Joining joists

[GMM]

I have been planning this for some time but circumstances have changed a bit so I need to work out a plan B. I need to add a floor to my loft, clear of the ceiling. The ultimate use will be storage but first it has to serve as a working platform for sorting out some issues up there. Plan A involved fitting wall plates (ledger boards) around the walls (there's an 18 inch wall at each eave) then taking the joists in through the roof after stripping tiles and before replacement. Unfortunately, the roofing job has been delayed a bit so I want to get a working platform over part of the space ahead of that, which will happen next summer.

There's no way I can get a 4(+)m joist into the space from indoors due to the angles on the stairs etc. It occurs to me though that I could get, say, a 3m length up there, so I'm wondering whether I can double up two lengths, bolted together (ie bolts, coach screws, serrated washers between, the works) once on the top floor. (This loft is above the first floor but accessed from the second floor, so there's space to assemble.) The result would be 3m then 1m in one direction and 1m then 3m in the other. Where I'm struggling, however, is in trying to rationalise whether this would be stronger or weaker than a single intact joist.

The interweb doesn't seem very useful as information ranges from exhortations never to join joists, ever, to one site that said a simple scarf joint would be fine (without doubling up or anything!). Certainly, there are plenty of sources recommending repairs of single rotten joists by bolting a short length alongside a butt join between old and new. Similarly, metal splice plates are available, but I'm not sure I would trust these without some support from whole joists flanking the join.

Just to be clear, this would be a temporary arrangement and full length joists to be installed later when the roof job happens. I'm just looking at making something to work from in the short term without taking a dive through the old and fragile lath and plaster ceiling. I'm also well aware that it's an expensive solution but sometimes it's worth spending a bit to move things along.

Any informed (or entertaining) views on this chaps?

[GB]

I have never done this, but in the tradition of USENET should that stop me from having an opinion? And expressing it?

It depends a lot on how much weight these joists are going to have to bear. Ordinarily, joists are much stronger than they really need to be. That is because people do not like springy floors. It makes them feel queasy and worried. Also, they do not like chunks of plaster falling off the ceiling below. If you do not mind a bit of springiness and your new joists are clear of the ceiling, you can probably get away with it.

I would definitely overlap the join by at least half a metre either side.

[Tim Watts]

On Tuesday 17 September 2013 20:13 wrote in uk.d-i-y:

IANASE (not a structural engineer) but 3m joists with a 2m bolted overlap whould do as something to sit on as you say it is temporary.

Use glue before bolting through - should be approaching the strength of an unbroken joist I would have thought - at least long enough until you get the roof done.

[GMM]

Absolutely...and with conviction too!

Well this is around a 2m overlap, so plenty of opportunity for bolts. Certainly a little bounce wouldn't be much of an issue when working on stuff.

[GMM]

Good point Tim. I hadn't thought of glue, partly as I was thinking about how easily they could be dismantled to replace with complete lengths, although if they work well, they might stay (flanked by otehr full length timbers). Presumably, the glue should take some flex - any suggestions? I seem to recall a link from here recently that said PVA was as good as it got.

I was thinking about keeping bolts toward the 'ends' (so in the 1/4 -

1/3 span region where boring is allowed) and using (say) staggered 3 inch coach screws in the middle, as this 'feels' a stronger solution.

[GB]

Ah, I hadn't quite cottoned on that you were going to be joining two 3m lengths to make each 4m. I'd focused on "bolting a short length alongside a butt join between old and new". Should be fine, but there might be a little bit of a tendency to twist, as the two ends will be staggered. You'll be screwing a decent floor firmly down onto these joists, though, won't you?

If you do it properly, it'll cost a bit. You couldn't just take a bit of roof covering off and get some 4m joists in between rain showers? You could do some sums - cost of all the wasted materials plus the time spent building the temporary version, versus cost of putting up some scaffolding, removing some of the roof (so you could poke the 4m lengths in) and then making good again.

[Tim Watts]

On Tuesday 17 September 2013 20:46 wrote in uk.d-i-y:

The other solution is to not bolt 2 full section lengths but bolt perhaps 6 of thinner material, eg 6" high by 1" wide nominal and stagger the joints to form a single engineered joist.

If you plastered the planks with PVA and clamped and bolted it would be nearly as good as a single 6x3" - remove whole when the roof is being done or saw in half. Use for something else less critical.

You may use no more wood than your theory but it would be stronger.

[GMM]

I think the only difficulty with that would be the dimension: I was thinking of 2 off 9 x 2 joists, which would then fit in a 100mm hanger. I'll have to take a look at whether 150mm hangers are available anywhere.

[Tim Watts]

On Tuesday 17 September 2013 21:39 wrote in uk.d-i-y:

Where did 150mm come from?

[Tim Watts]

On Tuesday 17 September 2013 22:21 Tim Watts wrote in uk.d-i-y:

Ddi you think I meant 3 off 2" wide planks? No, I was suggesting much thinner material like 3 off 1" wide leading to a finished size of 6x3"

Having 3 or more layers sandwiched into a uniform cuboid is a lot better than an overlap of 2 layers.

[GMM]

Ahh...with you now! I was thinking you meant six parallel 6 x 1 planks

- get it now (!). It's been a long day...

It would have the advantage of being easier to get up the stairs too.

I'm guessing that it would be important to source the right timber as this sort of stock won't have been stress graded but I certainly have the gut feeling that 3 layers is better than 2.

[John Rumm]

With such a substantial overlap, I would expect the whole joist to perform better than the equivalent length single section joist. The strength in sheer at the ends will be the same as the single joist, but the performance in bending at the centre will be better with the double thickness.

"Never join" will be a non starter for many applications since there is a limit as to how long a single length can be. Its quite common to join things like ridge beams etc.

For a temporary solution, a longish stopped scarf, bolted through the vertical a few times would also be adequate (although less reusable once dismantled)

Well, "educated guess" is all I would claim for the above. I have done a reasonable amount of joist fiddling and floor building [1].

Note also GB's comment about "over design" is very relevant. A normal floor joist is specced to dip no more than 14mm or 0.03% of its length (whichever is smaller) when subject to loads of 1kN/m^2. This makes for a nice rigid floor that won't disrupt the decorated surface of the ceiling under it, rather than for a design that is at the edge of its failure limit. So in this circumstance where you can tolerate far more bounce, and are nowhere near a ceiling etc you can use significantly less rigid members and still stay perfectly safe.

[1] e.g. loft conversion:

[Fredxx]

What is wrong in using the current ceiling joists?

Are they that puny? If so, can the ceiling be reinforced with acrow-props from underneath?

[Tim Watts]

On Tuesday 17 September 2013 22:32 wrote in uk.d-i-y:

Something like:

======================-----------

--------========================= ======================-----------

Arrange for the same 2m overlap, and plaster it with PVA (as in brush it on

100% coverage) - you are using new PAR wood so PVA will be stronger (and cheaper) than most other glues.

I would *think* a reasonable follow up to applying the glue would be screws (self drilling Reisser/ScrewTite type) where the screw is about 65mm long (ie almost but not quite teh full with of the assembled joist).

Screw every 50cm in 3 rows in a staggered fashion and repeat the other side staggered with respect to the first side.

That should clamp it against the glue and is not that many screws with a power driver.

Alternatively bolt though the centre line at regular intervals, but I suspect some extra G-clamps and spreader blocks would be wise whilst the glue sets.

I'd feel happy with that - but hang around in case anyone else thinks it's a crap idea...

[GMM]

Well, I was thinking I might fill in the gaps on the ends with the other

1m or so, so they should be pretty evenly mounted in the joist hangers. As you say, flooring should stabilise any tendency to twist.

I sort of agree but it's not an easy scaffolding job as it's over the conservatory and likely to cost the thick end of a grand. By getting materials in from the inside, I can take my time and do a bit when I ahve a free day or so.

[GMM]

If that is indeed the case, then they might stay there in the long term, and be supplemented by intact joists installed when the roof finally comes off. The less there is to do at that stage the better as Sod's first law of the weather is bound to operate, regardless of the time of year.

I'm sure that's right in principle but it just 'feels' a little bit dodgy to me and a little over-engineering never hurts in these situations...

I don't have any concerns about a bit of wobble in the platform as such but my main concern is that it doesn't (in either its temporary of permanent form - and the permanent version will be at the same level) sag sufficiently to come into contact with the ceiling as that could lead to another job that isn't needed at the moment. If I have to allow for too much spacing below these joists, then it will increase what will already be a step up into the space and restrict access etc etc. Given all that, it seems the simplistic approach, of designing broadly in line with specs for a habitable floor is the most reliable way to go, even if it is a bit over-done. Of course, it's highly likely that the ceiling already sags a fair bit, as shown in your photos, so that issue may be less serious than it seems.

[meow2222]

2m overlap seems excessive to me

PVA is not a good structural glue. It fails if wet and has no gap filling strength

NT

[GMM]

Yes they are. Very puny. And I'm not sure I'd be very popular if I have jacks all over the bedroom for the whole winter - having them there while I get the ledger boards up is already a matter of some negotiation.

[GMM]

My instinct is a combination of bolts through at around 1/3 span and then plenty of coach screws might do it. I've a feeling the way to go with this design is to try one out and see how rigid it feels.

One thought I did have was of fabricating some boxed I-joists from plywood and C-16 stud timber. Having found an interesting thesis covering this on the web, it's clear that these have some potential and seem to perform well in stress and deflection tests. I have a horrible feeling, though, that they really need a proper press for reliable assembly, rather than being strapped and clamped on the top floor bedroom floor.

[John Rumm]

Yup, with an inch spacing at the edge you are unlikely to get anything like enough sag to use all of that. Over a 4m span, that would need to sag twice as much as the maximum permitted to even come close, and that assumes that the existing ceiling is dead flat. (which I thought mine was until I could eye it up against something straight!)