Window performance with frames

19 posts / 0 new
Last post

Does anyone know of a window manufacturer web site that has literature
on its windows including the frame? I am using Trace 700 for EA Credit 1
and the reviewer asked for manufacturer data sheets for window
performance values (U-value, VLT and SHGC) including the frame.

Your help is appreciated,


Otto Schwieterman's picture
Joined: 2011-09-30
Reputation: 200

You can use LBNL Windows 5 or 6 to obtain this information. Just go to
their website and download the program, which is free. You then enter
you window parameters, area and apply a frame. The program will
calculate the whole window properties for you.

Rimes, Christie's picture
Joined: 2011-09-30
Reputation: 200

NFRC (National Fenestration Rating Council) on its web site has publicly
available certified product directory (CPD) for residential windows with
around 200,000 products. To access CPD go to Each product's
performance includes framing. You can search by a manufacturer, product
type, etc. There is also new CPD for commercial products, which can be
accessed from


D. Charlie Curcija2's picture
Joined: 2011-10-01
Reputation: 0

Hi Otto,

Serious Materials provides full frame performance data here: I have also have some of our commercial packages modeled in curtainwall and storefront framing systems that I can provide upon request.

Firstly let me say that Window5 is a great tool - I use it often and love it. The only thing I would advise is that there can be a wide distribution in frame U-factors. The default U-factor for a thermally broken aluminum (TBA) frame in Window 5 is U-1.0. Using THERM and Window5 I've seen TBA frame U-factors vary from as low as 0.8 to as high as 1.5. The resultant full-frame U-factor using these 2 differing TBA frames (and the same glass package) is: 0.29 (R-3.4) for the poorer performing frame, and 0.22 (R-4.5) for the better frame.

I would say that putting either of those full frame u-factors in your energy model will yield pretty different results.

I think it's really important that we, as simulators and modelers, look at full-frame performance because this is the system that will be installed in the building. Center-of-Glass data can be useful comparing two different glass packages, but at the end of the day, it's a framed system that will be installed and a framed system that should be analyzed. My 2?.

Alex Krickx

John Dossmith's picture
Joined: 2011-09-30
Reputation: 0

Hi Otto

You can download software from the web called window 5 or 6. I think
that can solve your problem. It got a great database of all the
different manufacturers and their products and you can create your own
window with a frame to get the values like u value etc to input into any
software. I used that software to input values into IES-VE and it worked


Manjinder Singh

SINGH Manjinder's picture
Joined: 2011-10-02
Reputation: 0

A quick tool that includes many glazing combinations with fairly good
estimates of frame performance based on break thickness. Includes
different spacer types, etc.

Chris Jones

Chris Jones's picture
Joined: 2011-09-30
Reputation: 0

Hey all:

We do not know Alex at all but completely support his position that it is
the entire assembly/full frame in use that counts. We are a building
envelope consultancy and deal principally in the built market and see this
issue all of the time where the performance of window, storefront and
curtain wall units is dramatically different than that ?calculated? or
provided. As a matter of fact, it is the unfortunate norm. At the end of
the day it is about actual in use performance and that performance versus
what was calculated. I will stay away from installation difference in this
one J.

Anyway, we do agree with and have seen the truth to his assessment of frame
differences. It is only the performance of the entire assembly, at every
level of a structure as a whole, that counts in the real world.

Thanks and take care,

Andy Hoover

Andy hoover's picture
Joined: 2011-10-02
Reputation: 0

Good morning Everyone,

I agree with both Andy and Alex that the installed assembly is what must be modeled because the assembly values can differ significantly from the center of glass values. However, more often than not, the fenestration is designed using glazing from one manufacturer which is set in frames by a different manufacturer. This creates difficulties in determining the assembly value, and modelers have a very hard time accessing testing data for a particular combination of glazing and frame. Windows5 is often the only tool a modeler has to approximate the assembly values needed for the models.

When using Windows5, we never use the default frame values. We go to the frame manufacturer to get the frame parameters including the U-value and create a custom frame for that project in Windows5. The same goes for the glazing assembly when the data is not already in the Windows5 database. By using the actual manufacturers' frame and glazing data, Windows5 can get close to the installed values.

Keep in mind that energy models can never exactly match reality. And, there comes a point in energy modeling when we have to say "close enough". In any model, that point is reached when the amount of time needed to get a fractional improvement in accuracy becomes burdensome with very little gained in accuracy. Using Windows5 as outlined above can help to minimize the margin of error with a reasonable amount of effort.

For EAc1 submissions, you can submit the glazing data sheet which shows the center of glass values, the data sheet (or other form of response) from the frame manufacturer, and a copy of the Windows5 report (or a snapshot of the inputs/outputs screen) showing the entire assembly calculated values. The reviewer will accept this approach.

I hope this clarifies things from the energy modeling and LEED perspectives. Have a nice day.

Sheila Sagerer

Energy Engineer, EIT, LEED AP

Sheila's picture
Joined: 2011-09-30
Reputation: 0

Important thing to remember with windows is that the real number is the one
that account for actual size and that takes into account angle dependent
properties of glazing. In most cases, single performance indices at
standard conditions are used (i.e., U-factor, SHGC, and VT at normal
incidence and standard environmental conditions) in an energy model of a
building, which can produce errors anywhere from 10% to a 100%!! The
correct procedure is to either specify correct configuration and size of the
product, with glazing specified layer by layer (possible in EnergyPlus) or
use what is called DOE2 or EnergyPlus reports from WINDOW program (DOE2
report is used in DOE2 simulation program and E+ report is used in E+
simulation program). This is the only way to fully account for changes in
size, angular properties of glazing and actual weather conditions.

NFRC (National Fenestration Rating Council) has started to provide DOE2 and
E+ report files in their new CMAST certification system, however number of
products available is still relatively small and is limited to commercial
projects. EnergyPlus has also implemented in version 5.0 option to specify
standard numbers at standard conditions (like the ones available from and E+ uses then internal
algorithms to provide glazing angular information and correction for actual
conditions. While it is best to model specific configurations for a window
in THERM and WINDOW program (THERM is required to model frames and edge of
glazing), this is not something that average user will be able to do, so
getting whole product information from the NFRC web site and using E+ option
to convert this data to something resembling real performance is still the
best and easiest way to get actual information. For commercial products provides full information that does not need any
simplifications or corrections.

Our energy simulation tools DesignBuilder and EFEN use EnergyPlus simulation
engine, with the full range of options mentioned above and also provide
hundreds of predefined fenestration products that can be matched to your
actual product.

D. Charlie Curcija

D. Charlie Curcija2's picture
Joined: 2011-10-01
Reputation: 0

Mr. Curcija and everyone:

If I may be so bold as to ask some "dumb" questions,

- I'm an individual who regularly relies on the DOE2 Glass library (a large excel file attached to any eQuest installation) to find window constructions whose NFRC framed U-value, visible transmittance, and SHGC/SC values best match those prescriptive values given for actual design models and for 90.1 baseline models. Am I to understand all of my past models have a "10% to 100% error" with regard to the modeled behavior of the glazing? Please explain!

- It sounds like the use of Window5 is pretty popular, and this ability to export a file which can be used by DOE2 has me intrigued - has NFRC or any other body decided to make representative report files to match 90.1's baseline envelope requirements to encourage this procedure for comparative studies?

- Is Window5 the same thing as WINDOW?

- Furthering Otto's original question, my reaction to a reviewer requiring "VLT including the frame" would be to first roll my eyes and then explain the frame has a VLT of zero, and is modeled distinctly from the glazing which has a VLT of ##%... am I misunderstanding the request there?




Nick-Caton's picture
Joined: 2011-09-30
Reputation: 805

Hi Nick,

I think I know the answer to some of your questions, for some others I'm happy to share my thoughts, and for the remainder I have no clue.

I have always assumed that for the DOE-2 Glass Library that those files have Window5 report files behind them - i.e. they're not just single values for U, SHGC and VT. I could be mistaken though.

Skipping number 2...

I believe that Window5 is the same thing as WINDOW (technically, I think it is called WINDOW5 and there is/soon will be WINDOW6).

Lastly - I've looked into some detail at how full frame U-factors are calculated and how they vary by size. So as a brief tangent, I thought I would look at whether SHGC varies by size too. I found that if a window had, say 0.5 SHGC COG and the window was 90% glass, 10% frame (by area) the resulting full frame SHGC was GREATER than (90%)*(0.5 SHGC). This means to me that the frame has some non-zero SHGC contribution. Granted, it was small (if I remember correctly on the order of 0.1) but still - non-zero!!! In my head I assumed that the frame would reflect additional heat through the window into the space - that may not be correct, but it made sense with my results. I haven't looked at VT that closely but I would assume the same is true - that the frame adds some small VT to the whole unit.

Alex Krickx

John Dossmith's picture
Joined: 2011-09-30
Reputation: 0

To answer to program name question, WINDOW is the generic name that is
sometimes used for the LBNL software used to model the optical and thermal
performance of windows and other fenestration products. There are currently
2 versions available:

- WINDOW 5.2 which is used by NFRC for their fenestration rating

- WINDOW 6.2 which is the next version of WINDOW, with all the
modeling capabilities of WINDOW 5.2, but with the major enhancement being
the ability to model ?complex glazing systems?, which includes venetian
blinds, uniformly woven material (such as bug screens), and frits.

Both versions generate data files for windows that can be used by DOE2 and
EnergyPlus, but only for specular glazing systems, not for the complex
glazing systems.

These programs can be downloaded at no cost (paid for by the US Dept of
Energy) on our website:

Robin Mitchell

Robin Mitchell's picture
Joined: 2011-10-01
Reputation: 0


I haven?t used eQuest in a long time, so I am not sure how is that
windows/glazing list incorporated into the program. If the list shows NFRC
U, SHGC and VT and internally eQuest uses DOE2 report file, then the answer
to your question is no, there is no 10-100% error. If eQuest just applies
those fixed values without using DOE2 report, then yes, large errors are
associated with window treatment.

NFRC lists specific products for which rating is available. They are not in
business of providing baselines. I guess that is more appropriate for
software developers. That is the reason that in our software tool we have
extensive list of products matching ASHRAE Fundamentals list of typical
windows, which would easily enable user to select 90.1 compliant window.

Window5 is WINDOW 5.2, so yes that is the same program. Please note that
there is already WINDOW 6.2, which now models shading devices and other
diffuse layers, in addition to specular glass.


D. Charlie Curcija2's picture
Joined: 2011-10-01
Reputation: 0


Frame VLT is 0, unless it is made of transparent material of some sort.
Frame SHGC is NOT 0, because portion of the heat is absorbed by the frame
and re-directed to inside. The more conducting the frame the higher the
frame SHGC and vice versa


D. Charlie Curcija2's picture
Joined: 2011-10-01
Reputation: 0

We have been debating this as well.

As I understand the ASRHAE code, there is no spectral performance specified for the baseline glass.

This leads modelers to use the "simplified" specification for the baseline in an effort to match the ASHRAE numbers exactly.

I second Nick's request for the baseline spectral performances to be published by a National body.

Aaron Dahlstrom , PE, LEED? AP

Dahlstrom, Aaron2's picture
Joined: 2011-09-30
Reputation: 4

Dear All,

Are we talking about the assessment for simple regulatory approval or are we
talking about what a window really is. We have found dramatic variations
between NFRC values and what actually goes into a building. I will not go
into differences in measured infiltration rates as that is too random but
the window framing factor is fairly parametric.

The NFRC tables work on a sample window size (lest say 900mm x 1200mm). This
is quite a small window when compared to most installations but;

- For windows smaller than the sample size the frame to galls
factor is much higher,

- Conversely for large format windows commonly used in residential
& commercial projects the framing factor is much less. In some test cases we
have calculated framing factors changing from 18% to 8% or even less.

Depending on Glass system and framing system this can create some
interesting changes to thermal simulation results. One of my colleagues will
hopefully publish something on this late 2010 or early 2011.

Mark Dewsbury

mark dewsbury's picture
Joined: 2011-10-02
Reputation: 0

Hi Mark,

I agree with the point you make, but I want to expand the conversation a bit. Specifically, I aim to discuss the difference between how to use NFRC values and how to input values into energy modeling software.

The NFRC value should be accurate for a window of the NFRC standard size. So a standard sized 1200mm x 1500mm fixed window will have a certain U-factor, SHGC and VT as part of its NFRC rating. But a different size will certainly have different performance values. I think NFRC is very useful in comparing Window A to Window B in a standard size, but you are right: once you've selected Window A we need to make sure that we have the right performance parameters for that sized window.

I've been working on some projects that are aiming to be Passive House certified in which case they DO need to know how performance varies with size. We've worked out some formulas to calculate the full frame performance (only for U-factor at this point) on a range of our products. It's tricky because you need to know the U-factor of each component (edge of glass against each part of the frame, each part of the frame, and COG) as well as the areas of each of those components which of course vary with different sized windows.

I'm not sure what other manufacturers are doing to be able to provide this, but we have some of our product lines in a format that we can provide an NFRC-grade value for a non-NFRC sized window.

If you want to see the calculations, you can follow this link: and click "Download SeriousWindows for Passive House" at the bottom of the page or click here:

The document describes how the window is broken into different sections and includes all the necessary info needed to calculate full frame performance.

The last thing I'd like to add: It is probably simpler to upload WINDOW5 data (I know eQUEST can import COG spectral data), specify your window size in the modeling software and specify your frame U-factor. Your software should take all that data and calculate what the full-frame data is pretty accurately.

If anyone has any questions about this, feel free to contact me: generating these formulas took quite a bit of time and I'd be happy to share my learning.

Alex Krickx

John Dossmith's picture
Joined: 2011-09-30
Reputation: 0

Dear Alex,

I work in the HER validation field , so we need to know that we are
modelling the right fabric matrix (framing factor and windows included). The
software available in Australia only has the NFRC type table for selection,
which is unsuitable for us. The get around the problem we can modify the
data in the software?s scratch file. This is a file the software creates
post data entry before simulation. We open this file and modify lots of
parameters based on ?as-built? observations.

Once you start playing at this level, you actually create as many windows as
you wish. You just need & checking with the data entry. We mainly modify the
framing ratios of a suitable NFRC window to match what is really in the
wall. It is laborious, so I do believe that we need a parametric model
within the HER software which can ultimately make the changes for us. The
only other addition would then be a mullions tool to nominate how many
vertical and horizontal mullions.

We are to discuss this in Australia when I finish my current windowless
project. But I have been advising another researcher with a windowed

Mark Dewsbury

mark dewsbury's picture
Joined: 2011-10-02
Reputation: 0


Just a couple of comments. I became frustrated with the glass library data
years ago when I could never find a window that met the Oregon Code values
or the ASHRAE values. So I use Window vs x. It's quick and easy and you can
input data for all your sizes thus you have real output. Size makes a big
difference! I have neglected the frames up until lately because it's been
harder to get data related to them. When you get into it you find that
things like thermal breaks vary wildly in their structures. Who knew? I
thought a thermal break was a thermal break, but no, they can be made out of
various materials and sometimes don't really "break" the heat path at all.
Manufacturers like to use terms like "thermally improved". Just try to find
out what that means. Anyway, since so many windows are custom built in the
field it's good to get as much glass and frame info as possible so you can
accurately model them. Call it one of the final frontiers: a good thermal
break can really save energy. I always suggest it as an opportunity because
I know most developers will not consider them otherwise. They can cost a lot
although how much is debatable.

Last comment, and I apologize in advance Robin, Window is not paid for by
the Department of Energy; nothing is. It's paid for by us using our tax
dollars. At least they are going for something useful wrt Window.


cmg750's picture
Joined: 2010-10-05
Reputation: 0