Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Complicated Hip roof 4

JStructsteel

Structural
Aug 22, 2002
1,350
Project is re-designing a roof that was damaged in a fire. Whats the best approach for design? Model in software? Obviously what they have works, so is it worth spending the time, or replace in kind with some simpson hangers, angles, etc?
1731192126775.png
 
Replies continue below

Recommended for you

I usually design roof framing with obvious load paths, as opposed to what's shown here. So, if it was me, I would not reuse this plan. I can somewhat understand the opposing argument, though, that the roof was built like this and "works," but I'm not sure I could justify it from an engineering perspective, at least not without a very complicated analysis.

For example, how are the top of the valley beams being supported? The one towards the bottom of the page has no clear load path that I can see. I know this type of thing can be debated endlessly, as there's strength due to the roof sheathing (folded plate theory, etc.), but I would want a post or another beam supporting this point.

If I had to use the exact design shown here, I would probably start by modeling it in RISA3D or a similar software. In general, I'd expect the hips to be in compression with there being an outward thrust at the building corners which would need to be resisted somehow. Maybe you could rely on the sheathing to some extent, although then it starts to get really complicated.
 
I’d base my decision on two major factors:

(A) Project Fee: If your fee is healthy then you probably should redesign since that would be in most agreement with what one ought to ethically do. If it’s a light fee, then I would tend towards replacement in kind with the caveat of B.

(B) Loading History: I would only replace it in kind given a low fee if the roof had seen a design level event. If the area has never had snow / wind / whatever near modern design levels, I would not just replace in kind regardless of fee. In that case we cannot possibly know just by the fact that it stood for this long how it would perform during such events.

In general, I’m a big fan of load testing, and if it passed muster during a design event, I would trust that more than I would trying to analyze it using our simplified methods / tools. But if not, then a redesign (or at least reanalysis) it is.
 
I do a huge amount of remediation and disaster work. Assuming its possible in your area I highly recommend replacing the roof with a truss system. It will likely lead to needing to do some footing extensions where the girder trusses are but you get a far more predictable and reliable diaphragm long term. In my area its also 20-30% cheaper than a hand framed system that meets code requirements.
 
I did a roof replacement after a fire, similar to this. I redesigned the whole thing because I felt that the original design was very empirical but wasn't engineer-y enough. My design costed more money. Architect and owner weren't pleased, and we got into a month-long argument about it.

I stay away from these jobs now, because I will inevitably be blamed for sticking up for myself. But if I had to do it again, I'd do the same exact thing. I would design it to what's correct.
 
How much of the roof is damaged? If it's a section (say less than 33% or not more than one major section of roof), you could probably get away with replacing members. Otherwise, replace it with a truss system. Simpler, cheaper, and up-to-date.
 
I’d base my decision on two major factors:

(A) Project Fee: If your fee is healthy then you probably should redesign since that would be in most agreement with what one ought to ethically do. If it’s a light fee, then I would tend towards replacement in kind with the caveat of B.

(B) Loading History: I would only replace it in kind given a low fee if the roof had seen a design level event. If the area has never had snow / wind / whatever near modern design levels, I would not just replace in kind regardless of fee. In that case we cannot possibly know just by the fact that it stood for this long how it would perform during such events.

In general, I’m a big fan of load testing, and if it passed muster during a design event, I would trust that more than I would trying to analyze it using our simplified methods / tools. But if not, then a redesign (or at least reanalysis) it is.
Its about a 100 year old building, so seen its share of loading.

There are a few interior walls that can be added to make bearing points, and the front hip can be replaced with a gable.
 
I do a huge amount of remediation and disaster work. Assuming its possible in your area I highly recommend replacing the roof with a truss system. It will likely lead to needing to do some footing extensions where the girder trusses are but you get a far more predictable and reliable diaphragm long term. In my area its also 20-30% cheaper than a hand framed system that meets code requirements.
This will be a livable space, so trusses not an option. There are some interior walls I can use as bearing, so that will help.
 
This is a small roof. If the attic is livable space, the roof slope is probably pretty steep. I imaging the thrust loads aren't Earth shattering in their magnitude.

A hip roof, despite what many of want to believe, can stand up just fine on its own without supports where the hips meet the ridge, so long the proportions are reasonable and the thrust is balanced and resolved. In other words, the hips react against the gable roof on both sides in much the same way the two sides of the gable react against one another through a ridge board. You just have to tie them at bearing - usually done by flat 2x4s across the tops of the ceiling joists, but also common to just let it hang there and the ceiling holds together (the hips don't have uniform thrust so it works a little better than a gable).

In this case, you're not quite balanced. But you're pretty close. I'd start by figuring out how out of balance you are and decide where to go from there. The tricky part is the valley...those generally need some sort of support since they're being pushed down and out of the roof, not pinned into by compression. But even there...it may be able to handle it. You can either run it by hand or model it. If you model it, make sure your hips are pinned at the corners since you'll have a wall in each direction to resist the movement. (It's a pretty common modeling mistake for this kind of roof.) The other alternative is go to go with a folded plate analysis, but that usually generates loads that are pretty unreasonable for most houses.

The other thing to consider is the strength of the materials. You may feel comfortable with the arrangement/layout, but the material you're substituting may not be as strong. Be sure you understand the species of wood from the original construction and roughly what it's modern grade would be. Then, look to see what the allowables are for that species. And I don't mean species group...I mean species. (Southern pine is a group, not a species.) To really replace in kind, what goes back needs to have equivalent strength.

Also, remember that 100 years doesn't mean it has ever seen a design event.
 
PhamENG
Thanks.
So breaking this down in parts, the ridge board is just that, a ridge board. No real reactions on the hip beams. The reactions from the hip beams at the ridge, is going down the rafters closest to the joint?
 
OH.

MY.

GOD.

Here goes an hour of my day.

/rant mode

I can't believe KootK hasn't jumped on this yet.

Obviously what they have works,
No. No, NO. NO. All caps 968 point font size NO. It's most certainly not. You simply cannot make that statement***. MAYBE what they have works.

*** Unless you're some kind of autistic savant who can look at a framing plan and KNOW 100% that all the calculations are going to work out, the wood species and sizes that are in place, the connections, etc. etc. etc.

This looks a lot like hand-framed 1950s or earlier framing. That really indicates pre-code in most parts of the country. Your specific comment about "model in software" (oh good luck with that~) indicates you're not able to confidently eyeball this and determine it complies to the current codes. You must "trust but verify", you are entrusted by the public with life-safety, you are not a carpenter who does not know any better getting beyond their depth and replacing "as-is" deficient 2x4 floor framing spanning 10 feet, you are an engineer.

Furthermore, you can't copy another engineer's design (if it even was engineered versus your average "pre-code" roof framing), unless they are under your responsible charge and direct supervision and you were involved in the decision making from the outset, which is quite likely not the case.

In which case, to re-use the design by reissuing it as-is, you're on the hook to either a) provide prescriptive framing and bypass your engineering skills or b) do the calculations on the existing to the depth that you can confirm the members are all appropriately designed per the standard of care, and the connections (you specify) work, to the standard of care.

A full replacement with trusses may be less expensive and also get the building re-occupied faster, in that case you are on the hook for the trusses to some degree in that you specify the loads, designate bearing points (and widths and species for the top plate) and the connections are "yours," you review the shop drawings and the sealed truss designs from the truss design engineer, and the headers underneath the trusses, as applicable, and the walls underneath the trusses, as applicable, are confirmed to work.

Most engineers at some point will stop this analysis, but where you stop is on you to defend (or rather, on the engineer your insurance company hires to defend you) if there is a problem. (i.e. if you stop at the foundation no engineer can "out-do" you, and claim you did not meet the standard of care, but I'd expect most engineers would stop somewhere around the bottom plate to the floor below.

Your drawing has an "EXSIT" typo, too.

I’d base my decision on two major factors:

(A) Project Fee: If your fee is healthy then you probably should redesign since that would be in most agreement with what one ought to ethically do. If it’s a light fee, then I would tend towards replacement in kind with the caveat of B.
Maybe you don't mean it that way, but LOOK at what you WROTE.

"you probably should redesign since that would be in most agreement with what one ought to ethically do."

EYES bulging out of my head. This is a MAJOR ethical lapse in judgement. The statement itself indicates you are aware fully of the fact that you, even by your own admission, are not meeting your ethical obligations.

Just to blast this out, I really don't want to see anyone EVER suggest that you approach health, life-safety and welfare off a PROJECT FEE. If you underbid it, that's on you, you MUST NOT compromise life-safety because you ran out of budget, you can perhaps ask for more fee, or pass on the job, or put appropriate special inspection on something like an epoxy anchor to improve the design values some, (not with the expectation of getting paid that work to offset some of the bleeding, because that surely looks like either enrichment or conflict of interest), but under no circumstances can you phone in engineering and do half the job because you agreed to do it too cheaply. Take your lumps, learn your lesson, eat your oatmeal and do a proper job.

Look at the FBPE Haug disciplinary action. The guy "designed" masonry walls and foundations for a church with zero calculations. That collapsed TWICE.

Also, this Loading History, I'm going to have to go off on that idea as well. Because a structure has stood for 100 years is NO evidence that it's experienced a design level event (especially in a seismic zone, this entire premise is totally refuted by California's 1986 URM law and subsequent mandatory reinforcing/retrofit projects).

If the existing is constructed wrong, or defectively, that is the exact opposite of reality, particularly when it comes to a stability-induced failure, i.e. lots of things are stable when they have zero load on them besides their self-weight, up here in the North, older roofs that don't "math out" on the design side probably never saw design level snow because the air leakage and the heat flow through the under-insulated ceiling melts the snow so it never reaches a design level event. Hypothetically in a region where snow load minimums apply, maybe, as those loads are intended to "cover" a one-time extreme snow event, i.e. minimal opportunity for melting or no melting, then maybe, but where snow loads represent an accumulation over a season, NO WAY.

Imagine if you "designed" a multi-story deck this way. It's never seen the design loading that's why it hasn't collapsed yet. I'm sure most engineers at some point in their career have come across something they are a bit mystified it hasn't fallen down, regardless of how old it is, and yeah, if it's never seen the design loads, that's a potential argument, but to reverse it and say a) because it hasn't collapsed it HAS seen the design load, that's simply speculation (technically, I think this is a contrapositive statement). Maybe if you can find enough weather information, maybe. Provided it's a wind/snow issue that controls the design.

Boy am I glad I took typing and word-processing in high school.

/rant over.

and five gold stars and two snaps up for milkshakelake and ENG16080. Welcome to the standard of care.

I usually design roof framing with obvious load paths, as opposed to what's shown here.
Now that I'm done with my little broadside at enable, I can agree with the following:
In general, I’m a big fan of load testing, and if it passed muster during a design event, I would trust that more than I would trying to analyze it using our simplified methods / tools. But if not, then a redesign (or at least reanalysis) it is.
This. But at the residential scale it's simply not practical (I think Algo center did load testing on their parking deck, for a large area).

Plus, given the fire, you have to half-a dollar sign dollar sign it and then test it. What happens when it collapses during load testing (as here in the States that's generally for more than 100% of the design load). This is much more of a rabbit hole that's never used, and it's "never" used for a reason, as you can actually destroy what you are testing to prove it's OKAY. That's "fine" for a component, so to speak, like a guardrail, as you're proving it was unsafe and gosh, darn, turns out it's unsafe, as it breaks free of the anchorage, oh well, (there's a longer story here, from personal experience, ask later), but for a roof, that's a partial collapse you are playing with, and it's hard to load it, too, when it's pitched.


Now, just to soften the tone slightly, a graphic..... 2015 International Building Code

2015 IBC - 1708 In-Situ Load Tests.jpg
2015 IBC 1708.3.2 Load test procedure not specified.jpg
This is a small roof. If the attic is livable space, the roof slope is probably pretty steep. I imaging the thrust loads aren't Earth shattering in their magnitude.

1731366556223.png

Steeper roofs produce lower thrust loads in the rafter ties (note a rafter tie, in the bottom third) is NOT a collar tie (in the upper third). Plus the OP said it was livable space. I feel like I've been in this building, metaphorically speaking, but my building was WAY less convoluted and the valley and ridges weren't structural (as there weren't any valleys or ridges).

Ridge boards - speaking loosely, are "mounts" for rafters and are not structural. They match or exceed the rafter depth (ideally), and allow some slight offset between the rafters on each side). I've seen roofs without them, (see above), rafters then must align, and given the thrust load, well, you know.

Ridge beams ARE structural and they support the rafters. They need supports at the end, like posts or a header, etc. etc.

(probably repeating something already said, there)

Now for the FAQ attack:

2008 - Raised Rafter Tie force calculation [ Journal of Light Construction ]



 
I did look at what I wrote, as, well, I wrote it. The implication of what I wrote was that the only circumstance under which I would not reanalyze it in detail is if it had undergone design loads and came out relatively unscathed, and even then, if the budget was healthy I would reanalyze it anyways. If you think that's a problematic stance, cool.
 
I didnt want this to become something else, clearly it has.

Admins, can you delete this thread.
 
My vote would be to not delete this thread (not that I have a vote here). I find it beneficial to see the differing opinions. It's the best chance to actually learn something new.

(No disrespect to OP though if it does get canned. It's their call.)
 
@lexpatrie - Thanks for that post. I feel the same and have had some fights with old timers in my area over these concepts. Felt too new to have a rant but you nailed it.

I've had a senior engineer in my area say that vaulted rafters to a ridge board with NO TIES was acceptable because the wall acts as a beam once it's sheathed and the perpendicular interior walls tie to the exterior walls to prevent them from pushing out. After some shocked silent thinking I asked if he would control the interior to exterior wall joint and have it be verified. He responded that wasn't necessary because there are millions of roofs like that in the world. This was a 2nd floor roof.

The exterior walls were 6 degrees out of plumb and the ceiling and wall sheet rock was cracked all over the place. Customer accepted his solution due to cost of adding engineered ties. Snow load at the property is 35 PSF, I was reviewing it in the summer.

I dropped a note to building safety but they said the other guy stamped it so it's all good.
 
@ lexpatrie & PlinkoEng,
Not an old timer here, and not looking for a fight or dramatics.

Just trying to get some advice on the best way to handle this roof re-frame w/o it being trussed.
 
@ lexpatrie & PlinkoEng,
Not an old timer here, and not looking for a fight or dramatics.

Just trying to get some advice on the best way to handle this roof re-frame w/o it being trussed.
When I do it I will typically use ridge and hip beams to support the rafters. At the ridge/hip connection I king post down to a bottom chord beam. If I can ill try to get 1 or 2 interior load points for the bottom chord to keep the size realistic and avoid exterior footing extensions (easier to put them in the crawl if possible). For connections I will use off the shelf Simpson connectors or 5/8" to 3/4" all thread thru bolt connections. I'll also use both in combinations. I will use ceiling joists as ties where possible.

It's a pretty overkill solution but it's easy to prove and I know it won't fail. Easy enough to also build in the floor system once you have a predictable roof system. Typically I do all that work, customer doesn't like the cost and then goes back to trusses like I suggested at the start. Attic trusses can also be used for the living space.
 
When I do it I will typically use ridge and hip beams to support the rafters. At the ridge/hip connection I king post down to a bottom chord beam. If I can ill try to get 1 or 2 interior load points for the bottom chord to keep the size realistic and avoid exterior footing extensions (easier to put them in the crawl if possible). For connections I will use off the shelf Simpson connectors or 5/8" to 3/4" all thread thru bolt connections. I'll also use both in combinations. I will use ceiling joists as ties where possible.

It's a pretty overkill solution but it's easy to prove and I know it won't fail. Easy enough to also build in the floor system once you have a predictable roof system. Typically I do all that work, customer doesn't like the cost and then goes back to trusses like I suggested at the start. Attic trusses can also be used for the living space.
This is an existing roof, being replaced. They want to keep the same basic framing and space inside, thus limited places for posts. At the top of the page, no place. At the bottom at the end of the ridge board, I can post down, and also change the hip on bottom left to a gable.

So that hopefully simplifies the framing, the dormer on the top right complicates the load path for the hip beams, they have to carry the additional load from the valley of the dormer.

Its almost like I want to move the valley beam from the top dormer to be opposite and in line with the left hip beam.

1731376024555.png
 
This is an existing roof, being replaced. They want to keep the same basic framing and space inside, thus limited places for posts. At the top of the page, no place. At the bottom at the end of the ridge board, I can post down, and also change the hip on bottom left to a gable.

So that hopefully simplifies the framing, the dormer on the top right complicates the load path for the hip beams, they have to carry the additional load from the valley of the dormer.

Its almost like I want to move the valley beam from the top dormer to be opposite and in line with the left hip beam.

View attachment 875
I would definitely agree with that simplification of geometry. I would also tell the client it gets them more height and useable width in the dormer. Value engineering is a sweet sweet phrase.

I live in a pretty code heavy place so when a client tells me they want a straight replacement in something pre-code that is empirical, or even indeterminate, I have the luxury of just saying no. At the end of the day it's my stamp and I don't use it unless I'm 100% comfortable with the solution.
 
@lexpatrie - Thanks for that post. I feel the same and have had some fights with old timers in my area over these concepts. Felt too new to have a rant but you nailed it.

I've had a senior engineer in my area say that vaulted rafters to a ridge board with NO TIES was acceptable because the wall acts as a beam once it's sheathed and the perpendicular interior walls tie to the exterior walls to prevent them from pushing out. After some shocked silent thinking I asked if he would control the interior to exterior wall joint and have it be verified. He responded that wasn't necessary because there are millions of roofs like that in the world. This was a 2nd floor roof.

The exterior walls were 6 degrees out of plumb and the ceiling and wall sheet rock was cracked all over the place. Customer accepted his solution due to cost of adding engineered ties. Snow load at the property is 35 PSF, I was reviewing it in the summer.

I dropped a note to building safety but they said the other guy stamped it so it's all good.
We had a guy around here like that. I have made a lot of money fixing his errors. Fortunately he has retired.
 

Part and Inventory Search

Sponsor