Timber Rivets and the NDS... 3

[walterbrennan]

Good afternoon.

The following is the definition of Pw for timber rivets, according to the 2001 NDS, section 13.2.1 (pg 116-117):

"Pw = tabulated nominal wood capacity design values parallel to grain (Tables 13.2.1.A-13.2.1.F) using wood member thickness for the member dimensions in Tables 13.2.1.A-13.2.1.F for connections with steel plates on opposite sides; and twice the wood member thickness for the member dimensions in Tables 13.2.1.A-13.2.1.F for connections having only one plate, lbs."

By this definition, one might assume that if the connnection had a steel plate on only one side (face) of the members being connected, one would enter the tables and reference values listed for a tabulated member which was twice the actual member thickness... which seems very, very counterintuitive, anyway.

Now, throw in the following blurb from the NDS commentary (1997 AND 2001; nearly identical), and things REALLY seem to get all confused:

"... Pw values tabulated in NDS Tables 13.2.1.A through 13.2.1.F apply to connections made with 1/4" side plates and to one plate with associated rivets... Where connections involve plates on two sides of the wood member the... applicable tabular Pw value is doubled to determine the reference capacity of the connection."

Man, oh man... I can't really tell for sure, but it seems like the commentary language reads exactly the opposite of the actual specification language...?!! Spec says, “… tabulated… design values… for connections with steel plates on opposite sides;…” Commentary says, “… values tabulated… apply to one plate with associated rivets…”

Help me out, folks… I am very flustered with this language. Has anyone seen a published design example for timber rivets according to the NDS provisions; by NDS or others…?

Thanks,

walterbrennan

 

[JAE]

OK, I'll take a stab at it (but no snickering if I get it wrong).

Read the first sentence of 13.2.1. It says (I have the 05 Edition) [green]"For timber rivet connections (one plate and rivets associated with it) where: ..."[/green]

It then goes on to say that the connection strength (i.e. the strength of ONE plate with its rivets) equals the LESSER of P_r or P_w.

P_w uses the tables. For a two-plate assembly you are to use the total width b, which is the total width of the member.

So if a plate/rivet assembly is pulling on a piece of wood of width b, then the capacity in the table is X based on just b since we have plates on either side of the member.

Taking an example of 5" wide member with 8 rivets/row and 8 rows per side, Rivet length 1 1/2" sp = 1 sq = 1 (Table 13.2.1A), using b = 5 (a two-plate condition) we get a value of P_w=12740 lbs.

Since we have two plates - the total P_w for the assembly = 2 x 12740 = 25480 lbs.

P_r = 280 x 1.5^0.32 x 8 x 8 = 20402 lbs for one plate.

Since we have two plates - P_r = 2 x 20402 = 40805 lbs and the value of P is the lesser value of the two and equals P_w = 25480 lbs.

If you have only one plate on the same total width of wood b, then the first P_w definition says to use 2b in the tables. We look this up for 2b = 2 x 5 = 10".

The value (for "8.5 and greater") is 11700 lbs.

The value of P_r = 20402 lbs from above.

Since P_w is less than P_r the value of P for the one plate assembly = 11700 lbs.

So for a 5" wide member
with 2 plates we have a P = 25480 lbs
with 1 plate we have a P = 11700 lbs.

This is intuitive to me.

Now - for the commentary. The quote is: [red]"... Pw values tabulated in NDS Tables 13.2.1.A through 13.2.1.F apply to connections made with 1/4" side plates and to one plate with associated rivets... Where connections involve plates on two sides of the wood member the... applicable tabular Pw value is doubled to determine the reference capacity of the connection."[/red]

The first sentence is consistent with the calculations above. The tables are based upon ONE plate. OK so far.

With two plate assemblies, you DO double the value you calculate. You calculate the capacity of ONE plate (using just b)...then double it because you have TWO plates.

With one plate assemblies, you DON'T double the value you calculate. You calculate the capacity of ONE plate (using 2b and compare that with P_r directly.

The commentary just doesn't mention that with one plate assemblies you use 2b to enter the tables....and you don't double your table value.

So the tables are trying to give you a WOOD capacity value and if you have two plates, one each side, the wood doesn't have as much strength to offer than when you have only one plate.


I think your basic confusion lies with confusing:

"doubling the capacity when you have two plates"

and

"using b or 2b".

They are two completely different aspects of the process.


OK - snickering can start.

 

[COEngineeer]

i havent read the whole thing but i think you deserve a star JAE.

 

[RockEngineer]

And a consulting fee. Excellent detailed answer. No snikering from me.

 

[walterbrennan]

Ah, 2b or not 2b; that is... alright, over-used... sorry...

No snickering from me, either, JAE. I really, really WANT to agree with you're approach, and I've turned it around that way more than a few times, myself; even prior to posting the initial question... But then, just as I'm beginning to convince myself that I have a handle on this thing, we hear back from AWC/AF&PA on the question.

During a three (3) separate telephone conversations with one of their technical liaisons, the following transpired…

First call, we read the spec language, chapter and verse out loud… got the first answer, which was, apparently that it stands correct as read.

Second call, the liaison gets back to us, after reading the commentary language; advises us that THAT language is in fact the correct variation; which is applied entirely different than the actual spec language implies (source of my first confusion…).

Third call, the liaison gets back to us to let us know that, between the spec and the commentary, the commentary is actually the flawed item… revisions will (apparently) be issued; via the AWC/AF&PA website, I imagine. We were also told that the table values were, “… already doubled...” We assume they meant for connections involving two (2) plates.

Finally, when asked if any practical design example for a timber riveted joint existed that AWC/AF&PA would be willing to endorse, we were told that some of these were, “… in the works,” and that they would fax us something one of these days. I don’t think I’ll hold my breath.

This is ALL very discouraging, to say the least…

The only bright spot is the location, after some searching around, of a fairly recent (2004) research report (

http://www.fpl.fs.fed.us/documnts/fplrp/fpl_rp619.pdf),

which brings some interesting facts to light, such as…

--- The NDS timber rivet provisions, which were first included in the 1997 spec, are based on theoretical modeling (and limited verification testing) performed in the mid 1970’s; which is the same testing on which the Canadian specification governing the use of timber rivets is based.

--- The Canadians have (apparently) been utilizing timber rivets for several decades with very good success.

--- Unlike the US specs, the Canadian model may include more than just tabular information for determining the limiting wood strength; it may include the (admittedly complicated) equations used to generate the tables. These alone would (probably) settle the question, here…

--- The above-referenced 2004 report includes testing which seems to validate, in larger part the original theoretical models.

I do not, as yet have a copy of the Canadian specifications, but I think I may seek them out. The 2004 report may have answered my questions about the physical behavior of the wood failure mechanism, but it did very little to improve my confidence in the NDS tabular values…

Anyway, that’s what I’ve discovered, to date… besides the fact that (apparently) we don’t really use timber rivets much, south of the (Canadian) border.

I appreciate your stab at this thing, and I would welcome any further insight that you happen upon, regarding this matter. I will report further, if I discover anything useful, as well.

Keep on truckin’…

walterbrennan

 

[JAE]

walterbrennan

Thanks for the information on the phone calls. I would still assert that the spec and commentary are not in conflict (the way I've interpreted it above) and I would suggest that the individual on the phone probably doesn't have direct authorship or intimate knowledge of the tables and spec language either...they are probably just initially flummoxed by the "apparent" discontinuity like both you and I are/were.

But never-the-less - do keep us updated here on this post if you get any further feedback. Not that I usually (or ever) specify timber rivets - but just that clarification for others who read Eng-Tips may be informed, or warned, accordingly.

Thanks,