GRK RSS Screws - Lateral Design Values - Test Data versus NDS Calculated Values 2

[Eng16080]

Based on a thread from last week in which the use of proprietary wood fasteners was brought up, I decided to take a deeper look into GRK RSS screws. I've noticed in the past large differences between manufacturer data and the lateral design values found in NDS Tables 12J and 12L for lag/wood screws.

Based on test data from ICC-ES Evaluation Report ESR-2442, Table 3, I created my own new table (see attachment) comparing the test data to values calculated per NDS 2018, Section 12.3. With a few exceptions, the test data indicates greater fastener strength than predicted by NDS. This isn't surprising, but the difference between the values in many cases is more significant than I would expect. In most cases tested values are at least 50 percent greater than NDS values. Several values are 2 or 3 times greater.

I'm curious if other engineers use this test data considering the huge difference compared to the NDS values in some cases. For a 5/16"x6" RSS PHEinox fastener, for instance, with perpendicular to grain loading (ledger type connection), it seems like you could used half or less screws versus standard screws calculated per NDS. (I suppose that's their whole selling point!)

The following are a few potential risks that I can think of with using the test values:
[ol 1]
[li]If the builder does not use the specific proprietary fastener called for and the substitute does not have similar test data, then the number of fasteners required could be significantly more, and the connection will likely not pass.[/li]
[li]If a connection were to fail and it was found that the field conditions did not exactly match the test conditions. For example, a 5/16" RSS fastener has test data for side member thicknesses of 1.5" and 2". If the side member were really 1.75", could an intermediate value be interpolated? Engineering judgement of course says yes, but if there was a problem that resulted in litigation, how would this hold up in court?[/li]
[li]What if a future evaluation report indicates that previous test values were too high? I assume that the evaluation report is an extension of the building code and that the engineer would not be liable for any errors, although I don't know.[/li]
[/ol]

I realize that I'm probably being overly cautious with some of this, but it seems a little crazy that we can suddenly use significantly less fasteners in most connections. Any comments are appreciated.

 

[Eng16080]

Attachment here:

 

[phamENG]

It's important to note that, while the test report values are based on testing, they are not the raw test values. They are modified with factors of safety as appropriate.

To address your stated concerns:

1) This is certainly a risk. But it's also a risk that they'll use 10d (or even 8d) box nails everywhere you tell them to use 10d common. There's still a fair amount of work done in this industry that's based on trust. Don't give the contractor too long of a leash and make sure your drawings are complete and easy to read, but at the end of the day you have to trust them to build it to the plans you provide. Verification is great, but not always possible ("Hey, I just finished building this building and the inspector has questions...you're name's on one of the sheets, what do I do?" are always fun phone calls).

2) Good test reports have conditions of use. If you don't meet them, don't use it. If it doesn't say you can interpolate, don't. Use the more limiting option.

3) Test reports, even ICC-ESR reports, are not extensions of the building code. Remember, the IBC is not the building code. It's a document produced by a company in business to make money (ICC). They have a very official sounding name that makes them sound like an authoritative body (good marketing!) but they are not. If you're in the US, the building code is a law passed by your state, county, city, or town as applicable to your locality. For instance, my drawings will get rejected by plan reviewers if I say the IBC or the IRC is the governing code. It's the Virginia Construction Code or the Virginia Residential Code. Now...those codes are 97% IBC/IRC, but that doesn't matter. The IBC and IRC don't mean anything unless a law has been passed making them (in whole or in part) part of the building code for the jurisdiction in question. The test reports are usually not included in that. What a test report does is a) provide the EOR/AOR the piece of mind that the product they have selected meets appropriate engineering standards and can be used in design and b) is usually a faster route to approval in jurisdictions with plan reviewers that pay attention and either know what they're talking about or pretend to know what they're talking about and just ask a bunch of questions.

I specify proprietary screws and use the capacities listed in their catalogs/test reports. It's more accurate than generalized screw calcs and allows me to take advantage of proprietary tech in the screw designs.

 

[Eng16080]

phamENG, thanks for the responses.

1. I agree that anything we specify on the plans has the risk of getting messed up in some way. Trying to minimize that risk is, of course, a good thing. Using the wrong nails is a good example and is another item on my list of things to look into a little closer. I assume we have a factor of safety of around 3 for most aspects of our structural designs before there start to become serious issues (serviceability aside).

2. Agreed. The report states exactly that. Basically, the tested screw strengths are only applicable to the exact conditions tested. Other sections of the report disqualify those values from being used in engineered wood products (except the noted glulams and CLT), etc, etc.

 

[skeletron]

Proprietary STS fasteners, GRK specifically, are particularly challenging here in Canada. The major diameter exceeds that of a #12 screw so CSA O86 kicks them over to the lag screw provisions which use a more restrictive reduction factor and also require the designer to check directional strength. GRK tables have fine print that indicate the values are based on "wood screw" provisions not "lag screw" provisions, which makes their marketing comparison a little unfair but *probably* justified with testing. Unfortunately, Illinois Tool Works (GRK's parent company) has not pursued CCMC compliance testing, which is kind of the equivalent to the ESR report. Here is an FAQ from MTC Solutions that goes further into this deviation.

I use the manufacturer's published data as a reference in design. My secondary check uses the standardized equations and geometry/material information available. Usually they are close and able to be correlated within reason. I typically use them in residential construction for ledgers or plate connections because contractors prefer to use a STS rather than a wood screw or lag screw.

With respect to your situational risks presented:
1. I call out the exact fastener that I want to see. I don't specify fasteners that can't be picked up from the local lumber yard or box store. The situation you are describing is usually covered by common general notes clauses.
2. I would avoid this situation by using the more conservative value. I think if your design is riding the line dictated by a 1/4" embedment difference, then your design may be unnecessarily value engineered for the tolerances of residential construction.
3. I think this is being overly cautious. Similar to #2 above, the design may be unnecessarily value engineered for the tolerance of regular construction.

 

[EngDM]

The major diameter exceeds that of a #12 screw so CSA O86 kicks them over to the lag screw provisions which use a more restrictive reduction factor and also require the designer to check directional strength.

Have you considered using the GRK RSS screws which have been developed with the lag screw provisions of O86?

 

[dik]

Does anyone know how a #14 screw using lag screw provisions compares to a #12 screw? Just curious...

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[Eng16080]

Does anyone know how a #14 screw using lag screw provisions compares to a #12 screw? Just curious...

I can't answer that question specific to Canadian requirements, but in terms of U.S. codes and specifically NDS 2018, I'll offer my understanding. The distinction between a "wood screw" and "lag screw" per NDS is that a "wood screw" has a diameter less than 1/4" and a "lag screw" has a diameter of 1/4" or greater. This is the overall diameter, not the root diameter (which is usually less).

If we consider a single "lag screw" with a diameter of 0.250" and a single "wood screw" with a diameter of 0.249" with all other properties being the same:
[ol 1]
[li]For loading parallel to wood grain, the dowel bearing strength of the "lag screw" is greater than the "wood screw" for a specific gravity, G, of 0.62 or less. The dowel bearing strength is the only value that would be different in the lateral yield limit equations. Therefore, the "lag screw" will be stronger in terms of loading parallel to wood grain for G of 0.62 or less. [/li]

[li]For loading perpendicular to wood grain, the dowel bearing strength of the "lag screw" is close to that of the "wood screw" for values of G below 0.50. For greater values, the dowel bearing strength is greater for the "wood screw." Besides the dowel bearing strength, there is also a reduction factor, Rd, used in the lateral yield limit equations. In this particular case, the value of Rd for the "lag screw" is greater than for the "wood screw" leading to a greater strength reduction of the "lag screw". I believe the net effect is that the "wood screw" will always be stronger in loading perpendicular to grain.[/li]
[/ol]

In terms of real numbers, I calculated the lateral strengths, Z_parallel and Z_perp, for a single shear connection with G = 0.50, a dowel bearing length of 1.5" in the side member, 2" in the main member, and a dowel bending yield strength of 70,000 psi. I compared a "lag screw" with D = 0.250" and Dr = 0.173" to a "wood screw" with D = 0.249" and Dr = 0.173". All other properties between the two fasteners are the same. Technically, the dimensions of the "wood screw" don't match those of a real wood screw but rather the "lag screw" for comparison purposes. The following were the results:

"Lag Screw"
Z_parallel = 153 lbs
Z_perp = 110 lbs

"Wood Screw"
Z_parallel = 140 lbs
Z_perp = 140 lbs

 

[skeletron]

Does anyone know how a #14 screw using lag screw provisions compares to a #12 screw? Just curious...

For #14 x 3.5" screw into 2x SPF (G=0.42), K=1.0
Wood Screw (CL 12.11.3)
Nr = 329 lbs with Mode F governing (directionally independent)

Lag Screw (CL 12.6.5)
Nr = 85 lbs with Mode F governing (perp side, perp main)
Nr = 128 lbs with Mode D/E governing (perp side, para main)
Nr = 138 lbs with Mode F governing (para side, perp main)
Nr = 192 lbs with Mode F governing (para side, para main)

In comparison, a 5/16" GRK in similar conditions would have Nr = 273 lbs under wood screw provisions and Nr = 194 lbs (max) to 86 lbs (min) under the lag screw provisions (directionally dependent).

The difference between CL 12.11.3 and CL 12.6.5 are:
[ul]
[li]phi = 0.8 for Wood Screw, 0.6 for Lag Screw[/li]
[li]Substantially different embedment strength values based on the direction of the force[/li]
[li]Slightly less embedment strength value for point side member when force is parallel to the grain[/li]
[li]JPL factor for reduced penetration between 5dF and 8dF[/li]
[/ul]

Have you considered using the GRK RSS screws which have been developed with the lag screw provisions of O86?

GRK does not list these values. I've found that if you run the numbers as a lag screw, it doesn't give any benefit to using them. This is counterintuitive, in my opinion, since an STS has design features that make it a better choice than a lag screw in many situations.

EDIT: Sorry to drive the tangent on this thread away from NDS.

 

[dik]

Thanks very much...

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[Eng16080]

If I'm understanding correctly, per the Canadian requirements, a wood screw and lag screw of roughly the same size have vastly different capacities, by a factor of almost 4 in perp. to grain loading. Seems like a big penalty for using lag screws.