Hello all! This is my first post. I’m a ‘part time’ builder (I do one house a summer, usually for myself or relatives). I want to build a distinctive home that can go up with the relative ‘ease’ of a colonial but is set apart from the sea of cookie cutter homes in my neighborhood. I’ve settled on a design that is basically an ‘L’ shape. The home is set apart by its mansard roof. I know that style doesn’t appeal to everyone, but my wife & I really like it. We currently live in a 2-story colonial and would like a little more space without a larger footprint. The mansard seems to cater to all of our needs with it’s historic appearance and by facilitating a 3rd floor. Unfortunately I haven’t been able to find any details on framing this type of roof. I have seen some on the ‘modern’ mansard, but they have an overhang which is not a traditional mansard. They also have a reverse dormer. Basically, they’re hideous! Of course the house will have an overhang to limit rainwater sheeting, but it’s not created by the steep vertical of the lower section of the roof, but rather an extension where the roof attaches to the top plate.
I’m assuming there is a perimeter knee wall that is set in from the edge of the exterior wall to carry the upper hip roof’s load. I’ve read about creating the lower portion of the roof (about 15 degrees) by building a tilted wall using wedges that is installed on the exterior plane without the knee wall, but that just doesn’t seem sound to me. I did a search here but didn’t turn up anything. I am going to conact my local historical society to see if they have any photos of the construction process of any of the local Second Empire homes, or perhaps even a tour of one. Hopefully one will have an unfinished top floor so the framing members are visible. But even then that may not yield a sound method that will pass inspection.
Any suggestions on framing this type of roof? Thank you all in advance!
Chad
Replies
I'm fascinated that you build one home every summer ;-) tell us more?
Can't help you on the framing details, but when I picture mansard houses I don't picture an l-shaped footprint...I'll have to think about that for a while....
Personally I wouldn't get too hung up on using a different roof style to make your project stand out from the neighborhood. So many other thing go into fitting a house to a site, and to designing something that has that certain "je ne sais quois" that says, this house is a house, not mish-mosh of styles, that a distinctive roof makes for a small part of the finished product. I've seen too many dead valleys, fake dormer windows in truss framed attics, and so on....
Here's the pic that inspired us...
http://www.historicplaces.ca/UploadedImages/14878.jpg
I have a 'regular' job so one house per summer is plenty! Like I said, so far it's been for myself or relatives. I just really enjoy building but would love for it to turn into a full time gig someday (pay would be nice!).
I'm not concerned with losing a little floor space so I really like the idea of the vertical wall set in from the perimeter. The first roof pitch would be created by tying the top of that wall with the edge of the floor deck. This also elminates the problem with hanging stuff on a 15 degree wall. I have read articles where this approach is alluded to but is never really spelled out so I'm unsure if it's correct. The 'tilted' wall and wedges just doesn't sound very sturdy for Maine snowloads. I do see how the framing for the dormers supports the tilted wall though. Sort of like a kickstand! :0)
As far as being overly concerned with styles, We just want to do something different. We like the Mansard style and it fits perfectly with our goals for the house.
Thanks to all!
Chad
View Image
that type of mansard is framed just like a gambrel.. but it has 4 sides instead of two
you have to carry the upper roof loads to the upper floor frameMike Smith Rhode Island : Design / Build / Repair / Restore
<<<
The key feature of Second-Empire style buildings, Mansard roofs are hipped roofs that are nearly flat on top and steeply sloped on the sides, generally covering the entire height of the top story to a building. The steeply sloping sides can be straight, concave or convex. It was named after the seventeenth-century French architect Francois Mansart, who first popularized the form.>>>
with gambrel roofs, one key to the correct proportion is that all 5 points of the gambrel should fall on a circular arc
i don't know what the key is for mansards, but i'd guess that a two pitch top would be the minimum ( which would neccessitate higher snow loads and specialty roofing )
the steeper pitched lower roof i would copy from some 2d empire examples in Paris
i'd also be very careful about the roofing material.. probably slate or.. i'd probbaly opt for one of Certainteed's slate look shingles like Grand Manor
the framing was not always well done.... i'd have an engineer help with the load carrying design of the upper floor that is going to support most of the entire upper roof & a lot of the lower roofMike Smith Rhode Island : Design / Build / Repair / Restore
From what I've read, the general guideline for the steeper portion is about 15 degrees. The hip portion on top can vary, but I prefer the look of around a 3/12 pitch. I will definitely be consulting with an engineer before anything is finalized. Might even see if it can be done with trusses.
it can definitely be done with trusses
View ImageMike Smith Rhode Island : Design / Build / Repair / Restore
Tried to post this question before but it failed to post.
Why the discussion about the second floor having to carry the load?
I am not seeing something about the load path apparantly.
I know I am getting old and my eyesight is going but isn't the flat portion of the roofs load carried by the sloped rafters down to the second floor walls? (This scenario assumes that the Mansard portion is being used to create a 3rd. floor wall roof structure)
Never built a Mansard roof and I looked in all my reference books here but there are none that give me any clues .
Can anyone draw a sketch that shows a section of stick framing for the Mansard that demonstrates why this isn't the case?. To me the tipped wall with the bevel shim sounds like a short cut to actually cutting birdsmouths and securing rafters to plates with the 2nd. floor ceiling joists acting as ties across the building.
Thanks
if it's trussed... then the load is carried on the FIRST floor walls
View Image
but if it's framed, the load is shared by the 2d floor walls, and the 2d floor
.... just as in a gambrel... UNLESS.. you frame it as a faux mansard.. with the kneewall bearing directly over the outside walls and the eaves extended out to the overhangMike Smith Rhode Island : Design / Build / Repair / Restore
That kind of attic is what I would call a gambrel - Not a mansard. I've done trusses for curved mansards before by building them out of short segments of 2X12 and then having someone cut them to a radius.
My next house will have no kitchen - just vending machines.
oh... well, i got it off a truss site and it was for a Mansard.. not a gambrel
the main difference betweeen a gambrel & a mansard is... the gambrel has gable ends, the mansard doesn't , ... in cross section they both look the same
as the OP pointed out... most "mansards" today cheat... they bear on the exterior wall and use an overhang to simulate the masard look
but they really don't LOOK like a traditional 2d empire mansardMike Smith Rhode Island : Design / Build / Repair / Restore
Great discussion guys! Mike, you're basically saying exactly what I've been getting at. I want to stay true to the Second Empire roof design as pictured above. I'll run far far away from the hideous 'modern mansard' (see this pic...http://www.savvyhomeadvice.com/glossary/images/mansard.jpg) where there's an overhang and the dormers are inverted.
It seems to me that the actual roof is the upper hip portion. The 15 degree tilted wall is still just that - a wall. Even if I have kneewalls, that tilted wall will still be there. So I guess the roof load would transfer to the exterior walls via the tilted wall and also to the 3rd floor joist (or ceiling joists depending on how you look at it!) via the vertical knee wall. What do you guys think?
i think i've screwed up some gambrels in the past by not paying enough attention to the load the kneewall imposes on the floor joists... most of that load goes straight down, not out on the diagonal
Mike Smith Rhode Island : Design / Build / Repair / Restore
So you're advocating the leaning wall method? I was concerned with too much load on the joists as well. As far as the overhang is concerned, it appears to me looking at most pics and local homes that the lower roofline (or tilted wall) bears directly on the exterior wall and the overhang seems like a separate piece altogether.
Driven,
I recently added a mansard addition to my mansard house. The original house was built in 1967. The mansard wall of the original was built as you described it. The mansard wall is on a 15 degree angle to the first floor walls and sits on a wedge nailed though the sub floor to the rim joist and floor joists. The rafters supporting the flat roof sits on another wedge nailed to the top plate.
The original house had two bedrooms, front and back that ran the width of the house. The reverse dormers (which, I agree look terrible) are there to support the angled mansard and to transfer load from the flat roof to the second floor joists. I eliminated the reverse dormers on the existing house converting them to shed dormers, but kept the triangular walls of the reverse dormers in place on the interior to maintain the integrity of the existing structure. As I understand it, the reason why mansards are now built with overhangs is that is the only way to have vertical interior walls without losing floor area in the interior. In our existing house the interior walls all had the 15 degree angle. No pictures, no bed headboard, the bathroom mirror mounted on the wall shows a good reflection of my knees on entry the the bathroom, worse as I get closer.
With the addition we built the second floor walls over the first then hung the mansards off that wall, giving us a consistent 15 degree slope, interior walls we could hang art on and overhangs.
I have the original builders drawings for the original 1967 house if they would help you.
I just spent an hour or so on the web looking at photos of historical Mansard roofs from around the world. Also included were a number of Second Empire styles.
Seems that the originals (assuming one would call a building built in the 1760-1890 range original) varied in whether they had eave overhangs or not. Most pictures I looked at showed an overhang , often with a detailed set of brackets as decoration or support.
The inset dormers do not appear in any of the photos I looked at , but in every case it appeared to me that the roof load was being carried either directly on the wall below , or in some cases it looked almost as if the ceiling joists for the lower floor were cantilevered out to support the rafters and create the overhang.
I would love to see some drawings/photos of how these roofs were framed so when you come up with some please post them .
if you drop a plumb line from the upper roof bearing point, in every instance i saw , this point will be substantially inside the building line
the 2d empire mansard has very little overhang.... the Italianate mansard has a more substantial overhang
i think the "cheater mansard " in a lot of modern takeoffs , a plumb line from the upper bearing point will fall over the outside building line, and the mansard is achieved by exagerating the overhang
a good source book for the variations is " A Field Guide to American Houses"... Virginia & Lee McAlester, 525 pagesMike Smith Rhode Island : Design / Build / Repair / Restore
I agree with the statement that the "true" Mansart has the upper bearing point well within an area contained by a line plumbed up from the lower floors walls. No question about that.
Again , most photos I looked up showed some eave overhang. No matter where the building or just which variation of the style of ornamentation.
What I still don't get is why a bearing knee wall that transfers load to floor joists? I guess what I am mentally picturing is a standard full hip roof structure at the wall plate line, no one ever puts knee walls under them.
Would not the load from the flatter pitch upper roof be transferred to the lower steeper rafters and thus down those rafters to the walls of the floor below. If this is the case no knee wall is needed.
I did find one reference that spoke of the upper roof using a "king post " style of construction and the lower pitch using a "queen post style" of construction , which suggests to me that you are correct in what you are saying and that I am not seeing the load path correctly and am wrong in my premise about how the load can be carried . I would just love to see some authentic photos or drawings .
some mansards were framed just as you said... no knee wall... just the rafter and the load goes all the way to the outer wall
but in a gambrel there is usually a knee wall , and so some ( or most ) of the load goes to the floor joists
i would think that most of the mansards were framed with a kneewall also... it's pretty hard to frame the lower pitch rafters without one... unless they were using falsework, which was removed laterMike Smith Rhode Island : Design / Build / Repair / Restore
Thanks,
That helps. I have framed a few Gambrels and all those I framed had either knee wall or post and beams for load paths.
I just didn't see it as necessarily needed in the Mansart.
Assuming that you want a flat roof to second as an observation deck or patio behind the mansard , then just build a flat roof over the second story; go to the truss plant with your plans and they will make a truss that looks like 12 of a truss. The bottom chord sits on top of the flat roof with as much overhang as you like. I did a lot of them this way in the Fla. keys back in the early 70's.
Nope, definitely want the hip roof on top! I did find some pictures of an addition that were somewhat helpful, but I'm still a bit confused. They show the tilted wall section being framed and supported by temporary bracing. It appears that once 4 of those wall are up that they just support each other? That can't be right! Here's a link... http://www.mansard.org/galleries/PraveenMutalikGallery/index.html
guy just posted 3 good mansard pics in"Along the Way"
here's a framing detail
View ImageMike Smith Rhode Island : Design / Build / Repair / Restore
Why can't that be right?
Thanks Mike! That helps... I've e-mailed mansard.org to see if they have the full sequence of photos from that mansard addition.Dovetail, I guess it is right, but after always building homes where load bearing walls are plumb it's hard to make myself accept that apparently they don't have to be... :o)
Yeah, really different from what I was thinking. There's probably a number of different ways you could go here depending on what you want your end results to be. I imagine that the original mansards were a post and beam type. Something like that could look nice with hand hewn beams and plaster. Lots of mortise and tenion joints knee braces and the like. That would be the high end. You also could just frame up straight walls with large openings by using some long headers. The openings would be used for built in sleeping areas or libraries or closets or desk areas, fireplace or any thing else you might see fit. Then attach additional framing over it for the shape, up to the transition point. After which you could just use trusses for the top part, unless your into framing that also,using collar ties. In any event your going to need plans and a framing schedule to get a permit.Maybe your architech will come up with some ideas of his own . Good luck on your project.
When I put the addition of my house, I added a 5' high peak roof to it and the existing mansard. When I took the plans for my addition to the town's building department, they had no problem with the design for the addition, but they questioned the original design. just what was supporting the snow load on the existing flat roof? It could not just be the four mansards leaning against each other. They questioned whether the original design should have been approved at all! I should add that the builder constructed at least five similar houses in town during the 60s.
My father-in-law architect and the Building Department sat down and worked out that fore and aft the 2nd floor interior walls were supporting the mansards by transfering load to the second floor and then to the first floor. Laterally those reverse dormers acted as gussets transfering load out to the first two floor joists on the second floor. The gussets straddled a main beam running for and aft down the centre line of the house landing on the joists about 4' on either side of the beam.
The floor joists and ceiling joists/roof rafters and load bearing walls all alternate in direction. That is the load bearing walls are at right angles to each other.