PDA

View Full Version : Fixing a Deck to a Curved Wall?



Clive McF
1st December 2007, 09:41 AM
Hi All,

I'm building a pool deck which will have a curved brick wall running along one side. Normally I'd just bolt a ledger to the wall, but the curve presents a few problems. The curve is convex, ie the wall curves in towards the deck, with the centre being further in to the deck than the ends. Anyone have ideas as to how best to do this? I wondered about putting partial saw cuts through the thickness of the ledger so I could bend it to the curve, but this might weaken it too much? Or I could use several short lengths, but even then they wouldn't have much area of contact with the wall. Or I guess I could fasten the joists to the wall with individual hangers but that's a lot of dynabolting and drilling. All advice would be very welcome.

Cheers

Clive

Gaza
1st December 2007, 09:47 AM
what about gal angle to support the joists, better than hangers as has more meat and would require less fixings than say a joists hanger.

Clive McF
1st December 2007, 09:53 AM
I wondered about that, but if using one piece of angle per joist I'd still need 2 bolts per joist, no? I guess I could get one long length of angle cut and welded up to match the curve of the wall, but this would be very expensive I'm thinking?

echnidna
1st December 2007, 10:10 AM
Omit the ledger altogether and put stumps in.
Bearers go from stump to stump so that the joists will overhang the bearers and go right to the wall.
I would keep the joist overhang to 300mm or less.

Clive McF
1st December 2007, 10:38 AM
Problem with that is that it's a new wall so the backfill isn't compacted and wouldn't take the loading of the stumps: I'd need to dig down about a metre till I hit the footings of the wall to take the load.

echnidna
1st December 2007, 10:58 AM
a metre's not that deep

Clive McF
1st December 2007, 12:00 PM
It's a lot of concrete...

echnidna
1st December 2007, 12:12 PM
use a treated pine soleplate and stump, no concrete needed

brynk
1st December 2007, 12:21 PM
you could direct-bury your posts then you wont even need concrete. a h4 treated pine post with the ends sealed with paint/wax/etcet and any cuts re-treated will go 40 years in ground. or you could consider any one of the aussie hardwoods rated at h1 durability - untreated they will do 25-30 years depending. have a read of the post on our deck (http://www.woodworkforums.com/showthread.php?t=58109) for directly burying ironbark posts, and a method of squaring them up that my brother suggested - placing a rubber mallet against the post and striking the face with a gympie (or lump) hammer while someone else pulls the stump towards square. get a rythm going and the vibrations displace the backfill around the stump slightly and it moves towards plumb.

here is a study done by the fwprdc - http://www.fwprdc.org.au/content/pdfs/PN04.1004.pdf - the in-ground natural durability of aussie timbers

i can also suggest a stumping shovel - they're like a giant pair of pliers with shovels on the end - you can fill the handles with sand to put weight in them & you will chew through down to a metre in no time. a decent set will set you back 80 bucks and a shyte one maybe 1/2 that.

r's brynk

journeyman Mick
1st December 2007, 03:10 PM
What about getting a length of sreel angle rolled to the radius of the wall and bolting that on as a ledger. This is often used on commercial jobs.

Mick

Clive McF
1st December 2007, 03:27 PM
I must admit, of the suggestions made so far (thanks, everyone), using the wall seems the least hassle. But as I mentioned earlier, getting angle made up would be expensive I imagine? Any thoughts as to what size angle would be needed?

journeyman Mick
1st December 2007, 03:33 PM
Probably 75 x 75 x 6. Places that roll curves would do it in a few minutes. I think if you offset the labour costs involved in the other methods then it would probably work out pretty cheap to buy the steel and get it rolled to your radius.

Mick

Pusser
1st December 2007, 10:15 PM
You could laminate a ledger. Use boards thin enough to follow the curvature and use enough concrete screws to hold it to shape. Glue nail subsequent thicknesses and use nails as clamps. Put dynabols through the whole laminated sections when dry. If you did it this way even a few kerfs would not matter as long as they did not line up with each other. Just a different option to consider.

Clive McF
2nd December 2007, 12:44 PM
That's an interesting thought: you'd use the concrete screws to hold the boards to the wall while building up the laminate? What size and type of wood would you use (I'd guess the radius of curvature is around 5m)?

Pusser
3rd December 2007, 12:15 AM
Please note my idea was just a sugestion as to how you might approach it seeing you were reluctant to get a fabricated steel piece and the stumps idea was not attractive. It was a left field solution I adapted from historical WW2 boatbuilding (no I wasnt there!). I thought someone might pick it up and run with it.

Having said that I would approach it using the species and sizing as if it were a straight ledger except I would move to the next size in depth as you may have to kerf the timber. The width would be the same finished size but a laminate rather than a solid piece. Assuming you wanted a finished width of 50 mm then ideally you would use 10 mm laminates assuming the timber you want will take a 5 metre radius. At 10 mm I think most would but not sure. Otherwise you could use 19mm and kerf it. Providing you do not line the kerfs up it should be no problem once glued and bolted. I would paint preservative on the back of the board going next to the brickwork, screw it into the brickwork at the centre and move outward arttaching it with those blue brick screws. You would not need to secure it to take the weight of the deck but still hold it firmly to resist the pull of the other laminates until you through bolt it. Use waterproof glue, (slow setting epoxy) and attach the next laminate working out from the centre again - a nailer would be useful here. I would glue in sections a metre or so at a time) The nails clamp the laminate so you will need a few and it would be best if they were sKewed. Repeat for subsequent laminates. You would not need to wait for the layers to dry. When the laminate is finished drill through the laminate. and into the brick work with some ramset bolts and do them up tight. You will probably have to clean up the top unless you use a glue that sets slowly enough so you can clean up each laminate.

As I say this is only an idea and I have not tried it but it is the way they used to build wooden minesweepers out of douglas fir and mahogany except they used copper nails to secure the laminates. the curvature on ships hulls was considerably more complex than the structure I think you are doing. The laminate should be as strong if not stronger than a solid piece even with kerfs. If you have to kerf make sure the kerfs are evenly spaced and start at a different distance from the end so none line up.

I would do it this way but I think you should get a second opinion before you try it on my say so.

Pusser
3rd December 2007, 07:52 AM
having slept on it it would be easier to use thinner laminates so you ar not fighting the laminate when bending. I would stick to thicnesses less than ten mm possibly 6mm - the less resistance when you bend would make a better laminate and the extra time would be compensated for by the easier working and less resistance to bending.

journeyman Mick
3rd December 2007, 09:35 AM
Unless you fill your laminations up with waterproof glue of some sort they'll end up holding water and rotting. Trust me, a rolled steel ledger will be cheap(ish) and quick.

Mick

Pusser
3rd December 2007, 05:33 PM
Unless you fill your laminations up with waterproof glue of some sort they'll end up holding water and rotting. Trust me, a rolled steel ledger will be cheap(ish) and quick.

Mick
I agree with you about the rolled steel being cheap(ish) and quick.

I thought most epoxys were water proof and as it works with ship hulls in a marine environment I would have thought it was ok. Rot is as much problem as rust. Certainly that way in ships - it all depends on the protective finish.

journeyman Mick
3rd December 2007, 07:39 PM
.........I thought most epoxys were water proof and as it works with ship hulls in a marine environment I would have thought it was ok. Rot is as much problem as rust. Certainly that way in ships - it all depends on the protective finish.

The epoxy is certainly waterproof, but even if you're paying yourself a pittance per hour to make up this laminated timber ledger the steel one will work out way cheaper.

Mick

Pusser
3rd December 2007, 11:29 PM
I still agree with you Journeyman.

thebuildingsurv
4th December 2007, 11:19 AM
Why not just bolt some vertical bits of 70x45 treated pine to the wall and sit your joist on those.

Clive McF
6th December 2007, 01:34 PM
Thanks for all the useful advice everyone: guess the ball's in my court now! So realistically it'll come down to time (making a laminated ledger, fastening joists individually) vs cost (getting a curved angle made - and hoping it fits my measurements!).

echnidna
6th December 2007, 01:41 PM
Why not just bolt some vertical bits of 70x45 treated pine to the wall and sit your joist on those.

Nice simple strong method :2tsup:

addo
6th December 2007, 06:41 PM
...So realistically it'll come down to time (making a laminated ledger, fastening joists individually) vs cost (getting a curved angle made - and hoping it fits my measurements!).
If you want to get it done before the man in the red suit descends your chimney, better call a few metal bashers tomorrow.

This time of year is hectic and the final stage (galvanising) you only have about a week left to get the piece in for dipping. If you get it made in car or ute length sections, you can do the legwork yourself between fabricators and galvanisers - this could be the difference between having it ready to bolt up on 21st December, vs being ready to bolt up in early February...

Regards, Adam.

Clive McF
7th December 2007, 02:17 PM
Sound advice: I never plan to get anything finished in time for Christmas, unless I start in Feb. And sometimes not even then...

Thinking about the vertical TP idea, presumably I'd need to make sure the upward-facing end grain was well sealed/covered since it'll get the worst of the weather?

thebuildingsurv
7th December 2007, 02:36 PM
u could do that and also put a bit of malthoid on top.

wads
6th April 2008, 10:58 AM
I'm about to embark on a similar project. Clive, did you get to complete your decking? Can you pass on any tips?

My garden perimeter is 5x5m. 2 sides of the square are set out with a raised flower beds with curved wavey walls both convex and concave. one side is flush against the house and one side is flush against a wall. I want the deck at 45 degress from the house and want to deck right up to the curvey walls of the flower beds.

Question: Do I fix ledgers to all sides of the decking perimeter including a curved ledger as discussed here?

Many thanks

autogenous
6th April 2008, 11:16 AM
Bolt a bearer to the wall on the parrallel wall. Run joists on top of the bearers perpendicular to one wall and bolt the other joist parallel to the other side of wall. Run one joist 45 to the curve. Then infill with trimmer joists in the curve to the wall.

Plan B. Build sleeper wall below deck to joist height. Run joists to top of sleeper wall as per above.

wads
6th April 2008, 06:06 PM
Thanks,

Just to confirm, please can you refer to this flickr link and repost your reply referencing A, B, C or D in the diagram. http://www.flickr.com/photos/25378405@N07/2391118337/.

Many thanks