Garage Portal Header Splices (plural)

[jar546]

I know the pictures don't tell the whole story. However, a lot of what you can see is obvious. What are your thoughts?
You can click on each image to expand it.

 

[Yankee Chronicler]

 

[Energystar]

Even without the splice, this in no way resembles any type of portal frame.

 

[fatboy]

No way!

 

[wwhitney]

Looks to be 3 plies, and both face plies are discontinuous over one opening?

Is the middle ply continuous over the opening, and if so does one ply have sufficient capacity for the gravity load?

Is the spliced portion of the beam part of the lateral force resisting system, and if so, is one ply sufficient for that?

Cheers, Wayne

 

[jar546]

Looks to be 3 plies, and both face plies are discontinuous over one opening?

Is the middle ply continuous over the opening, and if so does one ply have sufficient capacity for the gravity load?

Is the spliced portion of the beam part of the lateral force resisting system, and if so, is one ply sufficient for that?

Cheers, Wayne

Only an engineer can determine the neutral axis of the beam, therefore, without and engineer (which is the case here) the splice not over a column is a violation of the IRC.

 

[wwhitney]

Only an engineer can determine the neutral axis of the beam, therefore, without and engineer (which is the case here) the splice not over a column is a violation of the IRC.

If the spliced ply could have just been deleted, and still comply with the IRC header span (and braced wall segment) requirements, the presence of the spliced ply is not a violation. Hence my various questions.

Cheers, Wayne

 

[jar546]

If the spliced ply could have just been deleted, and still comply with the IRC header span (and braced wall segment) requirements, the presence of the spliced ply is not a violation. Hence my various questions.

Cheers, Wayne

The connection of the butt-end splices can be compliant if designed properly. They can add plywood or metal plates with through bolts, but it has to be designed.

 

[wwhitney]

The connection of the butt-end splices can be compliant if designed properly. They can add plywood or metal plates with through bolts, but it has to be designed.

It only has to be designed if that ply is structurally required. If it's a bonus ply put there for some other reason, like drywall backing, no design required.

Cheers, Wayne

 

[Yankee Chronicler]

Looks to be 3 plies, and both face plies are discontinuous over one opening?

Is the middle ply continuous over the opening, and if so does one ply have sufficient capacity for the gravity load?

Is the spliced portion of the beam part of the lateral force resisting system, and if so, is one ply sufficient for that?

Cheers, Wayne

I had the same thought, but sanity prevailed. LVLs are expensive. It looks like both outer plies are spliced, so the center ply alone would have to be structurally sufficient. If so, I can't imagine any building using expensive shorter LVL stock to build out a header when the same thing could be accomplished with scrap or cheap dimension lumber from Home Despot.

What do the approved plans show over the garage doors?

 

[ICE]

What do the approved plans show over the garage doors?

I would be surprised if the plans had random splices as we see here. Surprised enough to send the plans back to plan check.

If done correctly, lumber splices can be effective. I have made beams 16’ long from 5/8” T1-11 sandwiched and nailed together. The pieces were 10” tall and random in length, ripped from door and window falloff. I used them for a barn loft to store fruit picking boxes. They were strong.

 

[jar546]

What do the approved plans show over the garage doors?

I don't know, but I do know, it would never show it like that. This was shared to me by the BCO.

 

[tbz]

I have seen the reverse, with the inner used as a spacer and only the 2-outers required and full span the door opening.

But, the simple answer is, to check in the span tables if (1) beam is all that is needed, if so they are good and if not, it fails.

Thus, the tables should be able to cover this, but depending on its location and roof loads, I would have the same questions.

 

[jar546]

Maybe someday, there will be span tables for headers, built-up beams, and girders in the IRC that will have reductions for splices. Until then, the tables are not based on any splices.

 

[wwhitney]

Maybe someday, there will be span tables for headers, built-up beams, and girders in the IRC that will have reductions for splices. Until then, the tables are not based on any splices.

Agreed with respect to IRC Table R602.7(1) "Girder Spans and Header Spans for Exterior Bearing Walls," which includes entries for single ply members and would be applicable to the OP.

But as an aside, IRC Table R602.7(2) "Girder Spans and Header Spans for Interior Bearing Walls" includes entries only for two, three, and four ply members. And if for a given ply size and loading condition, you calculate the ratio of spans, you will find that the 4 ply span / 2 ply span is very close to the cube root of 3, while the 3 ply span / 2 ply span is very close to the cube root of 2. [I didn't exhaustively check this for all loading conditions, just a representative few.]

That would be consistent with the spans being deflection limited (so the cubic behavior) and based on one less continuous ply than the number listed. I.e. on one ply being disregarded, for whatever reason.

A quick scan of the IRC didn't turn up any text discussing this question--did I miss anything? If not, is one ply disregarded because it's allowed to be spliced over any given span? Or is one ply disregarded as an added factor of safety to account for other material/installation variations, and all girder ply splices must be over supports?

Cheers, Wayne

 

[jar546]

Agreed with respect to IRC Table R602.7(1) "Girder Spans and Header Spans for Exterior Bearing Walls," which includes entries for single ply members and would be applicable to the OP.

But as an aside, IRC Table R602.7(2) "Girder Spans and Header Spans for Interior Bearing Walls" includes entries only for two, three, and four ply members. And if for a given ply size and loading condition, you calculate the ratio of spans, you will find that the 4 ply span / 2 ply span is very close to the cube root of 3, while the 3 ply span / 2 ply span is very close to the cube root of 2. [I didn't exhaustively check this for all loading conditions, just a representative few.]

That would be consistent with the spans being deflection limited (so the cubic behavior) and based on one less continuous ply than the number listed. I.e. on one ply being disregarded, for whatever reason.

A quick scan of the IRC didn't turn up any text discussing this question--did I miss anything? If not, is one ply disregarded because it's allowed to be spliced over any given span? Or is one ply disregarded as an added factor of safety to account for other material/installation variations, and all girder ply splices must be over supports?

Cheers, Wayne

Wayne, the span ratios you mention are interesting, and I agree that they seem to reflect deflection-limited behavior. However, engineers put the IRC tables together, and their reasoning or intent isn’t fully explained in the code. They’re meant as a prescriptive method, so we follow them as written without knowing all the assumptions behind them.

When I need to verify spans that fall between the values in the tables, I use tools like StruCalc or BeamCheck. The results can sometimes be surprising compared to what the tables suggest, showing more going on in the calculations than meets the eye.

The simplest way to ensure load paths and structural integrity for splices is to ensure they are over vertical supports—prescriptively.

 

[wwhitney]

Wayne, the span ratios you mention are interesting, and I agree that they seem to reflect deflection-limited behavior. However, engineers put the IRC tables together, and their reasoning or intent isn’t fully explained in the code.

Just wanted to check that there isn't some text describing girder installation in part of the IRC I overlooked, which text might specify that a splice in one ply between supports is already built into the calculated spans and is allowed. I agree that absent such text, all splices of required plies of a girder must be over supports, even though the prescriptive spans in the Table seem to be disregarding one girder ply already.

But I hope we agree that if the girder span table says that (2) 2x12s is sufficient for a given loading condition, and I install (3) 2x12s but let them run wild, so that over each span there are 0 or 1 unsupported splices (always on an outer ply for simplicity, so that the two continuous plies are adjacent), that complies with the IRC. Yes?

Cheers, Wayne

 

[Yankee Chronicler]

But I hope we agree that if the girder span table says that (2) 2x12s is sufficient for a given loading condition, and I install (3) 2x12s but let them run wild, so that over each span there are 0 or 1 unsupported splices (always on an outer ply for simplicity, so that the two continuous plies are adjacent), that complies with the IRC. Yes?

Because I'm "that guy," I'm going to say that I don't agree.

This is a field-observed condition. We don't know what the approved construction drawings showed, but we do know that the code never prohibits building to better than code. This was a debate I had for ten years with a former State Building Inspector when he would stand up in front of classes and proclaim that, "The code is the least you can accept and the most you can require." He generally followed that statement with something to the effect that we could never require more than what the code called for, never mind what the approved construction drawings show.

So look at this as an example. The condition appears to be one of two single-car portals in a one-story garage. Even as a load-bearing (eave) wall, a 1-ply LVL is probably adequate. If it were a double-wide portal, for two cars sharing a single door, that might still work but would obviously increase the stresses in the LVL (both sheer and deflection) significantly.

But we can't know what the owner's future plans are. Suppose the garage is being built as a one-story structure today but the owner knows that he intends to add a second story a few years in the future -- or just wants to build in the option of doing so. So the garage door header is sized on the drawings to allow for the possible future second floor as well as the roof. The contractor doesn't know this, so he doesn't build the header as designed, and as shown on the approved construction documents.

As I understand it, when conducting field inspections we are supposed to be checking for conformity with both the code and the approved construction documents -- whichever has the more stringent requirements. Why else would we require that the approved construction documents be maintained on the site? Using this garage as an example, a framing inspection based solely on the code might find that a single full-length LVL is sufficient --for the initial, single-story structure. But the owner drew (or had drawn) a multi-ply LVL to allow for a future story, so if we approve less than that we are not following the approved construction documents, and we are not ensuring that the owner is getting what he's paying for.

 

[wwhitney]

Because I'm "that guy," I'm going to say that I don't agree.

You gave a good answer as to how the scenario I described might not comply with the approved construction drawings. But my statement was just "that complies with the IRC," not "complies with the construction drawings" or "should pass inspection."

So I don't think you actually disagreed with my statement. Having said that, thanks for pointing out that IRC compliance is not the only criterion of interest.

Cheers, Wayne

 

[jar546]

So the garage door header is sized on the drawings to allow for the possible future second floor as well as the roof. The contractor doesn't know this, so he doesn't build the header as designed, and as shown on the approved construction documents.

Thread Drift Alert

This is an excellent subject to bring up, and it is important to inspect the more stringent of the two, minimum code vs. approved as submitted.

We, as code inspectors, may not know what the future plans are for a building, but the owner and designer do. The contractor should be building to the approved drawings.

I've come across this on a few occasions, once when an electrician was upset because I would not approve a conduit run from the electrical room to the fire pump room because the plans specified two 3" PVC pipes. He only ran two 1-1/2" pipes which were adequately sized for the amount of wire in them. He said, "You can't fail me because I meet the NEC," to which I said, "I'm not failing you under the NEC, but I am failing you under the administration section of the building code for not building to the approved plans." I told him that he could either get the electrical engineer to submit a revision to the plans or he could follow them, but there were no other options. When I returned the following week for another UG inspection, there were two 3" conduits in the trench. The electrician was told by the EE that there is a future expansion planned for phase 2 and that the 3" pipes would be needed. Could you imagine how much money that would cost if something as simple as following the plans was not enforced and they got to phase 2 and found the undersized pipe under a finished slab in a retail store?

On another occasion, one of my friends, who is a BCO, called me years ago. He was sued because he inspected a new construction SFR to the minimum codes and not the approved plans. The owner and contractor got into a dispute, and a lawsuit followed. He asked me to be an expert witness, and I declined. He and the contractor lost the case. This was over the footing size.

 

[wwhitney]

But as an aside, IRC Table R602.7(2) "Girder Spans and Header Spans for Interior Bearing Walls" includes entries only for two, three, and four ply members. And if for a given ply size and loading condition, you calculate the ratio of spans, you will find that the 4 ply span / 2 ply span is very close to the cube root of 3, while the 3 ply span / 2 ply span is very close to the cube root of 2. [I didn't exhaustively check this for all loading conditions, just a representative few.]

To follow up on this aside, I checked the 2018 Wood Frame Construction Manual, and Table 3.24A2 has identical values as IRC Table R602.7(2), while also covering single ply 2x6 to 2x12 header/girders. So that falsifies the idea the table is based on only n-1 continuous plies in a built-up header, as 1 ply would then have 0 capacity.

I think the observations about cube root of 2 and cube root of 3 are just coincidences. I imported all the spans into a spread sheet, and looked at the ratio of allowable spans for n plies vs just 1 ply, for different member sizes and building widths. Pretty consistently the ratios are 1 for n=1 (obviously), 1.48 for n=2 (range 1.4 to 1.5), 1.86 for n=3 (range 1.84 to 1.87), and 2.15 for n=4 (range 2.13 to 2.16).

If we square those numbers (as would be appropriate for the strength controlled case, rather than the deflection controlled case), we get squared ratios of 2.2, 3.45, and 4.62 for n=2,3, and 4, respectively. The last two cases correspond well with n * 1.15, which is the repetitive member factor the NDS allows for 3 or more members. I'm still unclear on why the square ratio for n=2 is 2.2 ratio rather than just 2.0.

Cheers, Wayne

 

[jar546]

To follow up on this aside, I checked the 2018 Wood Frame Construction Manual, and Table 3.24A2 has identical values as IRC Table R602.7(2), while also covering single ply 2x6 to 2x12 header/girders. So that falsifies the idea the table is based on only n-1 continuous plies in a built-up header, as 1 ply would then have 0 capacity.

I think the observations about cube root of 2 and cube root of 3 are just coincidences. I imported all the spans into a spread sheet, and looked at the ratio of allowable spans for n plies vs just 1 ply, for different member sizes and building widths. Pretty consistently the ratios are 1 for n=1 (obviously), 1.48 for n=2 (range 1.4 to 1.5), 1.86 for n=3 (range 1.84 to 1.87), and 2.15 for n=4 (range 2.13 to 2.16).

If we square those numbers (as would be appropriate for the strength controlled case, rather than the deflection controlled case), we get squared ratios of 2.2, 3.45, and 4.62 for n=2,3, and 4, respectively. The last two cases correspond well with n * 1.15, which is the repetitive member factor the NDS allows for 3 or more members. I'm still unclear on why the square ratio for n=2 is 2.2 ratio rather than just 2.0.

Cheers, Wayne

This might be a great subject to reintroduce and expand upon in the engineering section of the forum. Thoughts?

 

[steveray]

It only has to be designed if that ply is structurally required. If it's a bonus ply put there for some other reason, like drywall backing, no design required.

Cheers, Wayne

As much as I want to agree....The prescriptive PFs require a min 3" wide header....Engineering...

 

[steveray]

I try to get everyone that comes in out of portals as I know 85% of the time they are done wrong....This one is at least 85% wrong and the part that is right is that they used framing on a foundation....