Are there walls on both sides?

I have another idea, but this could get tricky with the analysis and detailing. Can you design a stringer (out of heavy tube steel) that looks like this "<". This stringer would carry all the loads...torsion, bending and axial forces (like a truss). Each stair tread would cantilever out of this stringer.

I would think about a small main structural member in the treads, that is welded to a small embeded plate. The plate and main member would be hidden within the tread build out. Of course, embeded PL and holes in walls, even with the large contractors, tend to have issues with placement. Building in some reasonable tolerance is advised so any mistakes in field are not too problematic.

The tapered flange channels are the structural members. I could weld a endplate to these sections, which can be post fixed with chemical anchors into the wall. The flange channel is only 76mm high, to allow for the hardwood cladding. Any connection will need a lever arm action to form a rigid connection into the concrete. Minus edge distances, say I use 12mm anchors, gives me a 46mm leverarm. If I design each tread for say a 130kg load at the tread edge, that is a moment of 1.69kNm. With a 46mm leverarm gives a tensile force of 36kN which I can support with a HILTY HIT anchor.

So just need to design the endplate thickness.

The simplest connection would be a 400 wide x 200 high embed plate for each tread. The plates would be anchored to the wall and tread members would be field welded to the plates. A slight misalignment of the plates would not matter as they are oversize.

BA

From past experiences with these monstrosities be very careful about the deflection of the stairs as owners can complain that there is too much bounce when they walk (run...) up and down the stairs.

But I agree with BA, a precast end plate and site weld the treads to the plates is the way to go. I don't know the design loads but think you would almost struggle with chemical anchors.

Agree with both BA and jwilki, the key word being 'monstrosities'. Why is it that architects see themselves as sculptors?

I’d set the two conc. wall forms on the stair well side, and lay the stair out right on the inside face of the forms, in terms of placing any embedded hardware. Then install the rebar and the outside forms. You could use BA’s embedded weld plates, or you could embed steel boxes in the wall and stuff the treads into them and then weld them in place. You could also post tension the treads from their free ends into the conc. wall. Or you could embed a full length sloped stringer plate in the wall and then trim it out when you trim the treads, as you might an old fashioned stair way.

I don’t know quite how you are going to fabricate the tread boxes and then clad them with solid wood. Will expansion and contraction be a problem with wood joints, etc? You might consider tipping the free ends of the treads up a few degree to compensate for cantilever deflection. You might also taper the treads from 5" at the wall to 3" at the free end. Also consider, that at each free corner you provide a vert. connection (rod or 1" sq. tube, or some such) to the tread corner above and below so each loaded tread shares its loading. You may not want to interconnect the treads and glass handrail system because of the differential deflections. And, don’t forget the fairly sizable handrail loadings (lateral and vert.) required by codes, the Arch’s. design looks a little light. The landing looks much too light also. I’d review these stairs as they relate to your building code, sometimes an Arch’s. pipe dreams and the building code are not exactly in synch.

The other issue is how do you make the tread function as one unit? The architectural detail with the tapered flange channels at either end will not work. If I'm walking up this stair and I place my feet on the edge, then essentially all the force goes to one channel section (the hardwood is not stiff enough to transfer any load to the other channel), the force can be quiet large especially if someone is carrying a heavy object up/down the stairs and stops on the edge of a tread. I think you should investigate using square/rectangular hollow section prongs and connecting them together with flat plates top and bottom (that will help stiffen them up and transfer load across).

I like the idea dhengr had with the continuous sloped embed plate at each end of the treads. This can be post installed with bolts positioned to be hidden within the volumne of the built-up tread. The treads would be field welded. In my opinion this approach would be the cleanest and easiest for the contractor.

Thanks, Loads of useful suggestions!