Mounting timber joists on new steel beam

[nwnw]

Hi folks,

I'm working on the design for a new extension (all in Sketchup).  Loads of queries to come and I'll probably put my drawings up here soon for scrutiny.  My biggest headscratcher just now is how to support the existing roof structure when I remove a 6.5m length of structural stone wall.

I've seen various ways to do this with joist hangers and sandwiching timber on the steel, but I'm wondering why I can just mount direct as shown below.

I've coloured for clarity...Blue = existing roof trusses and joists.  Red = new extension flat roof joists and firrings.  Pink = Steel

 

 

Existing timbers I'm sure will not be perfectly aligned so my thinking would be to leave a nominal gap and shim to the steel, or to profile the face and underside of the joists a tiny bit to fit.

Thoughts?? Pitfalls??

[nwnw]

Ok, so I came the the conclusion that I was overengineering the beam a bit and it was going to be very heavy.  I've reduced in size down to a 305 x 165 x 40 Universal Beam (this will be run past a structural engineer).

Downside is that the bearing area on the inner rib is then not enough for the joists.  So I am proposing to add bearing plates as shown below.  These can also be differing thickness and used as shims.

Thoughts anyone??

[Big Jimbo]

I dont have a clue, but would it not be easier / cheaper to just have a wider plate welded to the underside of the RSJ. to allow a larger bearing surface ?

[Big Jimbo]

Do you need the goalposts ? is there not enough room to simply bear the RSJ on the wall either side of the opening ?

[nwnw]

2 hours ago, Big Jimbo said:

I dont have a clue, but would it not be easier / cheaper to just have a wider plate welded to the underside of the RSJ. to allow a larger bearing surface ?

I did consider that but I think plates are possibly a better option...

- For the smaller plates I can do it all myself.  I can order the plates pre-coated and simply drill all the holes and bolt.

- It keeps the weight down (saves maybe 20-30kg over a full width plate).

- It can also act as a shim

 

Regarding the 'goalposts'...my assumption is that I will need that to comply with the structural calcs and building regs.  The cottage is circa 140 years old so most likely will not have compliant foundations.

 

[gamestrolley]

Over doing solution, timber glulam or flinch beam and fix truss or beam shoes to face of timber or use a deeper C channel and will will have more  surface for the rafter to sit in 

[nwnw]

Thanks for the feedback and ideas @gamestrolley.

I like the fact that glulam or flinch beams would give an easy way of hanging the joists.  However, one issue with these methods is that the ends would be 'free' to pivot and therefore I would see more deflection in the centre of the beam (and therefore possibly cracking plaster, etc).  The Universal Beam/Column design would be bolted to the uprights, meaning that the ends would be 'fixed'.  Online calculators show centre deflection roughly doubles if the ends are free.  Thoughts?

 

C Channel or 'Parallel Flange Channels' (PFC) only go up to 100mm deep as far as I can see.  This is right on the limit for bearing area - and that's assuming all the existing old joists end on the same plane (unlikely).  Strangely, they also seem to be 20% higher price than UBs for similar size/strength.

[torre]

With a timber packer bolted into the web of your steel you could face nail joist hangers onto the packer. The hangers then also prevent any rotation and easily nail at different heights. 

 

Given you'll need SE calcs for this anyway, honestly I'd leave the design to them. The bolted on shims sound like quite a lot of work and will need additional specification by the SE.

[nwnw]

On 28/10/2025 at 14:27, torre said:

With a timber packer bolted into the web of your steel you could face nail joist hangers onto the packer. The hangers then also prevent any rotation and easily nail at different heights. 

 

Given you'll need SE calcs for this anyway, honestly I'd leave the design to them. The bolted on shims sound like quite a lot of work and will need additional specification by the SE.

I think you could be right.  Meeting with structural eng this week.  Will see what he thinks.

[saveasteading]

It's fun to think of ideas, but there is a more fundamental issue. If you put all that load on one half of the bottom flange, it puts a rotation into the beam. A beam out of plumb becomes very much weaker.

So this is likely to be 'knocked back'.

[Nickfromwales]

On 15/10/2025 at 14:23, nwnw said:

Regarding the 'goalposts'...my assumption is that I will need that to comply with the structural calcs and building regs.  The cottage is circa 140 years old so most likely will not have compliant foundations.

This may be worse as you’ll likely need 1/4 cube concrete pads set in to take these point loads. 
 

Load would likely be better transferred down through the walls so it’s distributed. 

[nwnw]

20 hours ago, saveasteading said:

It's fun to think of ideas, but there is a more fundamental issue. If you put all that load on one half of the bottom flange, it puts a rotation into the beam. A beam out of plumb becomes very much weaker.

So this is likely to be 'knocked back'.

I see your point.  I think in reality it would be fine as the flat ends of the beam would be flat onto the 'goalpost' tops.  But...proving this through simple calcs is another matter.

I think I agree that the engineer will much prefer the simpler method of a timber packer bolted to the centre rib to keep all the forces square.

Thanks for the input.

[nwnw]

20 hours ago, Nickfromwales said:

This may be worse as you’ll likely need 1/4 cube concrete pads set in to take these point loads. 
 

Load would likely be better transferred down through the walls so it’s distributed. 

My assumption was that I would need the goalposts and significant reinforced concrete pads.

Existing foundations are obviously fine for the load going through just now, but concentrating that load on 2 points would surely need a bit of verification, right?  It will be at least 6.5m clear span so quite a load.

Will see what the engineer thinks.