After the mystification of the previous example, I was wondering if there was much point in continuing, but, not one to let go of an idea too early, I decided I would forge on to the next dormer drawing puzzle in the Louis Mazerolle book from 1866. This next dormer is on a bias - that is, the entire plan of the thing has been pushed sideways to form a parallelogram. Here's the text illustration:
Why on earth would you want to build such a dormer? Well, even Mazerolle notes, d'une application assez rare. I could imagine that if the roof of the building went one direction, and a view or some such thing was angled slightly away from the line of the roof one might opt for such a dormer. Regardless, it is a layout conundrum.
Here's the plan view of the layout:
Note in the middle how the front posts are shaped into parallelogram sections, roughly speaking. Note how the rafters are arranges so as to be on a bias to the line fo the main roof. The layout method shown for determining the noulet, or little valleys, is the same as the method shown in the previous drawing, the Nolet Carré example, and I already found that method didn't work at all, so I wasn't feeling especially encouraged as I undertook this drawing to see the same method being proffered. My hope was that some new piece of information would come to light which would make everything miraculously work out.
Well, here is the dormer on a bias, as far as I have been able to take it:
I wasn't able to make the drawing method shown work for producing the noulet, so I tried to figure it out otherwise, without success. Then I draw it point to point using 3D and was able to make the part. Then I tried to reverse engineer it in the 2D development, for hours and hours I tried, to no avail.
I sent some information to a compagnon that I know, and am hoping he will be able to explain this mystery. In the meantime, I thought I could at least try to finish as much of the drawing as I could. Next stop was the molding which wraps around the front edge of the roof, the moulures de tympan. It's called a tympanium, or drum-shaped molding in English. Here's the section of the drawing that deals with the development of the section which runs along the long sides of the dormer, and the development of the section which climbs the gable at the front:
The slope line, rampant du fronton, indicated that the molding to climb the gable is to be set at it's slope for the development line transfer over from the normal section shown at the right side.
Well, I was skeptical right away when I saw this as it seemed to defy logic. Note how both moldings shown have top surfaces which are square to their inside faces. Now look again at the plan view, which shows the moldings wrapping around the front of the posts with a miter:
So, what the drawing is telling me, and the perspective view also shows, is a mitered connection. It is indicating the molding is to come to the corner, turn about 115˚ on the horizontal and then climb 45˚ uphill, and that the connection between pieces can be both mitered and have squared upper surfaces to the face wall line?! Welcome to the world of fantasy!
I couldn't see how these conditions were going to be met with a piece traveling through an unequal compound miter. Okay, okay, I thought, that doesn't seem likely to work, but let's put our preconceptions to one side and just follow along and see what happens.... Maybe I was seeing it wrong?
Nope, I was seeing it right. Following the development shown in the book, I produced both sections, and put them on the roof and extruded them into one another. They didn't match up at all, in any fashion. The shaped weren't even remotely coincident at the miter.
So, to hell with the drawing in the book!
Next, I used the 'normal section as a basis from which to develop the molding to climb the gable. For this I chopped the normal molding at the horizontal bisection of approximately 57˚. This process took quite a while, as I had to connect all the individual marks which make up the ogee profile on the molding and create a distorted version of same on the climbing piece of molding. Here's how the connection then looks, when connecting at a miter on the face and keeping the upper surface of the molding square to its backside:
As you can see, it doesn't meet at a miter - in fact the climbing piece ends up with a barbe on top. One could I suppose live with that.
Even worse though is what happens when the newly developed climbing moldings is extended from the peak down the other side to meet the opposite long side molding:
No way that's going to fly my friends. If I then reshaped the horizontal molding to fit to the one which descended to meet it, I'd end up with a fairly slim unit, I think noticeably so. It would also mean three separate shapes of moldings to do the job, hardly what one might call efficient. That dog won't hunt.
Then, putting that matter to one side, there's the issue of what happens at the peak of the roof with this arrangement on the climbing moldings:
Notice how the miter line at the peak is turned away from the dormer ridge and is actually perpendicular to the line of the main roof's travel. In fact, if you look closely at the drawing, it seems to indicate that this is the result sought. It looks ungainly to me and I wouldn't build something like that.
Okay, okay, let's see if I make some changes.... I thought, well, let's make the junction at the corners an actual full miter to see what happens:
Voila! problem solved, eh? Is it time for beers?
Uh, not exactly. The effect of rotating the top surface of the climber to meet the horizontal molding is that the climber's top surface is warped thusly:
Ya can't have something like that on a roof. And comparing this set up to the perspective view from the book shown at the top of the page, it would appear that this is not the intended outcome either.
Here's a look again at the method Mazzy shows to develop the tympanium molding:
And here's a comparison of the actual molding section that resulted from my point-to-point connections to produce the part (with a squared upper surface):
If you follow the lines of development, you will see that most of them do not coincide with the actual climbing molding. The drawing in the text is, once again, completely misleading and shows a result which is impossible or unusable. C'est merde.
WTF is up with this book? It's becoming sorta humorous now. How can it be that example after example is basically un-buildable as shown? Somebody help me out here, as I think I'm going to have to turn to hard liquor.
I'm going to have a crack at the next dormer in the book, despite all the issues I have uncovered, 'cause I guess I must be masochistic. The next one looks kinda fun, so I'll have a go, but if it is another unbuildable piece, I think I'll lay the book down for a good while.