Strap Reinforcement of Existing Rafter 4

[SRO]

I need to reinforce existing rafters for the addition of solar panels. The simple solution of sistering another rafter along side the existing one isn’t an option due to confinement issues of the attic access (rafter is too long to fit).

My solution:
Attach a Simpson coil strap “CS14” (1/16”x1-1/4” steel) at each end and run it under blocking that would be between the strap and the rafter to create a deeper beam (see detail).

My Process:
•Find required section modulus by solving Mmax = Fb’ x Sreq.
•Transform the steel to the equivalent wood section.
•Determine the neutral axis & new moment of inertia for the transformed section.
•Compared the section modulus above & below the NA to the required S
•Used the allowable deflection to determine the required moment of inertia & compared it to the one for the new transformed section.

My Question:
•Does anyone see a problem w/ this solution and is there anything I overlooked?

 

[SteelPE]

I know you already dismissed this idea… but have you checked to see if the interior partitions can take any additional load due to a strut being added in the attic? I would think this would me a much more cost effective option than ripping the roof off.

If that doesn’t work, is there anything else that can be done? Have you really dug into the code to see if you can gain any additional capacity or reduction in load? I know Massachusetts had a code change a few years ago. I also know that they have made multiple revisions to Chapter 34 pertaining to the addition of solar panels to an existing roof.

If you have exhausted all possibilities then I guess you truly are left with ripping the roof off.

 

[Prestressed Guy]

Sorry SRO, but I can’t agree with your assessment that torque is better for measurement of tension than elongation. Measurement of toque isn’t a good method of measuring tension due to unknown values of loss from friction and other factors. See the commentary on 16.2-49 of the AISC 13th for why this is unreliable. The blocks that I refer to use the principle that tension is proportional to elongation by means of (delta L) = (PL)/(IE) to determine tension. This is a very predictable and accurate way to pre-tension steel.
Determine how much tension (P) you need. Length (L) is the distance between the ends of the anchorage zones. The Moment of inertial (I) is the biggest challenge because you need to get an effective I due to the presence nail holes along the strap but you can make a good approximation with a ratio of (L*net I / L*full I). Once you determine how much elongation you need, some high school trig will tell you how tall to make the blocks to get that elongation.
Not very difficult to calculation and install, yet very accurate.

 

[SRO]

SteelPE:

I don’t think it’s an issue of ripping the entire roof off, just an opening in the roof or siding (maybe a gable vent) to pass the lumber into the attic. Which, by the way is starting to look like a better and better option.



Haydenwse:
I'll have to check out that section of the commentary.

My concern w/ the using the tension in the strap is that I was designing the strap to be 7-1/4” out from the bottom of the rafter. In order to obtain that distance, the strap would have to be attached w/ a lot of slack in it. My concern is that if I give them;
•the length of strap required between attachment points,
•the distance (along the rafter) between the attachment points,
•the depth of and distance between the blocks with the precision required to determine the tension,

is this going to be do-able from a construction stand point or is it going to be like giving out a foundation plan w/ dimensions that have 32nds of an inch on it?

 

[bones206]

Maybe you could get lumber in through a gable end louver to avoid opening the roof.

Just an idea - could you add bridging between the rafters along the fastener lines of the solar panels to share the load among the other rafters?

 

[Prestressed Guy]

If you want the strap to be 7¼” out you can install it in a taught condition over a pair of 2x8 blocks that are spaced closer than the final designed position. They would then be pushed outward to attain the correct elongation. The build would be fairly easy. The hard part is the calculating geometry to figure out where to place the blocks initially and how far they need to be pushed outward to get the correct tension. But that’s why we get paid the “big bucks” ;>)

If you run the strap over a pair of 2x8’s I would ease the edges to avoid a hard bend in the straps. You will also want to fix the blocks in place with plywood gussets on each side. I would also suggest running a short section of strap nailed to the sides of the rafter over the in-board end of the anchorage zone to take the vertical component of the tension. I personally would not trust a couple of screws in the last holes for that.

The suggestion from bones206 of distributing the load to adjacent rafters on the inside will also work. I wouldn’t use bridging though. I would install a strong-back to the under-side of the rafters at each intersection. Depending on the load, the connection could probably be made with Strong-Tie TS straps.

 

[dcarr82775]

Haydenwse,

I don't agree that it is that simple. In a perfect world yes, you could calculate elongation. This isn't a straight item (I didn't look it up, but I assume it is a coil strap). The amount of elongation would probably not rather small, and there will slack from installation and slip at the end anchorages. That would make calcing elongation a bit of a crap shoot.

If using a turnbuckle (and threaded rods), after the slack is taken out of the system you could calculate number of turns of the nut that would be required. That would give a little more accurate measurement of PT. But it seems to me there has to be an easier way to skin this cat (interior partitions, partial length sistering, etc)

 

[dhengr]

SRO:
You are kinda missing the point, Haydenwse is still right on the money, I generally agree with him. Except to add, reread my first post, the correct thoughts are there, but you have to read btwn. the lines a bit, so let me try it with different words. Your original scheme is O.K., but you are thinking about it from the wrong angle. However you do this, at this stage, it is a bit of a seat-of-the-pants process, but with some good engineering judgement and attention to detail you can arrive at a satisfactory solution, albeit not an exact solution. Applying the strap to the bot. of the rafter by hand, even over your blocks, would buy you nothing. The rafter would have to deflect too much to bring the strap into play and to overcome the relaxation at your end connections. And, the strap would only start picking up added load after that delta. Jack the rafter up enough so can call it a flat arch, apply tension to the strap by mech. means, nail it off, remove the jacks, and you have a tied arch and can make some attempt at calc’ing. its thrust (tension in the strap) and watch the end connections relax and the arch settle, again an approx. which adds some cap’y. to the rafter.

Torque on the turnbuckle is an approx. for the reasons AISC and Haydenwse suggest, friction, etc. But, if you know the pitch of the threads, then three full turns of the turnbuckle after you release the wrench and the rods or strap detwist, means so much elongation of the rod, which suggests a tension. T = E*A*rod elongation/rod length. Except, upward deflection of the rafter, slack in the rod, and end connection relaxation effects must be subtracted from your final tension calc., and these are somewhat unknown, thus your engineering judgement is needed, for a final approx. A string line on the rafter gives you upward delta, tick marks on the strap and rafter at each end gives you their movement, but add a little more because this is a long duration loading at the connection, thus more relaxation over time. After you’ve done a couple, I’d have a good feel for the rafter delta, but I’d continue to watch the end connection movement. For example, using 6" blocks, 4' on either side of the center of the rafter gives you rod or strap geometry, you make a reasonable estimate of the strap tension, and this imparts two upward forced on the rafter, increasing its moment cap’y, as long as it can take the new axial compression and stresses around the end connections and they don’t relax too much.

Haydenwse’s wedges and blocks method works too. I would apply some mech. tension to the strap to remove any slack, and nail it off at the ends. Now start inserting wedges 4.5' from the rafter center; watch the rafter delta upward and the end connection movement, and calc. the tension. The new strap geometry will suggest an approx. change in length of the strap once the above unknows are included, and thus the upward forces a 4' to give you upward moment you want. Insert the proper ht. blocks at 4' from the center.

 

[Prestressed Guy]

One correction to my earlier post. The equation should be (delta L) = (PL)/(AE) substitute A for all I’s. Oops, that must have been an old-timers moment. The delta L and geometry is directly calculated and dhengr is correct that you will need to add in additional elongation for losses.

I also agree that the turnbuckle idea will work just fine. You just can’t use torque to determine tension. You have to go back to elongation. The biggest deterrent to that system is that it is very expensive compared to other alternatives.

 

[SRO]

I’ve read the AISC commentary section and can see everyone’s point about using torque.

dhengr:
I love the idea of using the pitch of the thread to gage the tension.

I’m pretty sure I see what you mean, but just to verify, let me run through the process.
•Install the strap as tight as possible at approximately the depth away from the bottom of the rafter.
•Snap a chalk line along the rafter to later measure upward deflection.
•Put a couple of tick marks at the ends to later measure the slip at the anchor points.
•Start inserting temporary wedges slightly further from the center then the location of the permanent blocks.
•Continue inserting wedges to stress the strap to the tension required (while subtracting losses due to upward deflection, and slip at the anchor points.
•Once we get the tension required, cut and install the permanent blocks and remove the wedges.

I think I prefer the rods to the straps, but Haydenwse is right I think they are cost prohibitive.

 

[Prestressed Guy]

I have attached a detail showing how I would get the pre-tension into the strap.

 

[dhengr]

Haydenwse: Very nice schematic sketch, well thought out. And, we even agree on the formula for approx. tension now. The straps to pick up the vert. component are far superior to a few screws, and note that they must be designed to withstand the same vert. force your blocks impart on the rafter to give you the upward moment from two point loads (your 2x8 blks.) on the rafter. You might want these vert. point loads to approx. the solar panel reactions on that rafter, the new added loads. Then if you can design the end connections, quite conservatively to minimize their movement, and take those new end stresses on the rafter and the new induced compression in the rafter (now a beam column); the net result is adding an upward moment approx. equal to that caused by the new loads downward.

SRO:
You’re heading in the right direction. The thread pitch/turn of the turnbuckle works reasonably well, with a reasonable degree of confidence, when you start out with a slightly pre-tensioned rod (any sag already removed), with end connections which do not slip or relax, and where rafter deflection is not too great wrt the tension applied; since all of these are factors which you have some trouble accounting for in an exact fashion, on this problem. On larger jobs, the saddles which harp the tendons and the tendon end bearings are fabed. and much less yielding than what you have in your problem. And, they actually pull the rods and measure their elongation or read the pressure on the pulling jack, then run nuts tight onto the bearing to set the tension. You must pre-tension the strap a bit; whether it’s flat to the bot. of the rafter (then you use wedges to move it away and really tension it) or over precut blocks and you move them longitudinally, you don’t want to use half of your tensioning movement just to take the slack out of the strap.

All of the calcs. are done before hand so you know how to design your end connections, and they are the most difficult part of the problem. You can’t get enough nails in them, and their movement is so critical. We normally think nothing of a 1/32 or 1/64" movement in a wood framing joint, but that’s a lot in terms of steel elongation and final tension. What tension change does 1/16" elongation change cause? What elongation or length change is caused by a 1/16" rafter deflection change? Without knowing your rafter length, new loads and exact geometry, you will find that the strap tension gets pretty large to impart 100 or 200# upward on the blocks. Thus, the end connections become difficult quickly. The end tick marks must be sharp and accurately done because you will be looking for fairly small movements, with large effects. Do a calc. with 100 or 200# block loads, what deflection change does this cause? Snap the chalk line, or make a few fine marks, and leave the tight string in place so you can actually watch the deflection as you do the tensioning. As much as anything, this deflection might be the best indicator that you have achieved the result you wanted. But, you have to make adjustments on the fly, during the tensioning process, so you have to make tick mark readings on the fly too. Finally, after factoring these things together, you make the judgement that you have enough tension, and then hope no body asks you exactly what the tension is or will be in a year. But, again, think conservative end connection design, and remember nails in shear parallel to grain are not as good a nails perpendicular to the grain. And, on the first couple rafters I would check these movements and deflection changes for a few days to see that they hadn’t changed much.

You could use aircraft cables, one on each side of the rafter. Then you need matching thimbles, cable clamps, and long turnbuckles, but this might still be no more expensive than a good grade of strapping and all the problems associated with its end connections.


 

[SRO]

Thanks for all the help.

 

[Prestressed Guy]

Dhengr, it sounds like you also come from a prestressing background. With 7 years as project engineer for a major west coast prestress mfr, this stuff is second nature. I was not trying to derive all of the engineering and calculations but just show a scheme that would work. Your comments about small movement in the anchorage are exactly right. The one thing that helps greatly with this is that the modular ratio of steel/wood is so high that a little steel will go a long way, especially with the 7” increased depth.

The issue of losses of prestress can be a bit overwhelming the first time you look at them. That is why I suggested in an earlier email, that maybe a strong back under a group of rafters to distribute the load out to additional joists might be a simpler solution. There again, the challenge is stiffness. The strong back should be designed for stiffness instead of moment capacity.

 

[mbullism]

Just a small late add, slightly off point... MA has within the past couple of years adopted a new building code, and made revisions to chapter 34 regarding snow loads and PV panel installations ... I don't see it noted above, but the revisions were made to Chapter 34, the general code, not Chapter 93 covering existing residential work. If the rafters above are in a commercial building there is a reduction available... If they are in a house I'm not aware of a similar reduction available.

As to the issue at hand, I'm adding lumber... .02