Reduce thermal bridging in custom steel windows


We are installing a 15'x7' wall of custom steel windows in New York. Unfortunately, our steel fabricator, who is extremely nice but clearly out of his comfort zone with this project, already built out the frames and is not able to create a modern thermal break. Instead, he has used 1.5" steel tubing for the entire perimeter. We will be able to install large panes of double glazed glass, which should have excellent insulating properties. However, we are concerned that the amount of "unbroken" steel in the framing will lead to significant thermal bridging, condensation, and sweating. Since the metal framing has already been constructed and is a "sunk cost," we are wondering if there are any strategies to improve the thermal performance? (installing insulation in the tubing? Plastic shims between the glazing and frame? Exterior insulation and sheathing?)

Anyway, this is well beyond our area of expertise and so we would greatly appreciate any advice on how to make a less than ideal situation somewhat better.



Mar 4, 17 2:51 pm

What's your budget? (for these consults, not the materials)

Mar 4, 17 5:11 pm

The strategy would have been to work out the detail before fabricating.

Mar 4, 17 5:28 pm

Google "thermal break in steel frame windows" and then build something like that into the frame.

Mar 4, 17 5:44 pm

^^  David, there you go trying to use reason and logic.  Oh, well.

It's not clear how this window is detailed and finished out inside or out.  If the steel tube is exposed both inside and out, you are going to have a real problem particularly in a cold or humid climate.  Metal is a very efficient thermal conductor, which is working against you.  That's why you try to put a thermal break material, like rubber, between the inside part of the metal and outside.  Basically, two channels with a gasket in between.

All you can do now is try to protect the outside face of tube with some kind of inefficient material, and do the same on the inside.  Don't use wood because the dampness from condensation can result in mold issues.  Not fun.

Mar 4, 17 6:10 pm

He/she could split the tubes and then "glue" them back together with a thermal break. It's fun to watch if you make a video please.

Mar 4, 17 6:17 pm

or "float" the unit outboard a bit on anchor points and do the thermal break inboard.

Mar 4, 17 6:20 pm

Thanks David,

I wish we could split the tube and add a thermal break, but it looks like we are stuck with 1.5" steel tubes, so want see if we can come up with a solution to minimize condensation.

Right now, we are leaning toward adding insulation and sheathing on the outside, which should help somewhat with the heat transfer and condensation issues (see attached sketch). If we go that route, we still need to think through the details at the operable door and window. Pe

The only small silver living is that we are using very large panes of efficient double glazing (7' x 3'). Hopefully, that should improve the thermal performance of the overall units. That said, we still will be at risk of condensation forming on the interior frames. So, any suggestions on possible solutions that would still let us use the 1.5" tubing for most of the project would be very helpful. 

Thanks again!

Mar 4, 17 10:55 pm

It looks like in addition to the heat problem, there is a water infiltration problem. You'll need to think about sealant and flashing.

There is also a face of the glass touching the 1/4 material problem.

You should perhaps get a window person on board there.

I liked you in that bowling movie though.

Mar 5, 17 12:11 am

You'll need sealant where the "metal finish" meets the tab of the window frame. You'll need gaskets between the window frame and the insulated glass. The bottom tube will need the bottom of it removed so you can have weep holes at the bottom of the insulated glass unit. Those weep holes will need to shed on flashing. I made a sketch but it lacks a URL.

Mar 5, 17 12:51 am

I put the sketch in my "work updates".

Mar 5, 17 12:56 am

Thanks David!


This is amazing and extremely helpful! (Also, we love the retweet of Kellyanne Conway in Andrew Wyeth's Christina's World...Priceless).

Mar 5, 17 11:15 am

Your revised detail still has a thermal bridge problem (see image).

And I don't know how you're planning to secure the interior steel trim to the tube.  But (like David's sketch shows) they need to be removable, or if/when the glass breaks, it can't be replaced.

Mar 5, 17 7:13 pm

Let's try that image again.

Mar 5, 17 7:14 pm

I think most everyone here would suggest a thermal break. Perhaps one could start a crowd funding effort. :)

Mar 5, 17 7:23 pm

OP, why not get a diamond tip circular saw blade and cut the tubes then insert a thermal the risk of being redundant. I mean it's like $50 for the saw blade.

Mar 5, 17 7:51 pm

no wonder why architects can't make living...everybody bending over backwards to give out free work.

Mar 5, 17 10:16 pm

All the architects I know are making a living

Mar 5, 17 10:40 pm

You missed the point.

Mar 6, 17 9:24 am
Non Sequitur

Is the point that the sketches are in imperial units?

This is a forum, a guy wanders in and says hey look at this stupid thing we are doing, and then other people agree with him, that is what a forum does. 

Mar 6, 17 9:29 am
Non Sequitur

Last I checked, shop drawings reviewed and stamped by the consultants prior to fabrication likely would have fixed the issue... or is this another low-bidder win issue and too many people are bending over backwards to accomodate a contractor who could not be bothered to read the drawings/specs?


Has anyone suggesting that the tube be cut and a thermal break inserted actually done anything like this?  Every thermally broken window frame I've ever seen has the metal profile and plastic break locked together - basically a mortise and tenon joint.  I don't think there's some magic way to just do this on the fly.

chigurh...I assume you won't be making anymore posts to the forum, as asking a question = asking the rest of us for free services?

Mar 6, 17 9:43 am

you drew a free detail, maybe you are bored or something, but that kind thing devalues the profession.  

Mar 6, 17 9:44 am

A registrant is obligated to point out when a condition he observes is problematic. 

Mar 6, 17 9:57 am
Non Sequitur

I think you're stretching that asumption... it applies to professionals who notice and in the imediate presence of, ie. as observed on site... not from random wankers on online forums.

it makes me feel important :)


It is your duty and obligation to be self depreciating.

Mar 6, 17 10:03 am

Would need to tack weld receivers for the thermal break. Basically the OP needs to build an actual window rather than a steel frame that just looks like a window.

Mar 6, 17 10:05 am

David...agreed on real v. fake window.  But I'd be very hesitant to jerry rig this.  I could easily imagine the expanding/contracting steel tearing apart the thermal break plastic over time.  Maybe I'm naive, but I imagine window companies testing this very extensively to get it right.  The odds of my winging it and not having problems down the road.

OP...maybe the frame can be a nice dining table, and you could come up with a custom, minimal, wood frame?  It'll save you time/money/heartache in the long run.  Remember, when you find yourself in a hole, the first thing to do is stop digging.

chigurh...everyone on the forum has talked it over, and we all agree that you're to avoid being push overs, none of us will respond to any posts you may make in the future :)

Mar 6, 17 10:13 am

I had the exact same thought. It could make a nice table

Mar 6, 17 10:16 am

The price of that glass is going to be $$$$ compared to that steel frame. I would say someone needs to eat the cost and start over before you spend money twice on glass in two years. In addition to the thermal break the expansion and contraction of that steel frame may crack your glass if you don't accommodate the proper gaskets to frame. A 15' piece of steel is going to expand contract a fair amount I expect if exposed to the elements / indoors. Trying to make this working is going to end up costing someone a lot of money down the road.

Mar 6, 17 12:49 pm

opps just saw jcarchs comment. Agreed same concern.  15' of steel will expand and contract ~ 1/4" over a 100 degree temp shift !.  Glass won't.  

Mar 6, 17 12:51 pm

I've done a 100+ long window that was pinned under a steel tube, but the thermal break was accomplished at the finish connections. The glass was laminated, not insulated.

One can push on the wall and the whole thing sways, very dramatic.

Mar 6, 17 1:04 pm

sounds terrifying.

it's been there 16 years, no deaths

they have security guards to keep people from pushing on it :)


Boom! That's OP's answer. Hire a person for constant maintenance. a squggie, rag, bucket. Maybe a hairdryer. They should be able to keep that window free of moisture.

in my defense, the 100' window was under an interior partition (no wind load)

Wood Guy

Over a 100°F temperature swing, 3 meters of steel will move 0.08" (0.002m) and 3 meters of glass will move 0.06" (0.0015m), for a difference of about 1/32" or 0.5mm. There is plenty to worry about in the OP's situation but differential movement isn't one I'd be concerned with. 

Structural foam may help: It holds screws like wood but has an R-value of 3.1 to 3.8/in.

Mar 6, 17 1:23 pm

yes it seems like he could glue the half tubes together and call it an experiment


Doing the same thing here, so this isn't tested; We've just looked at the mockup, and mine are 9' wide counterweighted guillotine windows (single hung, but when they weigh 1200+ lbs, "guillotine" is a better description). 

What I detailed is basically at the glazing, I've got a neoprene break between the main frame and a custom glazing stop.  So at your sketch above, your square stop should be a flat tube (or just a plate) across the width then screwed through the neoprene into the main tube.  Mines operable, so the jamb/head/sill are two piece assemblies as well.   

Mar 6, 17 3:26 pm

A thermal break in the window frame isn't going to help you, even if you can figure out how to retroactively include one.

Instead, go to high-performace triple-glazing and improve the wall insulation around the window. The thermal bridging at the frame won't matter as much then.

Mar 6, 17 8:01 pm

that would imply there is money for the third layer of glass


You are incorrect assuming you haven't solved how to keep the steel warm. Warm moist air hitting a cold surface will condensate.


Yes it will. But there is no good way to thermally break steel window frames. They are what they are.


On the other hand, if you design a good thermal envelope and have your HVAC correctly positioned to move warm air across the windows, condensation problems will be greatly reduced.


As for the cost, triple glazed units don't actually have three layers of glass. The middle layer is usually a low-e film. You can also dramatically improve thermal performance of double glazing by having two low-e surfaces rather than one. If you can afford steel windows, you can probably afford triple glazing too.

cool, I've only used dual pane, low e so far, but I live in a warmish place

I like the 2x stuck inside the steel stud.
Mar 6, 17 9:19 pm

I've had similar problems and also see this non thermally broken tube steel window system in tons of NYC retail applications. I would try to gasket between the insulated glass and steel as much as possible. Also I haven't done this yet but definitely considered it - drill holes in the steel tube, fill with spray foam, weld and refinish. You would need to consult with spray foam applicator as foam in tight spaces and subject to condensation has to be considered. Also try to get duct work or perimeter heating onto the glass to reduce condensation which is caused by cold glass in the winter. And finally use a water resistant floor, if condensation is bad at least wood flooring won't get wet and warp.

Mar 7, 17 9:20 pm
Wood Guy

Good advice, but it drives me crazy that they would need to plan for a water-resistant floor due to expected condensation. What a failure of architecture to not use windows that won't condense. We have the technology, and have had for some time now...


No one mentioned the 2016 NYC energy code or Are these NFRC certified? I guess not because they are custom. What is the U-factor and SHGC? Did you use the prescriptive or performance approach for compliance with the NYCECC? 

Forget all of the above if your priject is exempt from energy code compliance. 

Jan 2, 18 11:42 am

Talking about thermal bridging...I love what SANAA did at their Zollverein project, simple massive concrete with "a few" pipes with running water from deep down below casted inside. No stupid foams or vapour bariers etc (as far as I remember) just concrete through and through, simply genius.

Not sure how that relates to this thread,  just wanted to share.

Jan 3, 18 3:20 am

has this been built yet? please post up the installed solution...architects are interested in how this turned out

Jan 3, 18 2:25 pm

Yes, pics or it didn't happen. And contractor license number...:)

Jan 3, 18 3:03 pm

I think the OP had an “oh shit” moment when I mentioned energy code compliance... 

Jan 3, 18 7:23 pm
Positive Pete

Are you suggesting that windoe is illegal? You can model it to be acceptable if the rest of building can level off the deficiencies

Are you suggesting that the OP was just sitting on the project for almost a year thinking everything was going to work out until you mentioned the energy code?


Yes- look at his sketch AND he is asking for free advice online... give me a break


The energy code issue is very real. But if you want to install non-NFRC windows you can still do it. You just have to switch to component performance approach and default the non-listed assemblies.

Sorry, I wasn't trying to point out that the energy code isn't important. I was trying to point out that the OP was asking for advice 10 months ago. I don't think the "oh shit" moment was because they read Bulgar's comment. They probably had that moment a day or two before they posted their problem here 10 months ago.


What I am suggesting is that if thermal bridging is only coming up now in the process, that perhaps energy compliance wasn't looked at well... not that the door itself is illegal. Of course it can be modeled... the question is if it has...

Jan 3, 18 9:24 pm

@proto, yes it's been built:

And here are all the pipes that run through the concrete, it's basically floor heating in the walls(!) with deep groundwater of a constant temperature that's being pumped through:

Jan 4, 18 1:23 am
Wood Guy

What a waste of resources. So the nearby mine has to be constantly pumped to keep from flooding, and they use that water to heat the building shell. I guess that's better than pumping the water into the river, as had been the case, but with a modicum of insulation they could use the same approach and district-heat the entire neighborhood, heat greenhouses, heat-exchange district hot water, or a hundred other practical uses. I really don't get how this building is considered genius-- they made a giant concrete radiator. Just because it's better than dumping hot water into a river doesn't make it good architecture. Someday the mine will be closed--what then?


Agreed. That's actually horrifying to me.


One does not exclude the other, still think it's genius.

Wood Guy

It certainly takes away most concerns about moisture accumulating in the wall assembly, and makes construction relatively simple. Hydronic radiant wall panels are nothing new, and radiant tubing cast in concrete floors is not new, but I suppose casting radiant tubing in concrete walls is a relatively novel approach. There's a fine line between genius and idiotic...


I know radiant wall panels or flooring is nothing new, but to have it inside a solid concrete wall that's also the facade that is just very smart imo. It makes great use of available resources and results in an elegant thin solution with a minimal material palette without the need for additional finishes etc. I would like to use this solution for my own house one day, simply pumping up ground water of a constant temperature as a temperature buffer year round, cooling in summer, heating in winter, so I can have simple low tech details with my concrete a la Sigurd Lewerentz' flower shop for example. Geothermal energy is a great carbon neutral energy solution, it's consistent and constant and also works when there's no sunshine or wind. It's the future of the past.

Wood Guy

Randomised, unless you live near a mile-deep well or have volcanoes near the surface as they do in Iceland, what you are thinking about for your house is a ground-source heat pump. People often call it geothermal energy but it's different technology.

Ground source heat pumps exchange heat with the earth via wells, but there is a lot of energy required for the electric pumps in an inexpensive "pump-and-dump" system, or you can use a sealed/pressurized system that requires less electricity to run the pumps but costs a lot more to install. A very common problem with them if they are not designed perfectly is that the ground cools off over the course of the winter and no longer works effectively. The ground eventually recovers, but heat moves slowly through the earth so it takes a while. Since efficient air-source heat pumps have hit the mainstream, ground-source heat pumps at a residential scale rarely make economic sense.

I could not care less what you decide to do at your own house, but as an energy nerd it seemed like a good chance to explain the difference. In case you're interested in more information: http://www.greenbuildingadviso...


No worries, I'm well versed on the subject. It depends on the depth you take the water from to have it at a constant temperature, year round.

^ I think the question was asking if the OP's project has been built.

Jan 4, 18 11:18 am

That's what I thought :)

go do it

excuse my simple mind.

but why can't the hollow steel tubing be sprayed with foam insulation?

Jan 4, 18 12:12 pm

look up thermal bridging and it should be clear. Assuming they dont float the entire assembly outboard, as David suggested above, the steel will​ still be a thermal bridge even if it has foam inside it or around the jamb

For posterity's future reference, here is how a steel window manufacturer reduces thermal bridging. They even have NFRC certification so the energy code people are happy.

Jan 4, 18 7:13 pm

But what's the U-value?

That depends on the glass you're going to use. NFRC is frame plus glass. The website notes they've tested to as low as 0.170 U-value (but it doesn't say what the IGU was). That's more than enough for prescriptive compliance in Climate Zone 7 in the 2015 IECC.

Block this user

Are you sure you want to block this user and hide all related comments throughout the site?

  • ×Search in: