Open Web Steel Joist Top Chord Diaphragm Strut - Tie Angle Detail 1

[r-struct]

When the top chords of open web steel joists (OWSJ) are used as boundary elements (struts/chords) of a diaphragm, one detail we are looking to use to transfer axial forces between OWSJs is that of tie angles as shown below.

This is consistent with the suggested detail from some joists manufacturer to have a more direct load path for the axial load from one OWSJ chord to the next OWSJ chord, rather than passing the axial load through the joist seat. One other option is to have a tie plate on top of the OWSJ top chords but this option would cause some slight interference to the steel deck above.

A few concerns/questions I have with the use of tie angles, for anyone who have used these in the past or who have looked at this in detail previously:

1. How is the minimum length of the angle at each OWSJ side established?
2. When checking the capacity of the tie angles, can its unsupported length be assumed to be close to zero or should it be close to 1200mm based on the sketch? The tie angles would almost be continuously supported by the OWSJ top chord which is also practically continuously supported by the steel deck diaphragm.
3. With the angles loaded only through the intermittent welds shown, I am a bit concerned with the eccentricity of how the axial load is delivered to and from the angles.
4. Should it be a concern that the angles will be held in place only by the same welds that are used to transfer the axial load?

Any further insights and comments with this detail is appreciated. Thanks in advance.

 

[HTURKAK]

Axial load 70 kN implies the top chord is a drag strut and part of the bracing system that is, should be designed for gravity loads and axial forces. The field welded tie consisting of a pair of plate near the neutral axis of the top chord angles should be preferred.

 

[Once20036]

I`ve done something similar, so i`ll chime in:
1) The 600mm? Minimum length here is based on getting enough weld to get the force out of the chord and into the angle, then back out on the far side.
2) For a plate member (with deck interference), I use the unsupported length as the clear distance between the two joists, plus an inch or two on each side. For your angle it might be a little bit different given that the vertically down leg is not welded for the entire 1200mm length. I suppose this gets into the question of why you`re using an angle, and if you could use a plate that doesnt have that downturned leg.
3)You could consider a rod member, tucked into the corner of the angles. With a double row of flare groove welds, you wouldn't have any eccentricity (or downturned leg) to worry about. We often reinforce joists to increase axial capacity for added loads in this manner.
4) Compared to the 70kN axial load, I wouldn't worry about the self weight of the angle.

Axial load 70 kN implies the top chord is a drag strut and part of the bracing system that is, should be designed for gravity loads and axial forces.

Agreed 100% - don't forget about this axial force at the joist midspan where you also have maximum chord axial force from bending.

 

[KootK]

@OP: first off, these are some great and insightful questions. Wherever you're working, they're lucky to have you.

1. How is the minimum length of the angle at each OWSJ side established?

A) Force transfer in and out as Once20036 mentioned.

B) Maybe give consideration to a little lateral shear lag from the welds to the centroid of the angles.

C) If this were my baby, I'd want to:

i) run the tie members back to at least the first, interior top chord panel point and;

ii) use a tie with some meaningful flexural stiffness perpendicular to the roof plane.

These thing would allow you to shield the top chord from developing applied moments between the top chord panel points to some degree.

2. When checking the capacity of the tie angles, can its unsupported length be assumed to be close to zero or should it be close to 1200mm based on the sketch? The tie angles would almost be continuously supported by the OWSJ top chord

D) It depends. At least one of the OWSJ top chord and the tie member needs to be designed not to buckle in the vertical direction when the compression introduced by the tie is considered. If you instruct the OWSJ supplier to handle this then the angle don't need to. Otherwise, I would assume the unbraced length of the angles to be the distance between top chord panel points for buckling in the vertical plane.

3. With the angles loaded only through the intermittent welds shown, I am a bit concerned with the eccentricity of how the axial load is delivered to and from the angles.

I agree and that informed much of my response to [#1] with respect to shear lag and minimizing the effects of bending in the OWSJ top chord.

4. Should it be a concern that the angles will be held in place only by the same welds that are used to transfer the axial load?

E) Yes, and in my opinion, this is really the Achilles heel of your detail as shown. The angle ties need to be braced torsionally and, as detailed, the only thing performing that function is your fillet welds resisting moment about their own longitudinal axis which is generally a weld design faux pas. This point alone is a great reason to use either a flat plate or or a channel.

Any further insights and comments with this detail is appreciated. Thanks in advance.

F) My preference here would be to use a channel with down turned flanges for the tie piece. That, in homage to the points that I made above:

i) torsional stability of the tie piece.

ii) meaningful, bi-directional stiffness of the tie piece to shield the joist top chord from additional bending and buckling effects.

With this, it's prudent to keep in mind that welding cost will dominate this and material costs will be immaterial. So, if it's spatially acceptable, there's little reason not to throw an MC10 in there and give yourself some stiffness and room to get things done. I was also thinking that that it might be nice to plug weld the channel to the OWSJ flanges from the top but I have a small amount of concern for temporarily melting the flanges as with transverse welding which we generally avoid in these situations.

G) I've not had great luck getting OWSJ suppliers to design their products for special conditions. They'll do it, for sure, but the price tends to shoot up in a hurry. For that reason, I try very hard not to ask for anything that would truly require any modification to the base design.

H) Keep in mind that you won't be able to keep the tie force in the top chord. A share of any imposed compression is going to shoot down to the bottom chord of the truss as soon as it encounters the first OWSJ diagonal web. What that means is that there may be an increased demand for OWSJ bottom chord bracing/bridging at the ends of the truss where the bottom chord flexural tension may not yet have overcome the imposed compression. I don't expect that additional top chord compression will cause you any grief because the top chord will normally be designed for the mid-span flexural compression demand which may well be an order of magnitude larger than anything you'll be adding with your collector design.

 

[KootK]

Just some sketches to go along with my previous post. Obviously, one needs to consider the impacts of:

1) Overhead welding.
2) Weld access.
3) The welding temporarily compromising the chord.

@OP: I actually suspect that your detail has come about from a desire to preserve a top-side welding condition. That's a big deal and may well keep you on that path.

 

[r-struct]

Thanks for your responses HTURKAK, Once20036, and KootK!

KootK, I appreciate your kind words and your detailed and thoughtful response. All the insights you've sshared will be considered and will not be taken lightly. Every EIT would be lucky to work with a senior engineer like you. Will probably go with the tie channel or tie plate option. Some follow up items though, if you're still up for another round of sharing your thoughts:

@OP: I actually suspect that your detail has come about from a desire to preserve a top-side welding condition. That's a big deal and may well keep you on that path.

1. You are correct that the aim is to have top-side welding condition only.

D) It depends. At least one of the OWSJ top chord and the tie member needs to be designed not to buckle in the vertical direction when the compression introduced by the tie is considered. If you instruct the OWSJ supplier to handle this then the angle don't need to. Otherwise, I would assume the unbraced length of the angles to be the distance between top chord panel points for buckling in the vertical plane.

2. Say the added top chord compression/tension (i.e. collector force) was instructed to be taken care of by the OWSJ supplier, do you think the 'meaningful flexural stiffness' in the direction perpendicular to the plane of steel deck can be slightly relaxed? Example, with the given condition, if a C8 with only top plug welds is used and overhead welding from your detail is omitted, will KootK still approve (WKsA)? I am thinking that if some additional axial compression due to the collector is specified to be resisted by the OWSJ, some thicker than usual chord material may have to be used - and this can somehow mitigate the concerns of compromising the OWSJ top chord due to welding.

3. If it exists and not only from experience, will you be able to point me to the reference where the 5/16" thick max. limit for plate thickness to minimize interference with the steel deck can be found?

Thank you.

 

[KootK]

2. Say the added top chord compression/tension (i.e. collector force) was instructed to be taken care of by the OWSJ supplier, do you think the 'meaningful flexural stiffness' in the direction perpendicular to the plane of steel deck can be slightly relaxed?

Yes, it could be entirely relaxed I think. And now that I've noodled on it a bit longer, I think that may well be a zero cost ask of the joist fabricator. I would assume that, for vertical buckling, it would be the top chord, mid-span panels that would govern the design of the compression chord. With similar panel point spacing at the ends of the truss, one would expect considerable reserve capacity there automatically. This is kinda pushing me in the direction of liking the flat plate option.

...if a C8 with only top plug welds is used and overhead welding from your detail is omitted, will KootK still approve (WKsA)?

Frankly, KootK is not sure if KootK approves of the plug welding. I'll have to noodle on that a bit more to decide if I'd really be comfortable with plug welding the tops of the chords. Obviously, when deck is spot welded to the top chords of joists, it is a very similar situation (Link). Maybe look into the minimum diameter of a true plug weld and see how that compares.

Given that you're in Alberta, like me, you also have to face the possibility that there's a 30% chance that the bidding process will result in your joists switching to Omega joist rather than Canam etc. Then you've got to deal with those pesky, cold formed hat sections for chords. Yum.

If it exists and not only from experience, will you be able to point me to the reference where the 5/16" thick max. limit for plate thickness to minimize interference with the steel deck can be found?

It exists in print somewhere and isn't just a KootK experiential thing. I'm at a loss to point you to the source from which I relieved that guidance however. Joist roof system design and detailing is an awesomely rich subject. I've probably got half a dozen articles that are permutations of "427 Ways to Design Joist and Deck Roofs More Efficiently". The source may well be the document shown below which can be downloaded here in the 2019 edition if you don't have it already: Link. It's hands down the best single reference for this stuff in my opinion. Dedicate a month of slow paced reading to get through it and you'll find that you're the smartest joist & deck guy in most rooms. You know, as long as you don't physically visit any fabricators and spend time in their rooms.

 

[Triangled]

I've seen angles like you have, rods tucked up into the elbow of the top chord angle as mentioned above and also "knife plate", a vertical plate that nestles between the two top chords requiring a "down" weld.
I've seen joist manufacturers create seats in each of three ways: a) as you show with seat angle vertical legs "outside" of the top chord vertical legs, b) seat angle vertical legs "inside" of the top chord vertical legs and c) directly on top of each other. Which configuration seems unpredictable and therefore the advise I have received is on plans/detail, direct joist manufacturer to "Hold clear" for indicated connecting plate etc. When using "knife plate" rule of thumb around hear is to use plate thickness 1/8" less than gap between top chords to facilitate field fitting. I bet a call to a local joist supplier would get you assistance on the specification part (not the design part) to be sure it is field delivered as intended.