and hard tooling during bonding, and doesn’t allow much margin for error. A larger bond gap is not as strong, but gives you a little wiggle room to accommodate small amounts of misalignment during assembly. With careful measurement during assembly and CNC molded parts, it should not be necessary to allow for more than a .125-inch bond gap.
Consider incorporating alignment or other devices into your templates. These ribs have 1/8-inch holes that allow mounting a tool to properly space split-rib halves as well as ensure that ribs are bonded into the wing at the correct angle. They also serve for laser alignment of ribs spanwise.
Making the Mold
Once you have your templates cut out, it’s time to make the male molds. Cleco the template to a piece of 3/4-inch melamine-laminated MDF, and use a flush-trim router to shape the male mold. Don’t forget to include a couple of washers between the MDF and template so that the bearing can ride on the template without the cutter touching it (you’ll know if you forget these!). Corian also works well as a mold material, and you can usually find scrap on Craigslist.
Depending on how wide a flange you want, you may need to stack two or more pieces of MDF. Since flush-trim router bits typically only have 1 inch of cutting edge, for molds deeper than 1 inch, it works best to cut one piece of -inch MDF using the template, then screw a second piece of MDF to the first, and router it flush using the top mold as the template. Screw the lower piece of MDF to the top piece from the bottom to avoid having screw heads on the top mold surface.
Routing the upper male mold. The template is Clecoed to a piece of MDF for rough cutout on the band saw. Note the use of flat washers as spacers between the template and MDF to avoid damaging the template while cutting.
Additional mold depth (for deeper flanges) is provide by screwing a second piece of MDF to the top mold and routering it flush using the top mold as the template.
For straight-edged parts, you can simply use a ruler as your router template, as is being done here for a drag spar mold.
Once you’ve routered the molds, you should put a radius around the top edge of the mold to help the carbon lay down over the edges (we’ll discuss allowable bend radius for composites in next month’s article). You can do this by hand with a sanding block, or if you want to be more exact, you can drop a radius bit into the router and do it that way. On parts that incorporate uni into the flange as a cap strip, I also like to put a small radius at the corners to help the uni wrap smoothly around the mold. Note that you can undercut or overcut the mold by using different diameters of router bits and guide bearings (see photo).
Next, apply foil tape to the sides of the mold to act as a release surface (you don’t need foil on top of the mold since the melamine releases easily as is). You can now mount the mold to a base. I like to use the 1/8-inch fiberboard sold for use as a whiteboard at the hardware store. It’s cheap and the white helps you see when you have enough PVA on the mold (the PVA gives it a green tinge). I then bag everything directly to a sheet of 1/2-inch tempered glass. I highly recommend this method if you plan to do any amount of composites work. Bagging directly to the glass is easy and saves bagging material, cleanup is a piece of cake with acetone (or a scraper for cured resin), and you can usually find tempered glass tabletops inexpensively on Craigslist.
For flanges that incorporate uni tape, radiusing sharp corners will help the tape wrap around the mold.
Router bits used for this project (L-R): 3/8-inch bit plus 3/8-inch bearing, 3/8-inch bit plus 5/16-inch bearing (.031-inch undercut), 1/2-inch bit plus 3/8-inch bearing (.062-inch undercut), 3/8-inch bit plus -inch bearing (.062-inch overcut). 3/16-inch radius bit, and 1/16-inch radius bit. Note that if you use MDF for your molds, it dulls bits pretty quickly, so have a couple extra bits on hand.
That’s all we have space for this time. We’ll wrap things up in the next installment. In the meantime, stay warm and stay motivated.