Swiss and German Roof Framing

[D L Bahler]

I've been having a bit of discussion of Don P about roof framing, started in another thread and made its way over to PM's. He suggested I start a public thread, so here goes.

We've been talking about roof framing practices in Switzerland and Germany. He had made some comments regarding the history of the Truss in Timber Framing, and I pointed out to him that the history is quite a bit different in the Holy Roman Empire than in France and England. 

I don't want to give too much information all at once. If there's interest in following this topic I'll put up more information in the future. It's good to keep our minds occupied right now. Especially for me since the state I live in has essentially shut down, and the county I live in has officially shut down. 

The basic idea is, all of these roof systems derive ultimately from simple posted ridge beam roofs, though the history is a bit more complex than that. It's sort of a fusion of Central European Celtic tradition with some Germanic influence, built roughly over a framework of Roman practice. So you start off with a sharp distinction between the roof framing in public buildings -churches, tithe barns, castles, town halls, etc. and the rural buildings, but by the late Middle Ages there comes to be quite a bit of crossover, especially in Switzerland where a somewhat simpler take on roof framing, driven by the framing techniques developed in the massive Swiss barns, takes over. 

Anyway, enough words. Let's get to some examples. First a couple real examples, or at least diagrams bassed roughly on real world examples.



 

First is the roof framing in the church of Scherzligen in the city of Thun. The current church was built in the 10th century over a much older building. This, as you might notice, is a King's Post Truss. Or perhaps more properly speaking, it's a crown post truss. The arrangement here is typical of the German system. There are primary trusses designed something like a flat English truss, but the German systems are always built in 3 dimensions. It also demonstrates have vastly different the German approach is compared to the English. 
So we have principle trusses that support a ridge beam (and on larger roofs also purlins) which in turn support common rafters between the trusses. So in this example, the ridge post is hung from the rafters, the upper cords, and the tie beam is in turn hung from the ridge post. 
So rather than having a lower truss to support the crown post, and then in turn have rafters above the truss, the Germans will always use the rafters themselves as the upper cords of the truss. 

This roof goes back likely to the 10th century. Originally it was built so that the trusses were visible from below. Later, probably 15th or 16th century, the ceiling was boarded over. 



 
This next example comes along quite a bit later. This is a diagram of the roof of the Stadtkirche in the city of Thun, which is within walking distance of the Scherzligen Church. This roof was built in 1737 when the entire church was renovated and enlarged, the older Medieval structure having fallen into severe disrepair. 
This is a classic "Liegender Stuhl" modified to function as a truss, which is quite common in Switzerland. The vertical posts are all hung from above. In the case of the long posts extending down to the bottom cord, they are strapped and bolted to longitudinal beams. The ceiling joists are bolted to the bottom of this beam.  This creates a 16m free span underneath. 
The use of steel strapping and bolting of the ceiling joist to the hanging support system goes back pretty far in Switzerland. For example, the village church of Unterkulm, built in 1576, has a roof system similar to this, though smaller and featuring only a single center-span hanging support. 

[Nebraska]

I did not realize steel came into play in building construction that early. Interesting....

[Don P]

I think the strong aversion to steel or wrought iron in some is probably more of a modern affectation.

Thanks for posting this David, and please do continue.
I've been a bit uncomfortable with my use of the term crown post to describe any compression ridge support that is in compression. I sure don't want to start rewriting the lexicon. In one of our pm's you used the term ridge post, I like that better. It doesn't tend to confuse it with the typical definition of a kingpost, typically a tension member hung from the ridge, or with a crown post, which is typically a compression post that runs up from a tie beam to a collar beam which in turn supports the ridge. I wouldn't call example one here a kingpost since the post appears to be a compression member supporting the ridge as a prop from below. Jim or others might have a better term.

I think I do take issue with what is supporting that post but it really revolves around the members and loads. If the top chords are supporting the ridge through the webs in tension, well, the webs are connected at points of maximum bending in the top chords, so as, or if, they sag the ridge post drops. The web connections would be in tension, quite a connection difficulty. However, in English crown post tradition if the tie supports the post, again the post is supported on the tie at the point of maximum bending so the same dropping problem but everything in that path is in compression so connections are far less stressed. That tie became quite massive and was often supported by large bracing arches from the wall posts. I may be misunderstanding your description as well. Don't let me bog us down early though, keep moving if you'd like to visit this later, I know you have lots of ground to cover.

One other thing that I think will help my understanding, if you would use terms such a liegender stuhl but right beside it translate the term. I think it might help it stick or help me differentiate.

Thanks for doing this .

[D L Bahler]

I don't recall the dimensions of the Scherzligen Church off hand, but I do believe the width of the building and dimensions of the tie are such that you should not expect it to bear the post in compression. Meanwhile, those rafters are massive. Knowing the situation, it's also likely the ridge beam is lashed to the upper cords. I'm trying to find some pictures of the actual construction but can't at the moment. All information on the building leads me to believe the roof above the nave is the original 10th century construction, with a steeper gothic roof above the choir and a gothic roof on the 9th century Carolingian tower, both probably built in 1380. 

I usually haven't translated the German terms because they lack English equivalents, and their translations would just be even more confusing. At least that's my thought. The Liegender Stuhl describes something very specific. Translated you might say something like canted purlin support but this in the American tradition implies something else entirely. But I'll try to come up with something helpful. I also use the German terms because I don't really know my English terms very well. 

Anyway, I can provide some pictures of other churches mentioned. 


 
First, we have the roof of the church of Unterkulm, built in 1576. 
You can see very clearly here the medieval-style "liegender stuhl" with hanging posts in the center. These posts, in turn, hold up a beam that runs down the center of the nave. You can also see the bolts attaching the ceiling beams below to this longitudinal support beam. 
The "Liegender Stuhl" refers to the heavy leaning posts just inside the rafters, along with the collar beam connecting them and the braces forming something sort of analogous to a cruck frame. 
You also see here one form of the "clamping beam" that is very very common in switzerland. In this example, the hanging posts are notched around the beam underneath and bolted to it, rather than using a through tenon like the English and French might. In other cases, this "post" would consist of two pieces of wood that clasp the beam and are either bolted or strapped together. 
I'd like to find some examples to show, but there is a great deal of cross-pollination between bridge and roof construction in Switzerland as well. The clamping beams being one example. In many cases, especially on the big Cathedral roofs, the carpenter will include a bridge-type truss running longitudinally though the cross-sectional roof trusses, or the roof truss itself might simply be a modified bridge truss. 



 
Then we have the Stadkirche (City Church) in Thun. This is the same basic principle, only here there are two rows of hanging posts instead of one. The beams from which the ceiling joists are hung are not visible in this example due to a different method of attachment, meaning they are completely obscured by the flooring.
This roof, again, spans 16 meters with no intermediate support underneath. That comes out to about 60 feet. It's also quite steep and built to carry clay tile and the heavy snow loads of the region. 
It's a much more elegant and simple approach when compared to the complex trestle systems employed in similarly sized churches in eastern Switzerland and Austria from around the same time (18th century)
(You can also see in the background a modern reinforcing truss that has been added to give more support to the junction of the straight nave roof with the round choir roof)
To be honest, I'm not actually clear what the central hanging beams on the upper level frame here actually accomplish. 

[D L Bahler]

Let's clarify things by working from the simple forms of German roof assemblies. We'll start with the "Liegender Stuhl" which describes specifically this assembly:


 

The principle here, as with all Central European roof systems, is that the main idea is the support of purlins, which support rafters. They'll use a "truss" (in quotations because it's not always a true truss, according to the Engineering definition of that term) to support the purlins if they can't give them direct support to the foundation below. 
In Switzerland, the Liegender Stuhl becomes the most common solution by the end of the Middle Ages. 
The main idea here is, the purlin posts are canted outward all the way to the exterior walls, carrying the entire roof load out with them. This requires some pretty substantial joinery where the "Binderstrebe", which is the leaning support that carries the purlin, joins into the tie beam. On many older buildings, this is the most common point of failure. It's a very high stress joint, and on the large roofs of Switzerland it doesn't work well to seat the "Binderstrebe" directly into the tie beam.
So to solve this problem, they added a beam underneath the "Binderstrebe" that can distribute the outward thrust into the tie beams of the common rafter pairs, and also generally be joined to the main tie beams more easily. This looks typically something like this:


 
In this picture, the red members are unique to to the support assemblies. Blue are common members, and green is longitudinal members. This will make sense in a minute. 



 
Between the support assemblies, you have common rafter assemblies. These consist of the rafters, a collar beam, and the "Aufschiebling" which I'm not sure the proper term in English, but it's the kicked out stub rafters to bring the roof out past the exterior walls.

Here's what the whole assembly might look like:


 
Hopefully the color system makes it easier to distinguish between the support assemblies and the common rafter assemblies to see how the whole thing goes together. 
You can also start to build a picture of how the German carpenters build in 3 dimensions with how the "Binderstrebe" don't attach directly to the tie beam, but rely on longitudinal beams to distribute the thrusting forces to multiple assemblies and make the whole thing much more durable. 

So functionally, the Ligender Stuhl consists of 6 parts:
The tie beam, which resists the outward thrust generated by the..
(2) Binderstrebe, or leaning support posts. These are canted outward to free up the interior space and carry roof loads to the exterior walls. This, in turn, completely frees up the design of the space below. 
Collar beam, which is captured between the tops of the Binderstrebe and here is a compression member. It keeps the assembly from collapsing inward. (In German this might be called a Zange, which translated to "Tongue")
(2) Braces which in this example are designed to work in tension, and make the structure more rigid by tying the rafter assemblies in to the support framework. In probably the 18th century carpenters switched to using compression bracing here instead. 

But as mentioned, it's usually modified to include a "Stuhlschwelle" which I'm not sure the best way to translate, but it's a sill beam to carry the "Binderstrebe" instead of seating them into the tie beams. 

The main function of the "Liegender Stuhl" is to hold up the purlins, which on a simple design like this are located somewhere between the middle and upper third points of the rafters. If you need more purlins, you can either tie them in somewhere in the middle of the "Binderstrebe" or better yet, you can just stack a whole other "Liegender stuhl" on top of the bottom one. 

(For reference, I modeled the wall section underneath to match the dimensions of the Stadkirche in Thun, 16m by 35m interior space. With this simple roof example, the tie beams would need to have intermediate support, and the roof probably wouldn't be strong enough for heavy tile and the snow loads of the Berner Oberland. But this serves simply to illustrate the basic ideas behind the Liegender Stuhl. I'll modify this same model to illustrate how you adapt this basic design to meet and address these problems. The roof as shown here would however be typical of a farmhouse in the Berner Mitteland and the Emmental, where open space underneath in not a concern, but open space in the "Dachstock" or attic space is greatly desired. The size also would be within the bounds of these buildings.)

[scsmith42]

Fascinating post!  Many thanks to David and Don for sharing their knowledge.

[D L Bahler]

So now let's solve some of the problems with the Liegender Stuhl...
First, if we want to have a span this wide, we need to come up with some way of holding up the cross beams. You can't expect a simple beam to span 60 feet, after all. So In this case, an open hall protestant church, we need to redesign the roof support system. 

Like the church of Unterkulm shown above, the simplest solution is to add a row of hanging posts down the center. We hang these posts from the collar and tongue beams above, and suspend the tie beams from them. 



 

 

BUT, since we're working with 3 dimensional framing, we don't have a hanging post on every tie beam, only on the support framework sections do we have these. So instead of holding the beams directly, once again we will instead use a longitudinal beam. Here the posts clamp the beam, and the tie beams are then bolted or strapped to this beam to give them support mid span. We now have a true "truss" for the first time. 

You'll also notice that the hanging posts are suspended from a cross beam that lays across the collar beams up top. Once again, this is easier and stronger than relying on tension joinery between the hanging post and the collars.

But this still is not ideal. There is still a 30' free span on either side of the hanging posts, and in addition to this they are suspended from the collar at its weakest point. On a roof this size or any larger, this is not going to be strong enough. 
So, we instead hang 2 rows of posts from nearer the ends of the collars, and their loads are more efficiently carried into the "Binderstrebe" 





 

This has also required us to lower the purlin height, and we're quickly moving outside of the bounds of what is acceptable. For a roof this size with this sort of loads, we'll need to adjust the design a little further. 
(Note, there is an error in this model. The "Zange" and the "Binderstrebe" don't come out at the right place. Compare to other pictures and you'll see how it's supposed to look) 

As mentioned above, for large roof we typically will want more than one set of purlins. In some cases we can set a purlin in the middle of the Binderstrebe, but in this case we're out of bounds for that to be a good solution. We still wouldn't have any support in the upper half. 

But there's another solution, which is a better one anyway. We'll make this a "Zweifacher Liegender Stuhl" -or a 2-section assembly. We'll do that by lowering the collar even further, and then stacking another Liegender stuhl on top. It looks like this: 



 


 


For some perspective, the bottom of the lopwer "Zange" or collar beam is 10 feet above the tie beam in this example, and there is about 24 feet between the lower hanging posts.
This is pretty close to the solution the carpenters who built the roof of the Thun Stadkirche came up with. The principles are all the same, the the execution shown in these samples is a little more straightforward. 

It's very typical of Western Switzerland that the roofs are fairly simple variations of the basic Liegender Stuhl design. Even the massive roof of the Bern Minster is little more than a liegender Stuhl, very much in fact like the last example shown here (though as a medieval basilica-style structure, the central nave is considerably narrower than later protestant hall type churches). In the east, however, things are a lot more complicated. They're a lot more like the complex truss designs of Bavaria and Austria.   

[D L Bahler]

Finally, to evolve the Liegender Stuhl without completely modifying it into something else, another common approach on larger buildings is to add a longitudinal truss system connecting with the already existing network of hanging posts and support beams. 
These systems tend to be based on the sort of trusses you see used on Swiss bridges, and it would look something like this: 


 

 

Note that I've also added a ridge beam, and the rafters have been hidden to make the longitudinal truss visible. 

This seems to be a pretty modern addition to the Liegender Stuhl, perhaps shortly after 1800. You could think of it something like a combination of (at the time) modern ideas with a proven old design. Though this sort of bridge truss does have a long pedigree, deriving ultimately from the medieval long-span kingspost bridges (which could have free spans up to about 20 meters) Bridges with this sort of truss seem to span up to around 40 meters max, and with the addition of an arch they get up to 50-60m

Another approach, this one from the 18th century, sees the design inverted. The hanging posts are instead compression posts, and angled steel rods running opposite of the wooden struts shown here are used as the tension members. 

[Hilltop366]

 

Thanks!

[btulloh]

Thanks for working this up and posting it.  It is interesting and educational.

[Don P]

This is very cool, I'm enjoying the education, the breakdown of how it evolves just helped make it click, good job, thank you.

Between the support assemblies, you have common rafter assemblies. These consist of the rafters, a collar beam, and the "Aufschiebling" which I'm not sure the proper term in English, but it's the kicked out stub rafters to bring the roof out past the exterior walls.

"Aufschiebling" = a sprocket
From what I'm seeing the liegender stuhl in its basic form is a truss. When the hanging posts begin to drop from points out in the span they cease to be trusses. The loading is no longer axial on the collar and tongue members. That is not to take away at all from what it is.

[D L Bahler]

I guess my thinking was, in its simple form the Liegender Stuhl is not designed to hold up the ceiling/floor beams in order to create a clear span, it's only designed to kick the roof loads outward toward the exterior walls. For spans of any significant size, or if you are using the space in the attic for anything where it needs to carry some weight, you still need walls or posts underneath the tie beams to hold them up, and any time you see a wide open liegender stuhl that's what's going on. 

But when you add the hanging posts, you give the whole assembly the ability to create clear spans. But also I see how it's not a conventional "truss" in that the system is forced to work in 3 dimensions instead of 2.

So the type of roof framing obviously is a consequence of how you're using the space. In the case of a large Protestant church, you want a massive open space below, like the example of the Stadkirche in Thun which looks something like this: 



 

 
You need  a complex roof frame that is completely self-supporting. But in the case of the Bernese Bauernhaus, which is a huge complex combining barn and house (actually usually the equivalent of 2 or 3 houses) into one enormous building, you will have lots of walls under the attic space, and instead they wanted an open attic to store more hay and equipment. So here you don't need so complex a roof frame, because there is plenty of support underneath. So it depends on where you want your huge open space. The liegender stuhl is great for creating one somewhere.


But my thinking is also influenced by the German language. The simple Liegender Stuhl is classified as "Sprengewerk" which is an assembly designed to carry loads outward with compression members. But what I think of as a "truss" is what German labels "Hängewerk", which uses tension members to "hang" a beam from above. That is, my thinking is that a truss involves hanging a tension cord in some way, which the simple Liegender Stuhl does not do.
Most complex assemblies end up a combination of the two. For example, a Swiss wooden bridge will have a "hängewerk" truss in the center of the span, with "sprengewerk" buttresses at the ends to shorten the effective span of the truss. This sort of construction gives them a rather unique profile. 
The final assemblies shown employ "Hängewerk" in the form of the hanging posts, which suspend the tie beams, and retain "Sprengewerk" as the primary means of carrying the loads outward toward the walls. It didn't take much for carpenters to realize they were recreating the engineering of their bridges, so it's no wonder they started adding bridge trusses to them. 

[Don P]

The textbook definition would be that a truss resolves the member forces in tension or compression carried axially, along the axis of the members... without inducing bending in them. As the hangewerk moves out into the span it is introducing bending into the equation so it doesn't meet the definition of a truss. Really that is neither here nor there, it works. It's largely what we were talking about in one of your pm's, this is a good example of empirical, experience based, design vs my definition of a truss above that comes from rational, mathematical engineering, based design. This probably makes a sliderule smoke... but as long as you know where and what the forces are and how to resist them it works fine. Enough of that footnote, onward

 

[D L Bahler]

After the Liegender Stuhl, the next type of roof framing to discuss is the "Stehender Stuhl"
First let's review the idea of a "Dachstuhl" liteerally it translates as "Roof Chair". Looking at the systems above, we can define a "Dachstuhl" -or also "Dachträger" or "Dachtragwerk" (which would both translate to something like, "Roof Carrier") as a complex roof support assembly, which doesn't have a real analogue in English and French carpentry.

To look at the Stehender Stuhl, we have to go back a bit further and look at the ancient predecessor to Central European roofs. The 'Hochstudbau". "Hochstud" refers to a high post supporting a structural ridge beam. The simple Celtic buildings from the bronze age on were built with a central row of posts that directly bore the ridge beam, and the rafters were hung from this beam and set on the wall. Since they're not attached to the walls, the rafters would naturally tend to press the walls inward, so these structures required a tie beam that works in compression, instead of the more familiar tension beams. It looks something like this:


 
This evolved to include multiple rows of posts to support purlins mid span. The most common arrangements are houses framed with 3, 5, or 7 support beams, though in the Canton of Zürich it's cmmon for the ridge beam to be completely eliminated, and instead there are 2 rows of purlins set very near the ridge, thus yielding a 4 or 6 row support system.
Here's an example of a large purlin framed roof like you might see on the Western Swiss Plateau:


 

This introduces another concept that the Swiss would exploit, indirect load pathing. The "Hochstud" is eliminated, but all other support posts still go all the way to the foundation. The weight of the ridge is carried over to the neighboring supports by landing the ridge post on the first cross beam it encounters.

This is not yet a "Stehender Stuhl" because it's not a "Dachstuhl" at all at this point. In order to be properly classified as such, the roof framing has to be more or less independent of the rest of the structure. In contrasst, the roof support posts here are continuous, and are the same posts that frame interior walls. It will become a "Stehender Stuhl" once we work out a few more concepts.

One principle advantage of this sort of system, though not illustrated in this example, is that it's a lot better suited to framing low roof pitches than the "Liegender Stuhl" is. Roofs in the Alps, for example, will be framed with a "Stehender Stuhl" set atop log walls, and carry roofs with very low slopes (typically between about a 4:12 and 6:12)

[D L Bahler]

Since the term "Stehender Stuhl" is less precise than "Liegender Stuhl", it's a little more complex to define how it works. In addition to the evolution of the simple posted roof as shown above, the "Stehender Stuhl" also includes roof systems that are derived from Kingspost Trusses.

In the German/Swiss/Austrian context, these roofs evolve from a kingspost truss that was built to carry a ridge beam. This was something the Romans left behind when they fled across the Alps in the early 5th century, and which was quickly adapted by the new rulers of these regions.

The Roman ancestor of German true trusses looks something like this:


 

In German, you call this "Einfaches Hängewerk" which translates to simple truss, more or less, (Though einfach here more specifically carries the implication of 1 piece, referring to a single central post)

There are many very old examples of this truss in Switzerland. For example, the Schlosskirche of Spiez, not far from Thun, has a 10th century kingspost roof. A short ways east of there the village church of Amsoldingen, built at the same time, has a larger example of the same roof system. This used to be the home of a major college of canons, and so it is a very large church. 
Here is the Spiez Church:


 
And the massive Amsoldingen Church:


 

 

Both of these were built by the Burgundian king Rudolph II in the 930's, and the structures from that time are almost entirely intact. 
They represent the work of "Lombard" carpenters from Italy, as well as the influence of the French kingdom of Upper Burgundy, which ruled the region until the kingdom's dissolution in 1033. This distinction contributes to a significant difference in the carpentry of Western Switzerland, such as Bern and Basel, compared to that of Eastern Switzerland.

The basic premise here is, the lower struts are tension members that pinch the ridge post, and the tie beams below are in turn hung from it. While the joinery of the upper cords of the truss appears to be designed to work in tension (lap dovetail) in fact this was done merely to ease in assembly. The carpenters knew full well these work in compression, but this joinery was used due to the way these roofs were put together.

In this case the upper cords terminate at the tie beam. In many cases, especially in houses, barns, and outbuildings in Switzerland, they will instead lap through the tie beams and be set into the wall posts below. SOmething like this:



 
This qualifies as indeterminate framing. It's not immediately obvious what the carpenter had in mind here. The "struts" might simply act as braces, and in this case probably do because the building is small enough that the tie beam won't deflect under the weight of the ridge. But this design is used all the time, and generally it is approached by the carpenter as if it were a kings post truss. You see this design applied to much larger buildings where it is more obviously designed to accomplish an open free span through trussing action.


But, we're still working with a building the size of the Stadkirche in Thun. That is, this is a 60 foot clear span. In such a situation a single hanging support is not going to be sufficient. So the simplest solution is to extend the basic principle, and make the truss more complex.
There are 2 common approaches. You can either make it into a queen post truss (Doppeltes Hängewerk) or a Dreifaches Hängewerk. This one is very very common. Not sure what to call it in English, but it looks like this:




This still isn't really the best solution in the world. It works, but it certainly has its disadvantages. The way in which the intermediate posts are hung from the truss cords is less than ideal, and creates joints that are under a great deal of stress.
A common solution is to just add more webbing. This goes all the way to the development of what is essentially a lattice truss. A great example of this is the Pfarrkirche St. Martin in Baar, Switzerland, built in 1557. This combines a Stehender Stuhl and a Liegender Stuhl in a very complex roof system (It would seem the roof of the Zürich Grossmünster is very similar)




This building has a complex network of struts in the cross sectional trusses, as well as a long lattice truss running down the center axis of the roof lengthwise. The cross sectional trusses feature compression cords that hold up the hanging posts and land on the tie beams, as well as tension webbing that suspend the structure from the "Binderstrebe"

It's a fascinating and extraordinarily complex roof, but I'm skeptical that it's any better functionally than the very similarly sized roof of the Thun Schlosskirche.

[Don P]

I'm not going to pretend to any expertise just passing on what I think I'm seeing and some terminology based on English carpentry.
This is a very rough sketch of one of the principle rafters at Chichester cathedral in the high nave roof based on Hewett's sketches (the cathedral is currently undergoing repairs and having a new lead roof installed, pics online)



As you can see it appears to be a crown post, a post supporting a collar from the tie, or is it... usually the post is a compression member. This has additional members, what Hewett calls raking struts, they terminate into jowls in the post, I've painted them blue denoting compression which would make the post a tension member supporting the tie, (typically tension is red) if so this is I think an early example of that shift of king going from compression to tension there ~1290. I've seen pics of wrought iron tension strapping around and under truss members in other parts of that roof framing so I believe that is what is going on. I'm only guessing at what the carpenters were thinking.

In the Cressing wheat barn there are similar looking doubled roof members that run from the peak, across the aisle posts to the exterior posts, these he refers to as passing braces. I'm going out on a limb and guessing that is probably what we are seeing in your pic of the outbuilding. Rather than calling those roof struts I'm tempted to think of them as passing braces. It does make you wish there was a diary buried in the walls sometimes :D

[D L Bahler]

Right, on the outbuilding, they are functionally braces. However, the profile and design mimics that of a truss system common in the area. You can tell, on the larger examples, they are designed to function as trusses  with the ridge posts acting in tension because of the joinery and because of the way things have given way over the centuries. 
The picture was just the only one I could find off hand of the basic assembly. My guess is, that structure wasn't built by a proper carpenter. It's a bake house, probably built by a farmer judging by the crudeness of the work. A proper Swiss carpenter would have put quite a bit more finesse into it. I'd guess the builder just copied something else he was exposed to, and comes up with an inexact imitation of a common roof design.

In the larger trusses typical of church and castle roofs (there are a lot of these in Switzerland) it's even more obvious what the members are designed to do. 

If your example was south German, I would say the "raking struts" are compression truss cords, and that post works in tension. But since this is English, who knows. Those canted supports for the purlins, that seems very odd to me. There are a lot of details there that would be very out of place in a south German or Swiss roof. Fortunately for us, the Germans did write things down, and in both Germany and Switzerland the carpenters guilds held on to a very good educational tradition that preserved a lot of old knowledge. Plus you have German and Swiss carpentry manuals going back to the early Renaissance. The village carpenters, not so much. So we have excellent record of what the design principles behind church and castle roofs were, but with farmhouses and barns and bake houses, we're left to make educated guesses. 

I mentioned the use of clamping hanging post pairs in an earlier post, thought I'd modify a previous example to illustrate this concept: 


 

 
You can see how the hanging posts are all in pairs, hung from support beams above and clamped to beams at the bottom. Typically, these connections would be made with heavy steel bolts. 
This also, incidentally, more faithfully replicates a Swiss bridge truss. Virtually all of the connections in the longitudinal truss assembly would be accomplished with steel bolts, and maybe even steel reinforcing plates. 
Such a design would be within the bounds of a Baroque era Protestant church, when the prevailing style was the construction of massive open spaces, free from any internal support, in stark contrast to the nave and aisle design of Catholic churches. 

[D L Bahler]

Here is a really good example of a protestant open hall church, this the Church of Göppingen Germany built in 1618. (pictures from Wikimedia commons)


 

It's a Liegender Stuhl with hanging posts, which work in clamping pairs. In the following pictures you can see the extensive use of bolts to lash everything together.


 

 

 


This church is wider than the Thun Church at 20.91 meters (about 69 feet) but is not nearly as long. 

[Nebraska]

 Those bolts were installed in the 1600's?  What is the wood species used in the construction?    I recall reading before that steel fasteners used in wooden construction would fail after 2 or 300 years.. atrributed  to the acids in the wood.... obviously not so in these cases. Thank you for the history lesson.

[D L Bahler]

Having reviewed the materials on this specific church, the current roof actually was built in 1770. So my original info was incorrect. 

In Switzerland the roofs will be of fir and spruce, in southern Germany they will also use pine. I would imagine the failure of metal fasteners is quickened when used with oak. That's why we always use stainless or galvanized when working with oak. But in the softwoods used in Central Europe this isn't so big a deal. Very low acidity. You have the Neubrügg north of Bern built in 1534/35 that has several generations of metal fasteners intact that can be used to quickly assess the date of the various repairs and modifications of the structure. This bridge is built of fir with some oak elements. 
That said, I know it is not uncommon for maintenance and repair work on historic structures to involve replacing some of the metal elements. 

The attic space in the Göppingen Church, by the way, was constructed to serve as a granary and storehouse. That's why it's so heavily built -it needed to bear several more tons of weight than just the roof itself. There are still the two windlass-cranes that were used to lift goods up into the attic space. 

[D L Bahler]

Taking some inspiration from another post,

There is a potential disadvantage to the Swiss/German methods of roof framing. Notably, the roof systems are typically entirely independent of the wall framing. 

When faced with wind loads, this can present potential problems. First, lets envision the structure with high lateral loads from winds. In Switzerland this is a serious concern, as they have the Föhn, which is a strong wind that comes down off the Alps, and will destroy underbuilt structures. This is one of the biggest reasons why Swiss roofs are so massively overbuilt, also combine this with this weather system's ability to dump several meters of snow at a time. (Why do you think the Swiss Alps are so famous for their skiing?) 





With a roof system just placed on top of the wall frames, the walls then bear the brunt of the wind load force. This is a pin connection, with no abilty to resist the lateral force at the pace where roof and walls meet. An exaggerated depiction of this dynamic looks like this:

 



But if we brace the roof to the walls in some way, or engineer some sort of rigid connection, the walls will better absorb the forces applied by these live loads.

 

 

So there are two types of failure we might encounter. With a pin connection, the high wind loads might just push the building over. Especially since traditional means of building have no connection between the frame and the foundation at all.
If there is a rigid or braced connection, the failure would come by bending the posts. 

In the case of Churches and Castles, the walls are typically a meter or more of solid stone. This thickness is there primarily (in the case of churches at least) the make the walls rigid so that live loads on the walls and roof don't break them. 

But what about wooden structures? Typically large Swiss buildings will have interior walls that brace the middle of the frames. 
But if they didn't, what might happen? Would the bracing in the rafter plane (pictured below) be sufficient to stiffen the structure and distribute the load forces to the braced gable walls? 


 

[D L Bahler]

One solution to this question is to include a horizontal network of bracing in the roof system, like below:


 

There are a number of different profiles this bracing might take. 
I can't say whether or not it's used before this, but this appears on buildings from the late 1700's and through the 1800's right up until modern times. This happens to be the point when the buildings constructed with this sort of roof system were the largest. Before this, this sort of bracing probably wasn't necessary.

[Don P]

Edit;
Ah, I was typing slow, there's the braced plane, way cool

Typically in modern conventional construction the roof and perhaps ceiling sheathing create a semi rigid diaphragm, a "plate". This fairly rigid plate then uniformly distributes the lateral load to the braced walls, typically no more than 25' apart, then to the floor diaphragm and into the continuously braced foundation, that's pretty much the current prescriptive lateral scenario. Engineered solutions can be more exotic.

In your sketch I see much bracing in both directions into heavier members that act as collectors. There is going to be a limit as to how far apart interior bracing wall lines can be safely placed, but the same thing is being accomplished here it looks like, show me what you are seeing.

Backing up and way aside. Someone made a comment about being surprised at the early use of iron fasteners, or, that was the journey I went on. I remembered nails and Romans from a couple of thousand years ago and had the day yesterday to read a book I've meant to for some years. DL had mentioned at some point that the Swiss had not "lost" roman information in the early Middle Ages the way much of western Europe had. This particular book was safely stored in a Swiss monastery, "we" rediscovered it at the dawn of the renaissance, so it was lost to western Europeans for about a thousand years. It inspired many men of that time who's names we all know. It was written by the Roman architect Vitruvius. I'd like to point to something in there, a reference to bolting together a trispast, a hoisting machine. Second sentence under "hoisting machines" here;
https://www.gutenberg.org/files/20239/20239-h/20239-h.htm#Page_285
This was a reference to structurally bolting something together about a hundred years before Christ. They have found a nut from around this time which is in a museum in Germany. Just like at my house they can't find the bolt. Many references to nailing various things together. Sorry for the side track, back to where we were.

[D L Bahler]

Looking at the roof system without the braced deck, we have this:
First a side view:


 
And a top view:


 
In German this is "Windrispen" which is more or less, Wind Bracing. It is designed with resisting lateral loads in mind, in both horizontal planes.
The "windrispen" in a "Liegender Stuhl" roof in Switzerland generally has this large X-shaped profile as seen here, and the braces themselves are designed to work in both compression and tension (so, basically, the braces are bigger than they need to be, with a strong enough dovetail lap to act well in tension, and still have enough of a shoulder to act in compression) 
In Germany you might instead see "Kopfband" braces -basically knee braces- that are set in plane with the supports. These are only useful in bracing against wind loading on the gables. 

The Library of Sankt Gallen is a pretty fascinating place. If I recall it's the world's largest collection of ancient and medieval books. Also more or less the birthplace of harmony in Western music, presumably under the influence of the Alpine yodel. It has German-language Bibles that predate Martin Luther by centuries. 

We've talked about old trusses before. It's possible the oldest truss in existence is actually in Rome, the support framing for the roof of the portico on the Pantheon. It could be a queen post truss, but it's not entirely clear. 
From what I've been able to learn through my attempt to study it, this support structure is made of hollow beams of hammered bronze with heavy bronze reinforcing straps. 

It's certainly true that Roman truss designs were retained in parts of Europe. Italian craftsmen definitely held on to the technology, and of course we shouldn't forget the Byzantines. I think you will find that some knowledge of the truss was retained in Italy, Spain, Greece, Switzerland, Austria, and even into parts of France (like Provence and Burgundy) and southern Germany. I think the real issue is that two distinctly different fields of carpentry departed from each other rapidly in the Middle Ages, before the "high" architecture of the (mostly Italian) masters eventually merged into the vernacular traditions near the end of the Middle Ages, and the English and Germans then replacing all of it with mathematically designed trusses in the 19th century. Anything post Bismark is heavily industrialized, and it might be fair to say that his modernization campaign dealt the death blow to traditional German carpentry. 

The Burgundian architecture in what's now Switzerland was definitely built under the direction of Lombard craftsmen -the Burgundian Kings looked to Italy instead of France as the center of their cultural and religious life, with Lombardy being their immediate neighbor and Milan being the most important city of the day. The French kings were their rivals, and when the Burgundian kings died out they ceded their lands to the Holy Roman Empire rather than to France. (which is why I don't speak French today, my family being from the part of Switzerland that was formerly Upper Burgunday) A lot of fairly low-sloped kings post roofs that were relatively narrow -the Lombard style church has a high nave, and two side aisles with separate shed roofs, 3 apses. Something like the Spiez church is a stereotypical Lombard design. This is in stark contrast to the more German style, which typically has a single roof, that you see in eastern Switzerland. I think it's fair to say that these more German style church roofs don't display the same connection to Roman trusses as their Lombard counterparts. The Fraumünster and Grossmünster churches in Zurich are both good examples of Romanesque German style architecture from just a little later. 

[Tom King]

Thanks so much for this thread.   It goes on my One of the Best Ever list.