Concrete Balcony Thermal Breaks

OK no takers.

Attached is an image of the isokorb for a dropped balcony. Anyone have any thumbs up/down on how comfortable they'd be using this?

1) I wouldn't camber the balcony for exactly the reason that you specified.

2) I agree, that stepped detail looks suspicious to North American eyes. That said, we know that the looping business is one of the more effective ways to make an opening moment connection. I'd be looking for some combination of the following:

A) test results.
B) explanation of the DIN design methods that make it go.
C) assurance that AHJ is likely to approve it.


I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

Thermal breaks at balconies always end up getting VE'd out of my jobs. Lucky I guess...

I know engineers who have used Halfen's products. Though I'm not sure if they can handle that large of a step down.

I'm on board with the looping in concept, but not used to it enough to have any gut feeling. Based on no experience it seems like the loop is small... They spec a min. of 220 beam width and based on that the loop appears to be ~100. Seems there should be some min. dimension requirements for the loop to work out. I also wouldn't mind threading a long bar through there for extra sleep at night. This double loop action also makes me a bit nervous - the 360 loop to go up and then a 90 right away. I'm guessing that they've got this worked out as these are common outside of N America - but makes me a bit nervous.

The additional deflection that comes from the break is annoying. Rough numbers look like close to 1/2" on a 7ft cantilever. Granted this isn't live load deflection so maybe you finish that out and no one knows the difference. Seems like I'd get a few calls saying the balcony is falling off when they first see that initial drop.

I guess that people using these all the time have grown comfortable with it.

Like I mentioned above, I'd want to know how to design one of these loop +90 myself before I'd trust it. I'm okay with the design being per a different code so long as I understood how it worked.

What makes you say that the isokorb only contributes to dead load deflection? Or are you just saying that dead load deflection is all that you're concerned about?

Part of my skepticism regarding the deflection comes from the fact that I don't believe that all of the isokorb deflection would be additive to regular deflections. The rigid body balcony rotation due to the isokorb would take the place of the rigid body rotation that would otherwise have occurred due to strain in the concrete and steel that would have otherwise occupied the same space. Moreover, the isokorb won't be prone to shrinkage or significant creep.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

That is one hell of a detail... What provides the compressive force in the cantilever, is the insulation loadbearing? And the confining stresses around that loop would surely be pretty serious if there was a significant cantilever to the balcony? And if there was a large cantilever, I reckon it would be pretty bouncy? have they done vibration analysis on this? And how does shear transfer through that interface? ze clever germanz, witchcraft.

Add to that I don't really get why it's required - does it get so cold over there that it can penetrate through concrete and significantly impact the internal temperature? It's like 40 degrees C here right now and I don't have aircon

It seems to me to be a solution in search of a problem but then, I'm here to learn. Like BowlingDanish, I didn't realize this was an issue.

Yes this is a serious issue up here in the great white north.

Cantilever balconies are becoming harder to do without some form of thermal break. it can get to 25-30 below on the balcony, and a toasty 20 above inside, asking for frost and condensation issues.

Concrete is not a good insulator and steel is worse. Nothing to do with balconies, but this is an article regarding the use of an insulated concrete sandwich panel system that is reinforced with FRP in order to reduce thermal transmission.


- infrared showing higher heat transmission at the grouted cores

Maybe a silly question, but can someone explain the purpose of the loop?

My gut reaction is that the bar could develop straight into the building slab or, bend up until it hits the top bars then bend parallel until developed. I must be missing something.

This is a picture showing the bar being developed straight in.


I believe the loop is useful for confining the compression strut and shows up a lot in German reinforcing details.

DETstru said:

Maybe a silly question, but can someone explain the purpose of the loop?

Click to expand...

My understanding of the loop is that it is simply a means of developing the bar as it enters and exits the joint and effectively redirecting the tensile rebar force around the dog leg.

One of the things about the detailing that I find disconcerting is that I can't really make rational sense of it from a strut and tie perspective which is rare. As I see it, the primary flexural compression forces would have to cut across the loop as they navigate the dog leg path up to the upper slab. As such, the primary flexural compression stresses aren't really confined by the loop at all.

@Bookowski: the clip below is from one of my few European references. It shows a bit of how concrete rebar stresses are evaluated within rebar bends of a given radius. It's not directly applicable to your situation but you may find it interesting anyhow.



I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

I do love when the building's envelope does this to my structure. I understand the concerns, but at what cost.

Maybe drop your slab and add some insulation for 4' into the room beyond the exterior wall. As someone who does forensic work sometimes, I would not purposefully camber that balcony upward.

KootK -
Bad wording on my part. I didn't mean that there is only dead load deflection, but it's the largest component. Slab + superdead is ballpark 150psf and live is 60psf. Even initial dead + isokorb dead is a decent number, enough that I'm worried it would raise a few eyebrows when they pull the shoring.

I can't really say whether or not that additional deflection is conservative - it comes from a formula that Shock provides with some factors you pull out of a table, there's no real way to rationalize it. However, I'd think that a similar small portion of concrete would exhibit less rotation/strain. Either way I have to trust their number - and they specifically say it's additive and in their design checklist they have a box 'did you remember to add the additional deflection'.

They describe cambering the balcony as if it's run of the mill stuff - but as I stated originally I'm not ok with it. The problem is that without a camber the add'l isokorb deflection eats up most of your allowable.

Bowling & Archie - Yes, it's a problem and kind of a new thing in my area that is now coming up often due to changes in energy code and also condensation/litigation risks. An alternative that I've seen so far is insulating the entire balcony, i.e. extending the building envelope out/around the balcony. This makes a very fat balcony and generally doesn't make architects happy.

Bowling - Vibration... maybe sort of. All I can tell is that they have a table with maximum recommended cantilevers, beyond which they say vibration should be considered. Beyond that they don't elaborate.

KootK - I had the same thoughts about strut and tie. Seems that it's not so much s&t as taking the tensile force on a train ride around the loops up and over. My guess is that it's primarily based on testing. I spoke to one of their tech guys in Germany but didn't get that far - I'm going to get into it more.

So far I told the architect that the product only has a straight bar option so we could at best drop the slab a few inches max. This at least takes the loop di loop out of the mix - at least until he looks into it and finds this option.

KootK -
Bad wording on my part. I didn't mean that there is only dead load deflection, but it's the largest component. Slab + superdead is ballpark 150psf and live is 60psf. Even initial dead + isokorb dead is a decent number, enough that I'm worried it would raise a few eyebrows when they pull the shoring.

I can't really say whether or not that additional deflection is conservative - it comes from a formula that Shock provides with some factors you pull out of a table, there's no real way to rationalize it. However, I'd think that a similar small portion of concrete would exhibit less rotation/strain. Either way I have to trust their number - and they specifically say it's additive and in their design checklist they have a box 'did you remember to add the additional deflection'.

They describe cambering the balcony as if it's run of the mill stuff - but as I stated originally I'm not ok with it. The problem is that without a camber the add'l isokorb deflection eats up most of your allowable.

Bowling & Archie - Yes, it's a problem and kind of a new thing in my area that is now coming up often due to changes in energy code and also condensation/litigation risks. An alternative that I've seen so far is insulating the entire balcony, i.e. extending the building envelope out/around the balcony. This makes a very fat balcony and generally doesn't make architects happy.

Bowling - Vibration... maybe sort of. All I can tell is that they have a table with maximum recommended cantilevers, beyond which they say vibration should be considered. Beyond that they don't elaborate.

KootK - I had the same thoughts about strut and tie. Seems that it's not so much s&t as taking the tensile force on a train ride around the loops up and over. My guess is that it's primarily based on testing. I spoke to one of their tech guys in Germany but didn't get that far - I'm going to get into it more.

So far I told the architect that the product only has a straight bar option so we could at best drop the slab a few inches max. This at least takes the loop di loop out of the mix - at least until he looks into it and finds this option.

bookowski, give us an update if you hear anything new from the Schock people. I'm really having trouble understanding why the loop is advantageous over a slightly bent bar.

bookowski said:

...that additional deflection ...there's no real way to rationalize it.

Click to expand...

The quote below suggests to me that the deflection is estimated by calculating the angular change that is incurred across the width of the isokorb as a result of the accumulated tension strain in the rebar and compression strain in the thermal break over that same width. That angular change times the balcony length becomes the tip deflection.

bookowski said:

I can't really say whether or not that additional deflection is conservative

Click to expand...

Regardless of how it's calculated, the isokorb deflection is a product of the angular change that occurs over the width of the isokorb (call this deflection component A). When you ran your SAFE/RAM model, the deflections that you pulled from it were, in part, a result of the angular change that occurred across the block of concrete that will, in reality, be replaced by the isokorb (call this deflection component B).

Your total deflection should definitely include component A but, technically, it should not include A and B both as that's double dipping. Deflection should rightly be SAFE PREDICTION + A - B.

bookowski said:

I had the same thoughts about strut and tie.

Click to expand...

I had to reach pretty hard for it but I came up with a strut and tie model that I feel might explain how the looping voodoo helps out with joint moment capacity. Per the sketch below, I believe that the loop reinforcement essentially coaxes the compression field to round the corner following the same path as the lower left quadrant of the loop. This would leave you with an effective flexural depth around the corner that would be:

a) a little less than the flexural depth outside the joint but;
b) a good deal more than your stock diagonal cutting across the joint.





I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

@Bookowski: what do you typically enforce for balcony deflection limits anyhow? I haven't seen much guidance on that and I tend to be pretty liberal.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

I agree that D = SAFE + A - B, but who wants to deal with calculating B. I also have a feeling that A is much larger than B (A >> B). The baseline deflection is a result of the integration over the entire length. The support point definitely has an outsized affect since you get the full L in angle x L, but still the original number included everything. To keep it thinkable lets say the cant is attached to a rigid support, you can look at replacing that therm break length by a rigid material to determine the component of this area. Lets say original cant = 84" (7'-0") and the break is 4" long. New deflection is (80"/84")^4 x full deflection which works out to 82%, so that 4" gives you 18% of your deflection.

Take the isokorb CM30 and a 200 slab, from isokorb tan alpha = 0.9. Assume that service moment is 2/3rd of ultimate (1/1.5 ish). Then isokorb predicts D_add = 0.9 x 84" x (1.5/1.5) x 1/100 = .75". We don't have the original deflection in this exercise - but I think this is a lot more than 18% of that number. 0.75" over 84" is using 2xl for cant is already L/220. In the above calcs for the ratio of Mser/PhiMn I assumed that Mu = PhiMn, maybe you design for some extra there but you've still got a high number. Lets say you allowed a nice even 1" originally (L/160 ish), take out 18% and that's .18" which is 1/4 of the .75" from isokorb.... so seems like A > B by a decent amount, how much probably depends on the specific case - but enough that I don't want to calc B each time or count on it too much.

There's also the fact that isokorb asks you to use 4" beyond your support as your new l, but that is just another wrench in the gears so not worth playing with that in the example here.

That ST diagram is funny. Seems like the mechanism is that the hoop is like wrapping a cord around a gear, the concrete becomes a confined wheel which 'turns' the force around the corner.