100 Year old Large, Conical Wood-Framed Roof - Compression Ring

Wow! Considering the scale that is impressive and gutsy!

Wouldn't that ring beam be in tension under gravity loads and resulting roof thrust?
Every roof sheathing board should also taking on some amount of tension, and probably do the majority of resisting thrust. I would think the connection of these boards at the hips is critical for this to work, especially under unbalanced snow or wind loads.

Is there a pattern to which struts are buckled? I am wondering whether the buckling is due to a distorted cone shape due to unbalanced loads. Part of the roof may have snow on the shadow side, or winds funneled.

Bracing the struts could help (or adding a 2x to make it a 'T' in section. Or adding tension ties across the space at the hip beams.



Eric McDonald, PE
McDonald Structural Engineering, PLLC

The inward lateral component of the bending deflection in the main rafters is larger than the outward deflection at the top of the columns due to thrust.
The columns are pulling inward at the top and pushed outwards where the struts are, which kind of lines up with the analysis.

I'm sure the sheathing board do take outward thrust, but if there isn't that interior compression ring, the main rafters have bending stresses in the 5000-6000 psi range. With the interior compression ring, it's mid 2400 psi plus minus a bunch, which seems much more realistic to me.

We are definitely repairing/replacing the struts. Trying to figure out of if the rest of the structure needs some retrofitting.

It's prior use was Carousel. So it was an unheated open-air structure, so I'm sure it has seen code-level snow loads before.

what is your snow load?

love seeing these older ones...
at first pass, it looks like the "ring" just outside the fan in the photo, is a series of simple span beams, which, in turn, support the 2x6 rafters. I don't see it functioning as a compression ring (if that's what you were referring too ??)
This simple span beam has a significant reaction felt by the main rafters 3.5x10" But this reaction appears to be about 2' or so away from the 3.5x6.5 'strut'... therefore, much of this load goes into the 'strut' without inducing nearly as much bending as it would without the strut.
So, I'm seeing the 3.5x10 main rafter as continuous over three supports, one at the column, one at the 3.5x6.5 strut and one at the peak.... plus axial load.

The order of construction, I'm guessing being, 1) main rafters, then 2) the struts , then 3) these intermediate simple span beams, and then 4) the 2x rafters and sheathing

you mention "The only issues are some bowing/buckling/warped struts."

this is the first place i'd expect challenges in this structure...a long slim compression member doing lots of work..

That's certainly a magnificent structure!

I think I see the following:

1. Two ring beams, one at the periphery and one at mid-span. The one at mid-span seems to have some type of splicing member going under the main rafters that might provide it continuity. It would be interesting to know how it's connected.

2. The 13'-6 external frame provides some degree of bracing akin to flying buttresses. And in the process, along with the 3 1/2" x 6 1/2" strut, forms something that mimics or evokes a hammerbeam "truss". There might be some combination of those two elements, perhaps, but certainly it appears that frame provides some external bracing.

3. Finally, is there any degree of fixity to the bases of any of the columns? Probably not, but I thought I'd ask.

Reality is more complex than it seems.

That's a beautiful roof, & the original builder probably did it like he'd done other things, only bigger. Probably no design went into it, in the sense that we mean design.
Hip roofs can be pretty amazing; the jack rafters (I think you're calling them common rafters) framing into the hips (I think you called them main rafters) are actually self-supporting if directly opposing and restrained at the bottom, so a hip rafter is merely a framing convenience. A conventional hip has rafters meeting at 90 in plan view, whereas these meet at 15 degrees. Obviously the self-supporting aspect reduces to zero at some point between 90 degrees and zero, but not linearly and maybe not in any way that you can quantify. This means that the hip rafters are not bearing all of the weight of the jacks, and the hips are acting in both bending and compression. The rim beams must be in tension, which means there must be good connections between the 6x6 rim beams at the columns, and between the triple 2x12s at the columns too (how are the angle clips fastened?). The 1" cable must be helping a lot, but looked at logically it would bend the tops of the columns out, not in, unless the struts/braces are carrying a lot of compression. You've said some of them are buckling, so it may be that they are taking much more compression than one would anticipate.
After all that nonsense, why are you analyzing it & what do you hope to show? Is it a change of use that requires some sort of proof that it works?

The dims are a bit off in the section - The main rafter is 3.5"x9.5" actual, the strut is 3.5"x5.5", the rim beam is 7.5" horizontal by 5.5" V, the cable is 5/8" and the the strut is closer to 9 feet than 12'-6".


Triangled - The ground snow load is 50psf, the roof snow load is approximately 29.2psf. The intermediate simple span beams lie in the same plane as the main rafters. They have these little stubs that are bolted to the main rafter the support then ends of the simple span beams.

I did treat the main rafter as a continuous span, and as such I still get about 25 ft-kip bending moment without the compression ring. The member clearly would have failed under that load, so my analysis is too conservative.

5 out of the struts are buckled laterally. We will probably replace the buckled struts and provide bracing at the mid, or 1/3 points (whatever is necessary). We also want to maintain the look. It's definitely a beautiful wood structure.

Archie - The periphery beam is just picks up the gravity load from the common/jack rafter tails. It's not braced properly to be able to resist out of plane bending. Much of that thrust is either picked up by the plank sheathing, or transferred to the column and picked up by the lower roof and the 5/8" diameter steel cable.

The foundations are another issue altogether. There's an unheated crawlspace, the column footings don't quite reach the frost line but do not appear to have heaved. The column bases just rest atop concrete piers without fastening.

The floor framing and columns for the roof are completely separate, the columns pass right through the floor framing. I'm guessing they replaced the original floor framing because the column and floor framing foundations are very different. Also, all the floor framing footings literally bear on grade. Frost is at least 3' in the area.

Old Building Guy

I understand what you are saying about the rim beam and jack rafters. The fact that rim beams larger dimension is horizontal makes me want to take a closer look at the column/rim beam connection. It didn't look substantial, but I'll have to go back through the photos. I could have missed something.

The only issue I see is that 4 out of the 6 jack rafters are supported at least one end by the transverse roof beam which is putting a large point load on the main rafters.

The cable appears to have been installed at a later date. I believe it should have been placed closer to the strut connection than it is.

The structure is used as a large dining hall. Some time ago they added what sounded like SIP panels on top of the roof. In addition, they added that large fan. I'm not sure if anyone checked the structure when it changed use (when it was insulated). I know nobody checked the connection of the fan to the roof. They want to replace the roof fan, and the suppliers won't install a new fan unless an engineer assumes liability for the connection.

Also, I don't have my notes in front of me, but one of the maintenance men had concerns over the columns (large check cracks) and the foundations. While checking all this out we noticed the struts are buckled and mentioned there may be some issues with the roof framing. And now here we are.

I know I won't get a perfect analysis, I just need something conservative enough to be safe, but realistic enough where I'm not saying "this should be in a flaming heap yesterday". The midspan compression ring helps me out.

It gets evermore interesting and I suspect OldBldgGuy was correct when he said the original builder probably just built a larger version of what he had built before. The more I look at it the more I think the sheathing probably contributes a lot to the overall integrity of the structure. A giant inverted ice cream cone, as it were.

Being used for a dinning hall, you say? Hmmm...I'm sure by that you mean a horse dinning hall or the like, because once you add high-occupany and potentially a school...well, sometimes it's hard to make those folks who read and write the building codes understand that it's held up for 100 years. It's a delicate spot to be in and there are probably no easy answers. Perhaps the "Existing Building Code" might help.

I didn't notice the cable before but it looks like a good idea. A retrofitted tension ring near the brace. Would it be worth adding another one at the base of the shell itself, next to the 6x6's? I've heard of modern coliseums being built using that principle to resolve the thrust from a concrete shell.

I think that giant fan is a monstrosity that detracts from the aesthetics of the structure. If they really needed the air movement several floor or window mounted units might have done the trick. And if aesthetics aren't important then just tear the building down and put up a metal building that meets all current code requirements. Just my opinion.

I agree it likely is behaving like an a cone, which is fine. A cone will still have bending moment though which is the issue. I believe there needs to be curvature in the roof like a dome to resolve the bending moment.

Yes the fan is ugly. They actually like the fan for the aesthetics, not the air movement .

They have events there and everything. Idk if wedding receptions, but stuff like that. The change in occupancy and everything is long done. It's been in use for a number of years as a dining hall. So really if we say its OK its OK. I just want the calculations to be in the same ballpark as reality so that we're comfortable enough with say just replacing the struts and leaving most of the roof intact.

Well, since it's 100 years old the wood could prove to be *much* stronger than what we now have. Some of that old virgin growth stuff is. It might be worth taking a small coupon and having it tested. That might yield a surprising amount of capacity.

And so they like the fan? Ok, well, it just goes to show that tastes differ.

i'm getting a significantly smaller bending moment from rough calc, looking at main rafter as continuous over three supports and supporting 4 point loads (from photo).

I'm not crazy about the fan either but I thought if I ignored it, it would just fly away.

What about Archie264's suggestion of another cable? The existing one does look much too low, can you put one at the eave, outside the 6x6? That's where I'd want to see it if you were to do anything.

In rereading my first post, I should clarify & correct a couple things: I should have said 30 degrees, 15 is between the rafter & the hip. And obviously the pitch of the roof is also part of it; 2:12 is drastically worse than 12:12.
When I said "after all that nonsense" I meant my own ramblings, not your questions.

Archie264

That's a great idea and I have heard of the concept that old growth is much stronger than new growth, although I haven't done much research into it. In fact, we can likely do this with some of the struts if we do go forward with replacing them.

Triangled-

My analysis is slightly conservative, I'm looking for order-of-magnitude. But I don't think I'm too far. I'm going from memory since I'm at home but here is what I remember from the analysis:

I get 290ft^2 actual area in the triangle, and 243ft^2 projected.

That simple beam is approx. midspan of the main rafter. The area above that simple beam is 1/4 of the total area and the area below is 3/4A. The area below, approximately half the load get transmitted to the simple beam and then the main rafter. The other half goes to the eave.

So I get a 29.2psf*(3/4)*(243)(0.5) + 10psf*(3/4)*(290)*(0.5) = 2661 SL and 1088 DL

Above the transverse beam, half the rafter load goes to the midspan beam and the other half goes to the main rafters

29.2psf*(1/4)*(243)*0.5 + 10psf*(1/4)*(290)*(0.5) = 887 SL + 363 DL

Total point load at midspan of main rafter = 3550# SL + 1450# DL

The other half i just distributed over the top**. Portion of the main rafter in the form of a triangular load peaking at the transverse beam.

120 plf SL and 48 plf DL

Also, don't forget that the strut is closer to 8'-6" than 12'-6", the drawings haven't been updated since our most recent site visit.

** It may be the case that all of the load goes to the midspan beam as the rafters may behaving like a ladder rather than a simple span beam for this upper portion.


OldBldngGuy

Yes I definitely agree with doing something with the cable, or adding another. I think if we replaced all the struts, and placed them closer to midspan, it might alleviate my worries and solve the problems. The struts need to be replaced anyways. Some our buckled out of plane as much as 8 inches.

A nice 4x10 or whatever size necessary strut with some bracing at midspan would look pretty good. I think I'll try and stay with timber in lieu of engineering products to keep the look unless necessary.

I have designed a 60 ft diameter pavilion and used a steel HSS compression ring towards the top. The tension ring at the perimeter columns was a glulam beam. How are the main rafters coming together at the very top?

Awesome looking structure!!

Seems like adding a compression ring will increase the tension being forces. (think stiffnesses)
That might be needed but I think I'd definitely want to improve the tension ring (cable) position.

Sometimes in old timber buildings, adding wood members is too cumbersome. I might consider steel with the expectation that one can "look beyond" the steel. It is obviously the modern element but it takes up less space and might not obscure the original elements as much as more bulky timbers might. That seems like a difficult choice/balancing act in the particular case.

BTW - I'm not persuaded by the "100 years" argument.... just sayin.

Snow loads are not always uniform over the entire roof area. In some roof shapes, snow may be removed from the windward side and dumped down on the leeward side. This could result in unbalanced snow load, further complicating analysis of the existing roof.

BA