All lessonsLesson 14
The cracked buttress
Why a buttress of many stones is weaker than a block, and what a pinnacle does about it.
Lesson 5 sized a buttress by the line of pressure. Ungewitter combines the vault’s thrust with the weight of the masonry above each joint and asks where the resultant crosses it. At the base, the question becomes one of overturning: does the weight of the whole buttress, about its outer edge, outweigh the moment of the thrust? The buttress is treated as one block. But a buttress is not one block. It is built of hundreds of stones in courses, and the joints between them hold nothing in tension.
One block, or many stones?
From the Lehrbuch
„Bei Druck durch die Kante erfolgt Umsturz! Bei Druck durch die Kerngrenze ist das Widerlager gesichert, falls die Kantenpressung nicht zu gross ist.“With pressure through the edge, overturning follows! With pressure through the kern limit, the abutment is secure, provided the edge pressure is not too great.Ungewitter 1890, pp. 150–151When a monolithic block tips about its toe, all of it rotates, and all of its weight resists. The stones low down on the inner side, nearest the vault, have the longest lever about the toe and resist the most. John Ochsendorf, José Ignacio Hernando and Santiago Huerta (2004) pointed out that a buttress of many stones cannot keep them. To lift the inner, lower stones along with the rest, the joints would have to carry tension. So at collapse afracture runs from the toe up through the courses to the inner face, and the wedge below it stays behind. Only the part outside the fracture rotates, and the wedge’s weight, with its long lever, is lost.
The tool draws the fracture as a straight line from the toe to the inner face, as in Ochsendorf’s figures. Where it ends is the subtle point. Here it is the lowest fracture for which the rotating block needs no further crack: below it, at every level, the block’s own section is wholly in compression. This is the model of Ochsendorf’s thesis (2002, ch. 3), where two equilibrium conditions fix the fracture height; for a rectangular buttress it has a closed form, which the tool matches. It reproduces his worked example of a church at Goa, whose buttress resists 112 kN as a block but only about 69 kN once fractured, to within one per cent.
As a block it would hold (1.38 ×), but cracked it holds only 0.97 × the thrust: it overturns. block: 83 kN; cracked: 58 kN; cracked : block = 70 %; fracture e: 5.77 m; GSF (cracked): 0.98.
- block
- 83 kN
- cracked
- 58 kN
- cracked : block
- 70 %
- fracture e
- 5.77 m
- GSF (cracked)
- 0.98
Over the whole range of depths the cracked buttress holds only about two thirds of what the block would. A buttress that the monolithic check passes with a margin of a third may in fact be at the limit. Since the geometric factor of Lesson 13 is measured on depth, the same buttress has a lower factor once cracked: it needs a greater minimum depth.
What a pinnacle does
From the Lehrbuch
„Von grösstem Wert kann eine richtig angebrachte Oberlast der Wand sein, die auch wieder um so wirkungsvoller ist, je grösser sie selbst oder ihr Hebelarm ist.“A correctly placed load on top of the wall can be of the greatest value; it is the more effective the greater it is, or the greater its lever arm.Ungewitter 1890, p. 124„Es treten sehr oft Grundrisse nach der Art der Fig. 349 auf, bei denen unten die Aussenkante durch Eckvorlagen gestärkt ist, während oben schwere Fialenaufbauten dicht an der Mauerflucht an günstigster Stelle belasten.“Plans like Fig. 349 occur very often, in which the outer edge is strengthened below by corner projections, while above heavy pinnacles load the buttress at the most favourable place, close to the wall face.Ungewitter 1890, p. 126, Fig. 349The medieval builders loaded their buttresses with pinnacles. Ungewitter values a load on top for its weight and its lever. Seen through the cracked buttress it does something more: the pinnacle stands on the rotating block, so its weight is never lost to the wedge, and it steepens the resultant in the upper courses, where the block is narrow. The second figure keeps the buttress at 1.9 m and adds a pinnacle on top.
With 0 kN on top the cracked buttress holds 0.76 × the thrust, 71 % of the block’s value. block: 64 kN; cracked: 46 kN; cracked : block = 71 %; fracture e: 5.66 m.
- block
- 64 kN
- cracked
- 46 kN
- cracked : block
- 71 %
- fracture e
- 5.66 m
A pinnacle of about 150 kN, a third of the buttress’s own weight, turns a failing cracked buttress into a sound one, and the ratio of cracked to monolithic strength rises as well. Ochsendorf and his colleagues conclude that monolithic checks should not be taken at face value. The pinnacles, it seems, were more than ornament.
Ungewitter’s rule against Amiens
Each reference building now carries this modern check beside the proportional rules of Lesson 10: its measured buttress against an estimate of its nave vault’s thrust, as one block and cracked. The estimates are rough (the thrust comes from Table I and the pinnacles are ignored), but the order is telling. At Amiens the cracked factor is about 1.5. At Beauvais, whose choir vaults fell in 1284, it is among the lowest of all.
Challenge
The Lab opens with an arch on a tall, slender buttress. Its modern view checks the buttress against the arch’s least thrust, as a block and cracked. Find the depth at which the cracked buttress reaches a geometric factor of 2, and compare it with the depth Ungewitter’s kern rule asks for in the 1890 view.