Recreating the Wright R-1820 “Cyclone” (4)

The fins of the air-cooled cylinder heads are a state-of-art piece of metallurgy (Figure 86‑1):

Figure 86-1 Fins of the R-1820 cylinder head

At the first glance, it is hard to believe that they were cast as a single piece. But when you look closer, you will discover that these fins “grow up” from the solid parts of the head as naturally, as the hair from the head (Figure 86‑2):

Figure 86-2 Close-up of the fin details

Try to imagine the shape of molds used in the production of these parts, and the challenges faced by their manufactures! (There is an interesting post about this. It describes production of the R-1830 Twin Wasp cylinders). Basically, modern producers of the heads for the air-cooled aircraft engines use the same technology as eighty years ago.

In my model I will recreate these fins in a somewhat simplified form, as a few separate Blender objects. I will also skip some fine details of their shape (for example the small features that I marked in the figure above). Such a simplification conforms the moderate level of details that I assumed for this model. It is always possible to make a more detailed version of this object later.

I began by forming the “external boundary surface” of the fins. After revising many photos I decided that it has a circular base. This base is combined with a shape extruded from a perpendicular arc (Figure 86‑3):

Figure 86-3 Forming the shape of the fins boundary

The rocker covers, formed in the previous post, helped me in estimating the shape of this object. In general, the cylinder head fins are not symmetric, since the exhaust valve produces much more heat than the intake valve. Thus there are more fins around this area. Initially I formed the basic shape, leaving gaps around the rocker covers (Figure 86‑3).

In the next step I filled these gaps (Figure 86‑4a):

Figure 86-4 Closing the gaps in the boundary shape

In fact, it was sometimes quite hard task that required careful analyzing the shape of the head fins in these areas.  Note that fragments of the rocker pushrods were partially “sunken” in this object  (Figure 86‑4b).

I cut out the areas around these pushrods using the Boolean (Difference) function. To do it, I placed along the rockers two simple “boxes” (Figure 86‑5a). Then I used them as the “tools” in a Boolean modifier that cuts out from the boundary shape the difference of their volumes (Figure 86‑5b):

Figure 86-5 Cutting out some space around the pushrods

(Note that I rounded the original sharp edge of the “cutting box” using a multi-segment Bevel modifier). Then I “fixed” (applied) results of the Boolean modifier. After removing unnecessary vertices and edges from this area, I obtained the shape shown in Figure 86‑5c). Finally I dynamically rounded the external edges of this cut out, using a multi-segment Bevel (Weight) modifier.

In similar way I created the hollows for the spark plugs. First I created two objects that have the shape of these cutouts. (As you can see in Figure 86‑6a), their shape was more complex than the pushrod “boxes”):

Figure 86-6 Hollows around the spark plugs

I used these two objects in a Boolean modifier, which I applied to the boundary shape object. Figure 86‑6b) shows, how this mesh looks like after “fixing” the results of this modifier. I also dynamically smoothed the resulting mesh using a moderate (level =1) Subdivision Surface modifier.

Finally, when the boundary shape was formed, I started to add the head fins. In the simplest case the mesh of a single fin can be just a single square face (Figure 86‑7a):

Figure 86-7 The basic idea of creating head fins

Then I obtained the results as in Figure 86‑7b) in a dynamic way, by adding to the fins object a stack of three modifiers:

1. Boolean (Intersect) modifier, which uses the boundary object as the “cutting” tool;
2. Solidify modifier, which gives the fins their thickness;
3. Bevel (Angle) modifier, just to “round” the external edges of the resulting fins;

As you can see above, their cumulative effect is quite interesting.

All what I have to do now is to add to this “fins” object subsequent faces. The “L”-shaped upper fins have somewhat more complex topology (Figure 86‑8):

Figure 86-8 Adding the upper (L-shaped) fins

It is built from a dozen of elementary square faces. I crated the rounded edge in the middle of this fin by adding another multi-level Bevel modifier to the top of the modifier stack. It rounds selected edges – those, where I set the so-called Bevel Weight coefficient to a nonzero value.

Sometimes the results generated by the Boolean and Solidify modifiers look strange. To fix these problems I had to be careful with the normal direction of the newly created mesh faces. Sometimes I even had to add an additional edge loop – because it alters the results of the tessellation that Blender performs for each face.

The R-1820 cylinder head also contains some “M” – shaped fins (Figure 86‑9a):

Figure 86-9 Adding the “M”-shaped fins

I built such elements using an outer “U”-shaped surface combined with the inner, flat face (Figure 86‑9b). The faces of the “U”-shaped surface have their normals directed outside, so the Solidify modifier generates the thick “walls” around it. The two edges at the “bottom” of this “U” are rounded (Figure 86‑9b), as in the case of the “L”- shaped fin. The inner surface extends a little (by less than the fin thickness) outside the original faces of “U”-shaped surface. After applying the modifiers, this “overflow” creates an impression that both surfaces are joined.

The final mesh of the head fins resembles somewhat a Minecraft object (Figure 86‑10a):

Figure 86-10 Complete fins (as the mesh and the “trimmed” object)

However, when you switch into the Object Mode, in a split second the modifiers transform it into the desired shape (Figure 86‑10b).

Do not be mistaken by this “smooth” workflow description. In practice I often had to make minor adjustments to the boundary surface mesh, to correct some unexpected effects of the Boolean modifier. Fortunately, the mesh of each fin is disconnected from the others, so all these issues appeared gradually, and I was able to resolve them in a systematic way. I had also made other adjustments: for example, in the middle of the work I discovered that the spacing between the fins was 0.215” instead of 0.220”. (I know that this distance seems extremely small, but for the 30 fins in a row, it really makes a difference!). Thus I shifted – vertically or horizontally – about two dozen fins. Fortunately, the Boolean modifier took care for the resulting adjustments in their shapes.

What’s more, it occurred that these dense, evenly spaced fins act as a kind of additional reference grid. While forming them, I found and corrected some inaccuracies between the fins and the valve and rocker covers.

For example: while forming these fins, I adjusted at least four times the angle, location and shape of the intake valve. And after each of these modifications, I had to fit anew the intake pipe. That’s why I prefer to keep such complex elements as this cylinder head split into various simpler objects as long as possible: you never know, when you have to modify them again!

You can check details of these fins in this source *.blend file. The model starts to resemble the real cylinder, but it still lacks many details. I will describe them in the next post.