Following Mazerolle: Lucarne à Croupe Nolet Carré (II)

Now, it wasn’t my intention to be devoting two posts to each of the drawings in the Mazerolle book, however, such is the case. I hope no one is falling asleep. Anyway, here we are again, with the hipped dormer problem detailed in the previous post:

Yesterday I mentioned that the drawing in the book for developing the view of the noulet was screwy, so I employed a different technique to produce the pieces. That method led to a perfect result – here you can see how the foot of one of the noulet (in yellow) meets the plate as it should:

My thought, always willing to give the benefit of the doubt, was that there must be some sort of new concept I was unfamiliar with in the Mazerolle drawing, and thought therefore that the best way to figure it out would be to place the completed piece, which I knew to be correct in shape and dimensions, back onto the drawing and reverse engineer it, as it were, to see if I could crack the code. Seems like a reasonable way to proceed, no?

Here’s a view of the developing drawing portion, which I situated directly below (ie., in plumb) the actual dormer so as to easily transfer lines roof to floor:

First off, here’s the developed drawing portion as shown in Mazerolle:

On the left side is the end portion of the plan view of the hipped dormer. Fig. 2, on the right, is supposed to show the development of the noulet. The piece LMQR is one of those noulet, KNPS is the other.

The weird bit is the establishment of the height, which is length AB. Trouble is, that length is not the same as the actual height of the dormer off the plate, but somewhat shorter. It should be the height off the plate as far as I’m concerned. I used a compass set to that measure AB anywayand looked around the drawing at a few likely places to see if I could find something the same, all the while doing my best in a carefully reading and re-reading of the text in its somewhat inscrutable 19th century carpenter’s French. The best I could figure it, length AB was supposed to be the same as the length CE in the following portion of the drawing:

This section is the remainder of the plan view, and shows the noulet looking down from above. Thus length CE is the distance in plan from the foot of the noulet to their heads.

That use of CE as the height AB wasn’t making sense, but what the heck – let the reverse engineering begin:

You can see that my rendition looks pretty darn close to the text. I have brought down a copy of one of the 3D noulet, and placed it on top of the plan. Trouble is, it doesn’t quite match:

In the above picture you can see that the dimension of the lower face of the noulet, where it meets the plan, is not the same as the drawing. If I drop points down from the corners on the upper face, they don’t met the plan at the right spots either.

At the foot end, the fit is also off:

So, I conclude that the method shown in the book on that section doesn’t work at all to produce a correct part, though I am open to the possibility of profoundly misunderstanding the purpose of that section of the drawing.

One more detail – the bottom of that drawing in the text showed the means by which a stick of wood can be cut obliquely through its cross section to produce two of the trapezoidal noulets. Here’s a focus in on that portion:

You can see two trapezoids touching at one corner. The lower one to the left side is the actual cross section from the 3D noulet which I know to be sized perfectly. The one on the right above is that produced by the drawing method shown in the text – not even close, either in width, height, or angle of cut. Look back to the Fig. two illustration a few pictures up, and see the parts labeled ‘a’ and ‘b’ with the arc swung between them. That is what I have reproduced in my drawing above.

Okay, setting that frickin’ mess to one side, I went back to the main plan section of the drawing, where it showed another development for the noulet, one where the sauterelle, or bevel gauges, could be used to pick up the relevant angles. I set up my drawing in as close a representation to the one in the book as I could, then dropped the 3D parts down to see what lined up with what:

After a bit of fiddling, more than a bit actually, I was able to achieve congruence between the development in 2D and the part itself:

Trouble is, the drawing in the book is slightly misleading in how it shows things. Take for instance this picture of the cut at the top end of the noulet:

And this is what it actually looks like, note the difference in the form of the yellow trapezoid to the shaded footprint in the drawing above (and the other problems with the way the end of the developed leg looks in that drawing compared with the actual piece):

It’s little things like that which can cause a lot of confusion when trying to sort out the connections between pieces and how something is supposed to look. Besides the error with shading the footprint, the book shows some erroneous projection lines. Things like that can lead you down dead ends for long periods of time. It’s frustrating.

So, I know more than a few readers out there might be laboring trying to make head or tail of these drawings – it isn’t the simplest stuff to get your head around, and that is why so many avoid it. Despite that, I am trying to show as best I can however that the Mazerolle book again seems to be, well, totally screwed up. I hope that you don’t need to understand the drawings I have shown to be able to see that there are some problems there. If it wasn’t for the luxury of being able to make the parts in 3D, then compare them with the drawing, I really doubt I would be able to sort these issues out.

There are so many pitfalls in following the drawings in the book.

So what to do? Should I keep going in this quest to draw so many of these pieces? It’s a major uphill battle, and after such a struggle, probably 20 hours work on the simple hipped dormer, I might anticipate that there will be drawings ahead I won’t be able to figure out even after many hours. Is it worth it?

4 Replies to “Following Mazerolle: Lucarne à Croupe Nolet Carré (II)”

  1. Too bad that developing the drawings are problematic, even for Mazerolle. I would document the problem by doing one or two extra drawings, and use my energy on other topics. The situation can still change, it could be that Mazerolle became better while writing his book, or that the drawings where made by more than one author.
    It is interesting to see this problem. As far as I understand there could have have been different approaches to making connections, Mazerolle sets the emphasis on working completely on paper, where the examples shown in French Connection 9 could be made by working more directly with the model.

  2. Damien,

    Mazerolle shows the primary French techniques in his drawings – either the piquage (scribing) or the full drawing development. Even if you want to do the scribing, some drawing development is still needed simply to place the pieces correctly in relation to one another on the floor. The full development is common to most models in the book – sometimes with the use of bevel gauges, sometimes without. All those models in French Connection 9 would have been done with full developed drawings. The building I was showing a model of has a developed floor plan which I made invisible for sake of clarity when posting the pictures.

    I don't know what to make of the odd method that Mazerolle has been showing in the dormer drawing. In the following dormer, which I am working on right now, the same method is shown, so I am left thinking there must be some way it works, and am going to try again to see if I can figure that out.

    I can of course figure the parts out in 3D if I need to and work backwards, so that is a big help.

    Thanks for your ongoing interest – I fear many readers might be lost in the drawings I've been showing and not getting much meaning from these recent posts.


  3. Hi Chris,
    I don't know if you are still interested in comments on this post at this late date. I've obtained a copy of Mazerolle and have worked through the first lucarne drawing. While the text is incomplete and has errors, and the engravings are problematic, the method does work. I drafted a lucarne based on the plate and then constructed a 3D model from that in Blender, and the model appears to be correct.

    The text is frustrating because it is directed at advanced apprentices, perhaps even compagnons doing their “Tour de France.” So, basic techniques, like constructing an auxiliary view of a roof surface (“la herse”), are simply not covered or are mentioned in passing. The errors I've found are mostly typos i.e., referring to wrong or non-existent letters on the drawing, or leaving a prime symbol off a letter. The fine details of the engravings should not be trusted. I don't know how the engravings were produced, but I bet that they were traced somehow from the original plan and, in that process, errors were introduced: straight lines got bent, letters were left out, lines were connected to the wrong point, etc. One of the most frustrating things about the drawings is that one has to guess which of several intersections a named point refers to.

    On the other hand, the methods described are rooted in descriptive geometry, and the math “works.” Elsewhere you've speculated that the text was deliberately misleading in order to mess up competitors, but I don't think this could be true after learning something about the compagnons de devoir and their ideals.

    Anyway, you comment that AB is not the height off the plate. There are two things going on here. One is that the roof line in the elevation is lower than that in the rendering at the top of the page. Someone got confused over whether or not there was roofing material or trim involved, or something, but in any event, that's just the way it is. The whole developed plan is consistent with the height AB in the elevation. Also, the “devers de pas” are constructed on a “virtual plate” that is at the intersection of the rafter and the wall, a bit above the real plate.

    The height AB is used to draw FO above the center line of the plan. That is used to construct a side view of the valley rafter lying on the roof: D is the intersection of the lucarne roof line with the main roof. The slope of the roof and the rafter width are used to establish E in the plan view, and C in plan is just D.

    It's true that the rendering is of the valley rafter cross section in the roof “herse” is screwed up, but this is engraving confusion. The sauterelle method does work to produce the correct angles on the square face of the rafter (as verified in Blender).

    I'd be interested to hear your thoughts on the value of these methods, assuming that they can be deciphered 🙂


  4. Tim,

    many thanks for your detailed comment. I see from your profile that you are skilled in computer graphics and have an interest in French, and live in France, so you have some good background for tackling Mazerolle.

    I haven't looked at that drawing in many months and am working on other drawings at the moment, however I will revisit the drawing in light of your comments and see what I can see.

    A certain amount of my exasperation in regards to that drawing had already been 'pre-loaded', so to speak, by struggling for many months with the treteau drawing. So, when I started finding more errors in the first lucarne drawing, I was more predisposed to conclude that it was more screwed up. work on subsequent lucarnes in that book only reinforced that idea.
    The ideals of the compagnons de devoir are I'm sure without reproach, however some of the errors in that text are so egregious that I have to wonder how they could have been made. In particular, when it comes to the treteau, the leg with a tenon axially rotated 90˚ from where it should be, and the perspective view of one of the long side braces being an inside-out mirror image of the correct piece.

    Like you, I place many of the errors as a result of copying and engraving goof ups. More than I would have expected to find, I will say that.

    Have a go at the next lucarne, the one on a bias, pgs. 56~7, and let me know (by email) if you can construct the molding profile wrapping around the wall plate and on around the gable as the book is showing. I would be delighted/astonished to find that it is possible!

    Looking forward to hearing back from you. Take care,


Anything to add?

error: Content is protected !!
%d bloggers like this: