Help needed with I beam construction roof

[joe90]

I had planned to build my roof from 300mm Finn I joists, hip roof, no purlins, two mtr ridge, osb lined inside for racking and vapour barrier, full fill with 300 mm batt insulation, felt, battens, slates etc. I have been told this is incalculable for the roof by the I beam suppliers! They are talking ( to my builder who I trust) about a raised tie truss roof battened out to 300 for insulation and using a glulam ridge beam or even needing steels in some way.

 

With regard the glulam ridge beam ( which I cannot see the need for) I have had this discussion before, with a pitched roof the ridge beam simply ties the two sets of rafters together and carries no weight as such. If ( like a loft conversion) one pitch was removed then yes a beam, steel or glulam is needed to carry the weight of the slope retained but with equal and opposite slopes surely this is not required.

 

Surely 300mm I joists over a 5 mtr span at an angle of 35 degrees is not rocket science to calculate ( I wish I had become a SE or at least learnt the theory before I started this project). We are also getting into the realms of paying for an S.E. to make these calculations which obviously I would like to avoid. 

 

So, am I barking up the wrong tree?, am I trying to be too simplistic?, I cannot for the life of me see the requirement for steels.

 

HELP

[Tennentslager]

31 minutes ago, joe90 said:

 

With regard the glulam ridge beam ( which I cannot see the need for) I have had this discussion before, with a pitched roof the ridge beam simply ties the two sets of rafters together and carries no weight as such. 

I'm no expert but my reading says a ridge BEAM carries half of the weight of the roof...the remaining half shared by the wall plates.

[joe90]

10 minutes ago, Tennentslager said:

I'm no expert but my reading says a ridge BEAM carries half of the weight of the roof...the remaining half shared by the wall plates.

But where does it carry it too?, each opposing rafter at the ridge leans against one another, the ridge board between them ( which in most houses is only a 1" board) simply ties them together . The wall plate also ties the rafters at their set distance and  transfers the weight directly to the wall below.

 

A steel beam or glulam is designed to carry weight over a span, loading  that weight onto whatever the beam is resting on at each end.

 

if in a ridge roof a very large velux or similar was installed then yes a beam is required to carry the weight over the gap in rafters.

[ProDave]

16 minutes ago, Tennentslager said:

I'm no expert but my reading says a ridge BEAM carries half of the weight of the roof...the remaining half shared by the wall plates.

In a gable ended roof, with the ridge supported on the gables, the ridge beam takes half the weight of the roof, and importantly stops the roof spreading allowing room in roof situations where upstairs you might have half height walls.

 

Joe is talking about a hipped roof. So there is no gable to take the weight, that is taken by the sloping beams at the hips. This type of roof needs ties at floor / ceiling level to stop the roof spreading, and usually intermediate support on purlins.

 

I can't see the loading on the gable beam being very high and am surpised someone thinks it needs a glulam beam.
 

[Tennentslager]

Again no expert but think that's a ridge board. A roof like that relies on the triangulation of the truss to transfer the weight to the wall.

In a hipped roof the *end* trusses are beefed up...but hey Joe don't listen to me, I've only read some log cabin websites and built one structure!

A fullsome explanation from the I-beam people needed I suspect.?

* just read some more and all of the rubbish I wrote above is bollocks, hope some more knowledgeable folks come along soon*

Edited January 9, 2017 by Tennentslager

[MikeSharp01]

Hi Joe. We had the same problem as we originally specified a I-joist portal frame - essentially they cannot calculate the shear forces in the joiner plate against the web - or rather they can but the manufacturing method would be critical and the joists are not built to be reliably calculated in this mode. Both the joist companies looked at and the SE and came to the same conclusion. I have done some tests and I tend to agree that its hard to get a reliable strength. Someone sent me off to find a steel portal structure that did not extend the beam depth in the centre down to take account of the 'flattening forces' and I couldn't find one, so have to agree with today's technology it cannot be done in I-joists.

 

Putting in a ridge beam means that the load is not transferred across the roof but down onto the the beam in the middle and down at the walls. In a traditional trussed rafter roof the whole thing is trying to flatten out and its stopping that where the ridge joint and associated calculation comes in. Add a ridge beam, itself supported by uprights somewhere, supporting the ridge and that problem goes away. (Edit: Although there is still a small component of force pushing the wall out)

Edited January 9, 2017 by MikeSharp01

[joe90]

Yes guys I understand the last two points. However at ceiling level there are ties in the firm of ceiling joists or more I beams so this would be taken care of, with regard the ridge I still don't understand as like with a ridge board they simply push against one another with equal and opposite forces.

 

looks like I need a structural engineer ( bugger) to work out the detail.

[joe90]

Mike, read your post a few times to get my head around the problem, are you saying that the problem is in the " joining" of the I beams, I.e. A satisfactory way of joining them to carry loads across the whole I joist?.  

[PeterW]

45 minutes ago, joe90 said:

Mike, read your post a few times to get my head around the problem, are you saying that the problem is in the " joining" of the I beams, I.e. A satisfactory way of joining them to carry loads across the whole I joist?.  

 

The issue I was told was getting a satisfactory join across the tie to rafter connection at the wallplate as you cannot put the webs into tension across their width so you need a connection method that ensure the top and bottom flanges are secured to the the rafters and the ceiling joists. 

 

A ridge steel is a comparatively quick calculation if it's a standard design and there are structural internal walls to place the steel onto. Other option is to use an IC at either end to stand the ridge beam on. 

 

[PeterW]

45 minutes ago, joe90 said:

Mike, read your post a few times to get my head around the problem, are you saying that the problem is in the " joining" of the I beams, I.e. A satisfactory way of joining them to carry loads across the whole I joist?.  

 

The issue I was told was getting a satisfactory join across the tie to rafter connection at the wallplate as you cannot put the webs into tension across their width so you need a connection method that ensure the top and bottom flanges are secured to the the rafters and the ceiling joists. 

 

A ridge steel is a comparatively quick calculation if it's a standard design and there are structural internal walls to place the steel onto. Other option is to use an IC at either end to stand the ridge beam on. 

 

[MikeSharp01]

Sorry Joe, just got to this in a break from bending Rebar! Yes its the joint at the top that is the problem. I was going to CNC cut several plates to go either side of the web to create the joint but its impossible to work out where the failure will be, although destructive testing might get to it, because the web is OSB the plates are Ply and the nails are nails, as the load is applied eventually the plate will start to turn, in trying to keep things in place, and will start to load up the flanges at the foot & head of the joist and the joint between the web and the flanges is not designed, or even contemplated, to be loaded up in this way, steel joists are all one material and so don't have a problem. If you glue the plates then the failure moves to the web itself just below the end of the plate because, essentially, of the turning action and the point loads it induces in the web and the associated flanges. In the end after looking at this for a long time, I was at first reluctant to believe the joist companies, and running my thinking past academic structural engineers of my aquaintance I gave up and accepted the ridge beam - ho hum. It adds about £500 to the roof cost for us.

Edited January 10, 2017 by MikeSharp01

[joe90]

Mike, thanks for this, a very comprehensive answer. I think I would rather be out bending re bar than swotting this stuff but you might disagree ?. I can live with the ridge beam as I could use a large timber/glulam supported on two 4" posts from structural walls below. Like you I have considered cutting ply gussets glued and screwed across both sets of Osb,s to tie the structure together at the top, another American website simply shows two joists butted at the ridge and metal strapping fixed over the ridge to stop them sliding apart.

 

on this same American web site it shows sloping joists glued and screwed to ceiling (spreader) joists through top and bottom rail only. I am sure a full fill of the gap and more glue and screwing would be more than adequate.

 

I think this is now academic as unless I get a S.E. to sign off what I want, it won't get passed so I need to find one with a sympathetic ear and a knowledge of I joists.

[Simplysimon]

a word with touchwood homes might be an idea as they build with jji's, i've been looking at them as an alternative to twin stud.

simon

[ProDave]

Two points of relevance.

 

One of the new builds near me has ther timber frame built of JJI I beams, including the 45 degree pitched roof, but crucially the roof is supported on a ridge beam.

 

the portable eco home builder close to me used I beam construction as well, and they do much like what is being talked about. When I asked them they would not give details about exactly how they made the joins but from what I saw, it looked like the gusset plates were made of OSB, very thick perhaps several layers, to the point that the "stack" of OSB at the joints was as thick as the fat bit of the I beam.
 

[joe90]

What I don't understand is why a ridge beam is needed with I beams but not with roofing joists , surely I beams are just engineered joists. Well the site that gave me most knowledge is this one. http://www.donnybrooktimberframes.co.uk/files/I-Joists - Boise UK Technical Guide.pdf and yes a ridge beam is asked for ( I think I have found a way of providing it in a hip roof so hopefully my problem is sorted,) if it is I will post my solution in case it helps others. Yes joints are all about packing out the web,s to match the thickness of the top and bottom string so the beam as a whole is loaded ( as far as I can tell from the above).

[Temp]

1 hour ago, joe90 said:

What I don't understand is why a ridge beam is needed with I beams but not with roofing joists , surely I beams are just engineered joists.

 

Perhaps engineered wood I beams can't handle tension? So the triangulation forces that tend to push the walls out have to be handled by a structural ridge beam instead of tension in the joists. 

Edited January 12, 2017 by Temp

[Temp]

Actually that can't be it because they will be in tension when used as rafters with a structural ridge beam.

[joe90]

Now you understand my problem ?

[joe90]

Sorry guys but this is bugging me so I needed to do a drawing, now from my very simplistic drawing I fail to understand why "W" needs support from below

1/.  flex is the bend within the rafter and it depends on the weight and depth of the timber/ijoist

2/. Spread is the tension in the ceiling joists created by the weight of the whole roof ( I admit this could be less if a ridge beam were installed)

3/. W1 is the equal weight on both wall plates of the whole roof ( unless a ridge beam is installed)

4/. W2 is the equal and opposite weight both sides of the roof create by resting on one another ( think of house of cards) 

 

The various pressures are dependant on angle a.

 

therefore if all the above is adequate there is no need for W to be supported I.e. No ridge beam.

 

 

 

 

[DavidFrancis]

This probably does nothing to answer your main question, but in a strong wind I'd have thought your W forces could be considerably different one side to the other.  Also possibly with wind-blown snow.

 

 

[joe90]

17 minutes ago, DavidFrancis said:

This probably does nothing to answer your main question, but in a strong wind I'd have thought your W forces could be considerably different one side to the other.  Also possibly with wind-blown snow.

 

 

I would have to agree but there is still no difference between  a lumber made roof and and I beam roof.

[ProDave]

What your drawing misses, is Purlins to take part of the flex load, and transmit it partly to the gable walls, and partly to any supporting props they may have.

 

But weren't you talking of a hipped roof?
 

[joe90]

The purlins are missed on purpose, I don't want them, need them.! The main reason for a clear roof space with a warm roof is to create a VCL with osb cladding and not have lots of "posts" etc poking through. ( yes I know I have to deal with ceiling joists) I want to stay away from tapes and rely on glued joins for air tightness. The span ( 5 mtrs) of the rafters is more than coped with using 302mm I joists, the walls and stretcher ceiling joists can also easily cope with the loads, it's just this ridge thing that has got under my skin as I don't see it's need ( enter an S.E. who explains why I am completely wrong ?)

[Simplysimon]

On 09/01/2017 at 20:49, ProDave said:

In a gable ended roof, with the ridge supported on the gables, the ridge beam takes half the weight of the roof, and importantly stops the roof spreading allowing room in roof situations where upstairs you might have half height walls.

 

sorry @ProDave, it doesn't, the ridge carries no weight. it is there purely to make the fixing of the rafter easier. a trad roof can be built without a ridge and in no way stops the roof spreading, a collar tie would stop the spread.

 

Joe is talking about a hipped roof. So there is no gable to take the weight, that is taken by the sloping beams at the hips. This type of roof needs ties at floor / ceiling level to stop the roof spreading, and usually intermediate support on purlins.

again, no. the hips will transmit a small amount of load but not much, it would need to be a point load at the junction of the crown, hip and first pair of commons. again a ridge is not required, however, the main part of the roof would need to be sheeted to avoid the roof falling like  pack of cards when the crown and hip rafters were placed. the best trad roof is a pole plate which cannot spread under any circumstance.

On 09/01/2017 at 20:49, ProDave said:

 

I can't see the loading on the gable beam being very high and am surpised someone thinks it needs a glulam beam.

i agree!
 

simon

[Simplysimon]

@joe90 , if the i beams are both cut with the plumb cut and well cut to ensure full contact, there should be no requirement for a beam. there would probably be a need for a ply/osb solid brace cut and fitted into the web on both sides probably 500/600mm long to tie them both together. again once the roof cladding is fixed there should be no movement but isuppose belt and braces.

 

simon