Fabricating a Custom Spinner

The spinner and cowl of my custom SR-1 race plane were drafted together to maintain curvature continuity. It was highly unlikely there was a spinner out there that would perfectly match the diameter and curvature of the CAD model, so I knew this meant I’d be fabricating my own. That said, I also knew that I would be pairing an Airmaster (AM) constant-speed hub with custom Catto blades to the Edge Performance 912STi. I therefore perused AM’s line of spinners and saw one with a backplate diameter within 0.1 inch of my design, so I tweaked the cowl CAD model to be able to make use of AM’s spinner backplate (but not their spinner). Although the backplate is the easier of the two components to fabricate, I wanted to use the AM backplate for initial flight testing for a couple of reasons.

First, although the wing has been located to minimize the need for any fore/aft ballasting, it’s possible that a small amount of ballasting may be required. The aluminum backplate weighs approximately 5 ounces more than a carbon fiber backplate would, and is about as far forward as you could hope to ballast.

Second, a carbon backplate will require some precise press mold fabrication (or post-process machining of the B-side) to achieve the required high degree of flatness, as it is sandwiched between the CS hub and the prop flange. This ensures the axis of rotation remains aligned to the crankshaft. The prop lug holes would also require machining. Nothing terribly difficult, but in the name of moving things along, the aluminum backplate will do until I have time to come back and fabricate a carbon one (if we need to take weight out of the nose or just want to reduce aircraft weight).

A carbon backplate certainly has advantages. As noted, it would be about two-thirds the weight, for a savings of approximately 5 ounces. That may not seem like much, but anyone who has followed construction of the SR-1 knows that weight is a huge deal for this airplane to meet the class weight requirements (less than 661 pounds takeoff weight, including pilot and fuel). I’ll often rework parts to take off a tenth of an ounce, so 5 ounces is a significant weight penalty if I don’t need ballast up front.

Carbon’s other advantage in this application is fatigue resistance. I once scoured the country for a replacement for my Grumman Yankee’s cracked aluminum backplate (I imagine there are many other GA airplanes with the same problem) and finally found one, albeit obnoxiously expensive. That said, this airplane is highly unlikely to fly enough hours to ever have that be a problem. (But if you’d like a carbon backplate for your [insert name of ancient spam can here], hit me up!)

With the backplate discussion out of the way, let’s dive into the spinner. The first step is to fabricate a mold. I had initially hoped to CNC a male buck upon which to fabricate a female mold, at the same time as machining the cowl molds (more on that in a future article). Paulo Iscold kindly assisted with this machining, and it was decided that we were unlikely to achieve the degree of concentricity I desired with his machine (a large five-axis Haas that is great for things like fuselage molds).

After returning to Santa Paula from Paulo’s shop in San Luis Obispo, I hit up Andy Chiavetta, who has turned a few spinners on an old lathe in his shop. Andy turns an initial shape out of foam, fiberglasses it, sprays it with Duratec 707-002 Gray Surface primer, and then final-sands it using sanding blocks and templates. He does all this without ever removing the buck from the lathe in order to maintain concentricity.

Armed with this info, I decided I would turn the spinner on my Bridgeport mill. I felt that keeping the part vertical during shaping would avoid the slight runout that gravity would impart if turned horizontally (I should note that neither Andy nor I used a live center at the spinner tip). For a buck, I used 4-pound tooling foam. A precision-ground 5/8-inch steel rod (McMaster-Carr) was epoxied into the blank for workholding. Rather than hand-shaping to a template, I 3D-printed a sanding tool that I mounted to the Bridgeport. This worked quite well, and after the initial shaping, I removed the buck and applied a heavy coat of Duratec. I skipped the fiberglassing, as in my experience Duratec over tooling foam is sufficiently durable for one-off molds.

While I appreciate the fact that whenever you remove or re-chuck a part in a lathe or mill you lose concentricity, I was unwilling to spray Duratec with the part still in the mill. Thankfully, I had applied enough Duratec to sand the tool back to concentric without any low spots.

Once I was happy with the surface finish of the Duratec, a pair of scribe lines were added near the spinner base trim line by chucking a pointer into the cross bridge and running it lightly into the spinning buck. To be clear, I wasn’t exactly sure where the spinner would terminate, so one scribe line indicated the termination point based on the CAD model, while a second scribe line would allow me to measure out a planar termination if I ended up trimming the spinner base past the first scribe line. The point is: you don’t want to lose your planar reference when trimming the spinner base. This keeps your axis of rotation parallel to the crankshaft’s.

The buck was then removed from the mill and mounted to a flat board for the next step of fabricating the female mold. The buck was waxed (any mold release wax will do; I used three coats of Partall paste #2) and shot with PVA. It was then coated with a layer of epoxy and cabosil. This mixture has the consistency of petroleum jelly, helping to avoid bubbles on the mold surface and ensuring good adhesion of the tooling cloth. Tooling cloth is usually a heavyweight carbon or fiberglass, but in this instance, I simply used scraps of the same 6-ounce cloth I use elsewhere on the airplane. Ten layers was sufficient for this one-off project.

Once the mold had cured, the plug was removed and set aside. A trial spinner was laid up in the female mold to get a feel for how the layup process would go (same release schedule as above). For this spinner, each ply consisted of three triangular-ish sections. Different designs might require more or fewer sections. For purposes of rotational balance, it’s generally best to a) reduce the number of sections (and thus overlaps), b) reduce the amount of overlap between sections (in this case, the overlap was half an inch ), and c) offset overlap locations at each ply to avoid overlaps from different plies aligning.

I can’t advise on what an adequate number of plies for a spinner will be—it depends on various factors like diameter, length, maximum rpm, number of blade cutouts, etc. However, as a starting point, one can aim for a laminate of thickness equal to a similar aluminum spinner (this is somewhat humorously referred to as “black aluminum” in the aerospace industry, since a dimensionally identical carbon replica of an aluminum part is likely sufficiently strong but hardly optimized). Recognize that the spinner sees a combination of centrifugal, shear, and bending forces , so ply orientation (i.e., 0°/90°, 45°, or some other combination of angles) is important.

The trial spinner revealed a few small blemishes in the mold that were filled with 3M spot putty. The whole process was then repeated to create the actual spinner. The trial spinner was later used to ensure the prop cutouts were the right size before performing the cutout on the actual spinner. (I often make trial parts to prove out the fabrication process. It’s great to keep these parts to determine final trim lines, etc., without worrying that I might ruin an otherwise perfectly good final part.)

The next steps were to create a forward bulkhead and a mounting flange. Although most spinners simply run screws through the spinner side into the backplate flange, I’ve never been a fan of this approach for composite spinners. Aside from the fact that the exposed spinning screw heads aren’t particularly aerodynamic (of no concern to most, I guess), the faying surface of a composite spinner usually does not perfectly match the backplate flange. Perhaps I’m overthinking this, as this is how it’s done on most aircraft and seems to work fine. Nevertheless, my approach is to bond a molded flange to the inside base of the spinner that seats the spinner on the backplate (see picture). This flange uses the backplate itself as the mold, so the fit is perfect. I drill a small dimple at one location of the flange, which serves as a detent to properly clock the mounting flange.

The aluminum backplate was released with packaging tape and wax, and the mounting flange was vacuum-bagged on top of it. Once cured, the mounting flange was removed and trimmed flush with the aluminum backplate. A drill-and-trim template was drawn up, which I had cut by SendCutSend. An assembly of the template, mounting flange, and AM backplate was then sandwiched together and the screw holes match-drilled. The AM backplate was removed from the stackup, and the template was then used to drill the nut plate holes in the mounting flange. Nut plates were riveted to the mounting flange.

Next, I fabricated the forward bulkhead. The forward bulkhead self-centers on the motor casing of the prop hub, which is concentric with the axis of rotation (this is an obvious assumption, but was still confirmed using a dial indicator). The bulkhead is a slide fit over the casing, and it was vacuum-bagged on a 3D-printed mold such that the outer mold line (OML) of the part is fay (precisely shaped to mirror and nest tightly) to both the casing and the spinner’s inside wall.

The AM backplate, mounting flange, and constant-speed hub could now be bolted together and mounted with an adapter to a rotary table on the Bridgeport mill. These parts were checked for concentricity with a dial indicator. Turning the rotary table by hand is too slow for a quick readout, so I chucked an Allen bit on a flexible drive into a hand drill to turn the table, and this worked well.

Once I was happy with the concentricity of the above assembly, it was time to measure the runout of the spinner and to shim the bondline gaps to minimize it. The forward bulkhead and mounting flange both bond to the spinner, and I designed this bondline at 0.030 inch. This is thicker than would normally be required for a good bond (typically 0.010–0.020 inch ), so the extra bondline thickness allowed me to adjust the gap larger or smaller to minimize runout. I superglued some small 0.030-inch-thick bond spacers every few inches around the perimeter of both the forward bulkhead and mounting flange to set the nominal gap. I then measured runout, and added or removed material to the spacers until it was minimized.

It was now time to bond the forward bulkhead to the spinner. I slipped the forward bulkhead over the Airmaster hub and positioned it a bit higher than necessary. I then spread adhesive on both the bulkhead flange and the mating surface of the spinner. I then dropped the spinner over this assembly and pushed down until it seated on the base flange.

Because I had added a few small pieces of foil tape to the Airmaster hub, the slip fit was now slightly tight. This ensured that once the spinner was dropped into place, it would push the bulkhead down the hub until the spinner seated on the mounting flange, and the bulkhead would stay firmly clamped against the spinner while the adhesive cured. Immediately after dropping the spinner into place and before the adhesive cured, I checked the runout to ensure it matched what I had seen during the dry-fit.

Once the forward bulkhead bond cured, the next step was to bond the spinner to the mounting flange. The spinner was dry-fit for a runout check to make any final small adjustments to the spacers if required. Note that for all these steps, clocking the spinner to the mounting flange is necessary since the wall thickness of the spinner varies by a few thousandths on the B-side due to ply overlaps, and it’s precisely this variance (as well as other sources of runout) that is being accommodated by the bond spacers.

Once I was happy with the dry-fit runout, the mounting flange was bonded to the spinner in the same manner as the forward bulkhead. Final runout was approximately plus or minus 0.006 inch (it varied depending on whether I was measuring at the forward bulkhead or base of spinner), which was slightly over my goal of 0.005 inch, but I think otherwise acceptable.

The final step to finish the spinner was to make the blade cutouts and fit the gap fill plates. Due to the small size of the spinner, the blade ferrule extends partially beyond the OML of the spinner. While not ideal from an aerodynamic point of view, it simplifies the blade cutout as blade rotation does not affect the cutout shape (it’s just a circle). I 3D-printed a scribe template that was slightly oversized to the ferrule and removed the cutout on each side (as noted above, I did this on the practice spinner first).

Once I was happy with the spinner-to-ferrule clearance, I made a splash mold of the inside of the spinner at each cutout. I used this mold to vacuum-bag a pair of U-shaped doublers that I bonded to the spinner to reinforce the cutout location. Lastly, I used the blade cutout remnants and trimmed them down to use as the gap fill plates.

Part of the design and fabrication intent was to minimize mass imbalance from the spinner, so it will be interesting to see how well the prop balances dynamically. As an assembly, it is obviously impossible to separate the contribution of the spinner versus the backplate, hub, and blades. However, the latter have been statically balanced by both Airmaster and Catto as part of their manufacturing processes, so my hope is that dynamic balancing isn’t too problematic.

The final weight of the spinner is significantly lighter than the Airmaster aluminum spinner that would normally come with the backplate, although as pointed out before, any advantage must be considered against ballasting requirements. That said, the reduction of rotating mass is a net positive. As with most projects on this build, fabricating the spinner was fairly time-consuming, but that is the price you pay for custom parts, and I’m quite happy with the final product.