[frdf250]

I am wrapping up a 20'x48' deck and I am having some issues controlling the sway. I will list a little about the deck below -

• Ledger Board - Deck is attached to the house with a ledger board that is fastened to the house using 1/2" galvanized hex bolts and nuts and washers on the inside of the house.
  
• Posts - Deck sits on (8) 8"x8" pressure treated posts and (2) 6"x6" posts. All of the galvanized brackets were attached to the concrete with J bolts and torqued to the specs recommended by Simpson in the product manual
  
• Post to Beam - Each of the posts are connected to the steel 4"x12" I-Beams using a piece of 1/4" plate bent on a 90 degree angle to form a clamp on the bottom flange of the I-Beam.(see picture) The plates are attached to the post with (6) 1/2" lag bolts and each post has two clamps. Plates were then welded to the I-Beam.
  
• Beams - The I-Beams are full continuous pieces straight from the mill at 50' long - I cut off 2' from each beam to make them the correct length.
  
• Joists - The joists of the deck are connected to the I-Beams through a 2"x6" board fastened to the top of the I-Beam and fastened every 16" with (2) self drilling TEK screws at each 16" location. Joists are connected to the house using Simpson brackets. Anywhere PT wood is touching the steel I dropped down a layer of adhesive rubber roofing flashing as a barrier.
  
• Deck Boards - Deck boards run perpendicular to the joists.
  
  
• Existing sway control -
  
  [*]5/8" galvanized cable connected to the beam and post running in an X pattern as close to a 45 degree angle as I could. I purchased a tension meter and have about 1100lbs on each cable.
  [*]3"x4" I-Beams were installed to brace the middle of each post preventing any bowing from the tension applied to the cable.
  ​

​

I installed 2"x4" underneath the deck in a "V" pattern and connected them to the joists with HEADLok screws.


I do not have any knee bracing or sway control on the beam closest to the house. I don't think it would do anything since the beam is only 4' from the house. If you stand over the beam closest to the house it has virtually no movement. once you go out about 5-8' you can feel the movement. Span between each beam is 14'.

Thank you in advance for your thoughts and suggestions.

 

[Bud9051]

Good morning and welcome to the forum.

I'm not an engineer, but at first glance the "X" cable bracing needs to be anchored at the bottom. Like the ledger board is not moving because it is anchored to the house, that cross bracing needs to be anchored to the ground. Hopefully the ground isn't moving . You could add some 2x4's from top to bottom temporarily to judge if it improved the sway.

Bud

 

[SPS-1]

Looks like a beefy design.

I think I would have done solid X members, but still that should be pretty stiff. You don't mention how the cables are terminated to the structure, but likely done as robustly as everything thing else.

You sure its actually swaying ? Sometimes we are so worried about something, that you could swear you see/feel/hear it.

 

[frdf250]

Bud - My concern with doing the full "X" would be applying all of that tension to the cables and relying on the anchor bolts that are holding the post bases down to keep the posts from moving. Valid point though.

SPS-1 - attached a picture of the connection point at the top of the "X" and at the bottom near the middle of the post its connected using a 1/2" x 4" structural eye bolt installed on a 45 degree angle sunk all the way to the head.

Surprisingly the "V" bracing seemed to provide the biggest improvement. I am wondering what would happen if I tighten up my "V's" to get one more in there or making them "X's"?

 

[SeniorSitizen]

The X bracing in that location is for accessibility and egress and with the post mounted in that manner it will always sway. Lowering the X's i'm sure will help but 6x6" or 8x8" posts in concrete with elevated X's is about the only way to make it acceptable.

 

[Bud9051]

Looks like those cables may be touching or close to it at the center of the "x". Get a step ladder or other so you can get up to one of the center locations and have one or some people up top wiggle to aggravate the sway. Those cross points should shift with any sway and you should feel it. That would indicate the posts are flexing.

Using 2x4's in an M or W pattern would be easier than more cables.

Bud

 

[frdf250]

Looks like those cables may be touching or close to it at the center of the "x". Get a step ladder or other so you can get up to one of the center locations and have one or some people up top wiggle to aggravate the sway. Those cross points should shift with any sway and you should feel it. That would indicate the posts are flexing.

Using 2x4's in an M or W pattern would be easier than more cables.

Bud

That's a great idea with the cables. I will take a sharpie and draw a line across the two cables to visualize any movement.

Is there a science on how far to space the M or W pattern. In my head it would make sense to have them as close to a 45 degree angle as possible. Right now I have 4 if I was to go to 6 it would look like the attached picture. I am not sure if that spacing is to close?

 

[HotRodx10]

I'm not sure how stiff the attachment between the ledger board and the joists is, but I suspect that may be where you're losing stiffness. I would suggest rather than the multiple Vs, run a single diagonal brace corner to corner, securely attached directly to the ledger. Obviously, it won't be a single piece, but putting the boards side by side and having them bypass each other (overlap), where end the of one and the beginning of the next are attached to the same joist, should be adequate to maintain continuity. To help with the continuity, you can just lay a wider piece on top where the diagonal boards overlap and screw both boards into the wider piece, i.e. with 2 2x4's overlapping lay a 12" long piece of 2x8 parallel on top, between the joists, and put 4 or 5 screws through each of the 2x4's into the 2x8.

 

[SPS-1]

I sketched out the posts and X bracing. (Just guessing at the actual dimensions.) You can get a feel for stiffness by how much the cable would need to stretch to shift the posts 5 degrees.

Indeed, Bud is right, anchoring at the bottom would have been stiffer. Not by a heck of a lot (18%) but it also would have allowed mounting the eyebolt into concrete rather than wood. Also, since a cable provides no strength in compression, it would have been roughly twice as stiff if those X members were solid rather than cable (half the cables are always providing no resistance to movement).

But I still wonder if it is really moving.

 

[HotRodx10]

Stiffening the columns at the outside edge, with longer or bigger cables and such, may be moderately effective, but the most effective way to stiffen the system, is to limit the racking of the deck itself. A continuous diagonal strut (or cable X bracing running corner to corner both ways), will accomplish this in a fairly straightforward way.


I took a closer look and realized you have a beam near the house side. The diagonal(s) could go corner to corner between the steel beams. If there's more than 2 or 3 feet between the ledger and the first beam, you may want to add a diagonal in there as well. The racking will be lengthwise, so the shallower the diagonal, the more effective it will be. the reason the V pattern you have is only somewhat effective is that the the angle is too steep (the angle between the bracing and the joists is too small).

 

[3onthetree]

You've done a really nice job with workmanship and I commend you for using steel, it looks absolutely fantastic, can't wait to see your railings and stair. However, here is a list of things I see, that can all contribute to movement and swaying on such a big deck. This isn't a nitpick, just observations to give feedback, verification, or more info.

- Beam span: The wideflange beam looks to be spanning approx 14+10+10+14 (I see to avoid the windows). You would not be able to do a 14' span with nominal wood, you would need a glulam. Don't know the size of the W, from the pic the flanges look thin but could be the perspective, so hopefully it is sized accordingly.
- Joist size: Not stated, but 14' between wideflanges a 2x10 for SYP is right there, so you may get some bounce unless bumping up to 2x12. Cantilever seems fine.
- Blocking: Joists need blocking between every space when cantilevered. I would also put blocking on the beam near the house because of the long 14' span. Assuming they are only toenailed into the 2x6 sill, you don't have a lot of meat to fasten into.
- Post height: Not stated, but 6x6 could range from 11'-14' depending, 8x8 probably fine up to 14' depending on calcs.
- Bracing: The eyelets for the cables are basically hinges. Even with them fully tight (to the limit of your strength), there is enough play for them to rotate when in compression. You need wood bracing with a rigid connection on the end posts to the beam. The middle posts would receive more bending from braces, so it is not recommended, but with 8x8s they should be sufficient (without doing calcs) to counter that. For tall post heights, I personally prefer the X bracing, but I see you also have the 3x4 W (or channel?).

 

[frdf250]

The support from everyone is much appreciated.

SPS-1 I am assuming you are an engineer? The calc you did on the cable stretch is great thank you.

3onthetree answers to your questions below -

- Beam span: The wideflange beam is a W12x14
- Joist size: Joists are 2x10
- Blocking: Right now the only blocking I have is in the middle of the 14' span. I am interested in understanding why the blocking at the end of the cantilever would help. Not challenging just want to try and understand what the does. Trust me I am all about safety so if its important i want to understand it and add it.
- Post height: the longest one I think is right around 13.5'
- Bracing: Yes there is a piece of horizontal 3x4(i think) I put that in to counteract the cables potentially pulling the posts in when under tension.

HotRodx10 - you are saying remove all "V's" and replace with one big "X" underside corner to corner? In this case it would be from steel beam to steel beam.

I really like the idea of running cable from steel beam to steel beam underneath but worry about applying tension to the cable and having it pull the beams together since the joist are only toenailed to the top 2x6 board mounted to the wide flange beam.

 

[Windows on Wash]

That thing is beefy!!!

 

[HotRodx10]

HotRodx10 - you are saying remove all "V's" and replace with one big "X" underside corner to corner? In this case it would be from steel beam to steel beam.

Yes, either a big X with cables or a single 2x4 or 2x6 diagonal spliced such that it creates a continuous load path corner to corner.



A crash course in one small aspect of structural engineering:


Triangles are very strong, because as long as the lengths of the sides cannot change, its shape cannot change. Further, the forces in the sides of the triangle (and their connections at the ends) are proportional to their lengths. So, in the case of the steep diagonals you have now, the force in the short side parallel to the beam must be resisted by forces in the other 2 sides proportional to their lengths. Thus, if the side parallel to the joists is twice as long as the beam side the force in that side (and in the connections at the ends) is twice the force trying to cause the sway, and the force in the diagonal, being longer still, is even higher (2.23 times higher, to be exact). If the diagonal is made shallower, to say a 45 degree angle, now the force in the joist side is the same as in the beam side, and the force in the diagonal is only 1.41 times higher. Decrease the angle even shallower, by going corner to corner, assuming 20' between beams (I forget at the moment what it actually is), now the beam side is more than twice as long as the joist side, so the forces in the joist side (and again, it's connections) go way down, and the force in the diagonal drops to 1.083 times the force introduced in the beam side (diagonal is 52' long - 52'/48'=1.083).


However, this only works if the corners of the triangles are braced. If you connect the diagonal out in the middle of the joist, for instance, then the joist can bend, the corner of the triangle moves, the entire triangle rotates, and the deck moves. That's why a said you may need a diagonal brace between the back beam and the wall. With the diagonal brace going from beam to beam, that part of the deck is held square, but it may still rack (go from a square to a parallelogram) between the back beam and the house. Probably not, since the joists are connected at the beams, so they would have to bend over that short distance in order for that to happen.

 

[Nealtw]

You have the horizontal holding the posts at a set distance

But they can all bend in the middle together.

Add one more cross brace in the lower section to stop that.
You didn't need the other three the one on the end locks it up.

 

[frdf250]

I added some wood on the bottom today to do a test. This significantly reduced sway I was actually surprised. Underneath when standing next to the cables you could still feel it but on top it was much improved. I only attached the 2x8 boards with about 6 screws each so I would argue that it would be even more improved if it was a through bolt. That being said I am wondering what is the bet option -

1) Add one more row of turnbuckles and cable below the horizontal steel mid post support.

2)Remove the existing cable and run a full X from top of post to bottom of post.

3)Remove cable all together and purchase steel angle iron and through bolt the angle iron to the top and bottom of the posts creating an X in every bay. I could even weld the angle iron to the mid horizontal support if it would help.

 

[Nealtw]

I added some wood on the bottom today to do a test. This significantly reduced sway I was actually surprised. Underneath when standing next to the cables you could still feel it but on top it was much improved. I only attached the 2x8 boards with about 6 screws each so I would argue that it would be even more improved if it was a through bolt. That being said I am wondering what is the bet option -

1) Add one more row of turnbuckles and cable below the horizontal steel mid post support.

2)Remove the existing cable and run a full X from top of post to bottom of post.

3)Remove cable all together and purchase steel angle iron and through bolt the angle iron to the top and bottom of the posts creating an X in every bay. I could even weld the angle iron to the mid horizontal support if it would help.

You got all you needed with the first 2x8 you nailed on.
When we build something like this we put up one angle brace and tie the tops together and get no sway. we brace back to the house for scaffolding.
I would just add one more of you cable set up in one lower area.

 

[Nealtw]

The top half of your posts are held stiff, you just need to lock up the bottom

 

[HotRodx10]

You got all you needed with the first 2x8 you nailed on.

I agree. What Neal showed in the sketch should be all you need, if the bottom of that outside post is well secured to the foundation so that it won't move sideways.



I made my recommendations based on the assumption that you wanted to be able to walk under the deck without obstacles. On that subject, if you secure the joists to the ledger and to the beams (if you haven't already), you don't really need the diagonal braces from the back posts to the front beam.

 

[3onthetree]

- Joist size: Joists are 2x10
- Blocking: Right now the only blocking I have is in the middle of the 14' span. I am interested in understanding why the blocking at the end of the cantilever would help.
- Post height: the longest one I think is right around 13.5'

- Start with American Wood Council guide to deck construction DCA6. This is the standard code, (IRC actually refers back to it) and you can check everything else out, like lateral constraints and joist toenailing.
- Check your 2x10 span, 14' is at the limit for SYP, but users might not be comfortable at that height feeling any 'bounce' (max L/360 deflection is 1/2"). Your midpoint blocking might help a tiny bit.
- It also shows you a detail about blocking the joists at the beam. An applied load causes a member to 'rotate,' so at each joist's bearing point (on the W12x14 beam) the blocking resists that (the rim joist is farther out on the cantilever). I would do it on the W near the house too, even though that cantilever is attached via ledger.
- The end posts only need a diagonal brace parallel to the beam to give the floor diaphragm a rigid connection to its support. If these don't connect to the beam, then you must rely on the beam-to-post bracket you have made to supply the rigid connection (where usually a bracket for this purpose would be in an "L" shape) which would not be adequate in this case. If you go with 'Xs', at least the end ones should attach to the W12x14 (if not all of them just for looks). And the taller the post, the more bending it will have from the loads above, so the 'X' braces are good in this respect (again using 8x8 post should counter the bending these braces will apply). I think you will be satisfied if you remove the cables and just use some steel shapes for the X bracing to give you rigid connections.