We are a leading glass manufacturer based in China, specializing in high-quality glass solutions for industrial and architectural applications. With years of experience and ISO certification, we provide fast, tailored quotes and responsive support for procurement professionals, engineers, and project managers worldwide.
We are a leading glass manufacturer based in China, specializing in high-quality glass solutions for industrial and architectural applications. With years of experience and ISO certification, we provide fast, tailored quotes and responsive support for procurement professionals, engineers, and project managers worldwide.
I know that sentence irritates owners, architects, and sometimes sales teams, but I’d rather be honest at the start than polite at the change-order stage, because most “glass quality” disputes are not manufacturing mysteries at all; they are contract failures dressed up as aesthetic outrage.
And what is “bad glass,” exactly?
I’ve seen million-dollar façades get dragged into argument because nobody locked the viewing distance, nobody separated vision glass from spandrel acceptance, and everybody waited until the building was occupied to discover that tempered glass, low-E coatings, and giant lites behave differently under real daylight than they do in a sample box.
Here’s the hard truth.
If your specification says the glass must be “free of visible defects,” you don’t have a standard. You have a lawsuit draft, because FGIA’s own guidance says glass is generally inspected with 20/20 vision, in the vertical position, at the distance specified by the applicable standard, and that those distances vary by product type and blemish. ASTM C1036 covers quality requirements for flat glass used in architectural glazing, but it also says reflective distortion is not addressed, and ASTM’s own published paper says C1036 is not intended to be your full field-inspection rulebook.
So I write acceptance language before procurement.
When I’m dealing with custom solar-control coated glass specifications, I want the spec to say who inspects, from where, at what hour, against which mockup, and whether I’m judging color uniformity, coating haze, edge deletion, pinholes, or reflected distortion separately. Otherwise the project team ends up comparing memory against sunlight. That is not quality control. That is group therapy.
Most people name-drop ASTM and stop there.
That’s lazy. ASTM C1036 is your baseline for annealed monolithic flat soda-lime glass and related blemish/distortion categories; ASTM C1048 governs monolithic heat-strengthened and fully tempered flat glass produced on horizontal heat-treating systems, and it flatly warns that strain pattern can become visible under certain light conditions; ASTM E2190 is about insulating glass unit durability, fog resistance, dew point, and gas-retention qualification, not every appearance complaint an owner invents at substantial completion. E2190 also says it does not cover appearance and is not applicable to IGUs with spandrel glass coating because of test-method limits. On the safety side, 16 CFR 1201.4 sends covered architectural glazing products to ANSI Z97.1-2015 testing.
That distinction matters.
If you’re buying bulk heat-soaked tempered glass supply or jumbo tempered glass panels for façades, you are not just buying “stronger glass.” You are buying a product family that can show roller wave, anisotropy, and strain visibility under polarized or low-angle light, and the standard does not magically erase those optics because a sales sheet called the product premium. I’ve watched teams learn that too late.
The Seattle case should have scared more people.
In the Fifteen Twenty-One Second Avenue lawsuit, a 38-story Seattle condominium alleged that approximately 7,850 IGUs in its exterior curtain wall had to be replaced, with allegations tied to JS780 Gray sealant, UV exposure, and shattering risk; the court order notes claims that some panes shattered and that one shattering event occurred as high as the 35th floor. That is not a “minor visual punch list.” That is what happens when long-term performance, certification claims, and acceptance discipline drift apart.
San Francisco got the memo the hard way.
According to San Francisco DBI materials, broken glass was reported from seven high-rises after the March 2023 storms, and DBI said a WJE investigation found 30 breakages likely caused by issues that could have been identified and mitigated before the storms. By August 2024, AB-110 tied façade inspections to ASTM E2270 and required comprehensive inspections for buildings five or more stories, with supplemental façade inspections for buildings 15 stories or taller. Reports alternate every five years between detailed comprehensive and more general supplemental reviews. That is what a city does after it gets tired of pretending exterior glazing is self-policing.
And nickel sulfide still keeps showing up.
Facade Tectonics said in 2023 that spontaneous breakage in tempered glass from nickel-sulfide contamination remains a live issue, that heat soaking is the known strategy to reduce occurrence, and that there have been increasing occurrences involving significant quantities of glass in building façades. The chemistry is not exotic: NiS is small, nasty, and very expensive when ignored.
Not all blemishes deserve the same panic.
A faint strain pattern visible only under a specific sun angle is not the same thing as trapped debris inside an IGU, and neither is the same thing as edge damage that can turn handling abuse into later breakage. I also separate vision glass from spandrel glass because owners judge them differently, façades read them differently, and standards don’t always evaluate them the same way. If you’re specifying patterned or privacy products like fluted architectural privacy glass, the acceptance conversation has to account for texture, diffusion, and intended obscuration instead of pretending it should behave like low-iron float.
Here’s the field version I trust:
| Condition | What it looks like on site | What I call it | What should control acceptance |
|---|---|---|---|
| Surface scratch | Visible line or abrasion on the exposed lite | Visual defect | Project field protocol + applicable ASTM product spec |
| Roller wave / anisotropy | Distorted reflections, zebra-like bands, stress patterns | Tempering optical effect | Approved mockup + tempered glass acceptance language |
| Dirt or fingerprint inside IGU | Contamination trapped between lites | Fabrication defect | IGU fabrication QA + replacement criteria |
| Fogging / dew / gas loss | Haze, moisture, or persistent internal clouding | Performance failure | ASTM E2190-related durability requirements + warranty |
| Edge chip / shell / spall | Damage near the lite perimeter | Handling or fabrication risk | Edge-quality thresholds + safety review |
| Spandrel mismatch | Color or opacity variation behind backpan/ceramic/frit | Assembly/appearance issue | Separate spandrel mockup and approval standard |
The industry’s favorite dodge is pretending all six conditions live under one sentence in Division 08.
They do not. C1048 openly recognizes strain visibility under certain light, C1036 says reflective distortion is not addressed, and E2190 is a durability qualification standard rather than a universal beauty contest. That is why one mockup never covers everything unless the spec writer does the work.
Write it down.
I want installed glass, cleaned but not over-polished, viewed in the vertical position, under agreed daylight conditions, by named parties, from fixed distances, with lite IDs logged and photos taken from a repeatable angle. I want separate acceptance criteria for vision glass, spandrel glass, coated glass, patterned glass, and heat-treated jumbo lites. And I want the mockup to control unless the executed contract says otherwise. Why do so many teams still act surprised by that?
I also want substitutions locked down.
In the Seattle case, the dispute turned in part on allegations about sealant choice and certification. So no, I do not accept “functionally equivalent” coating, interlayer, spacer, or sealant language on façade glass unless the revised build-up is re-reviewed in writing by the design team, façade consultant, and owner rep. That sounds strict because it should.
And for tempered façade work, I stop pretending heat soak is magic.
NGA states that heat soak testing is meant to reduce the risk of spontaneous breakage from nickel-sulfide inclusions, not eliminate it, and notes that North America still does not have a consensus heat-soak standard, while EN 14179-1 is commonly referenced for all-lites heat soaking. So if the project needs heat-soaked tempered glass for large façade runs, say whether you want 100% heat soak, what documentation you expect, and whether post-temper coatings have been reviewed.
I’d keep it blunt.
I would list the governing product standards by glass type, define field viewing conditions, require a project-specific visual mockup, separate vision and spandrel acceptance, prohibit unapproved sealant or coating substitutions, and create a reject/monitor/accept matrix for scratches, distortion, anisotropy, internal debris, edge damage, and seal failure. I would also require maintenance logging for towers, because San Francisco’s 2024 program now explicitly asks whether a building has spandrel glass, whether maintenance records exist, and whether there has been glass breakage in the prior five years. That is not bureaucracy for its own sake. It is memory with legal value.
Architectural glass quality standards are the contract, code, and test-based rules that define how glass is manufactured, viewed, inspected, and accepted, including which blemishes, distortions, seal failures, and safety characteristics are allowed for a specific product type, project condition, and installation context. If the project does not define those conditions clearly, the argument moves from engineering to opinion fast.
Architectural glass should be inspected in its installed vertical position, by a person with naked-eye or corrected 20/20 vision, at the project’s defined viewing distance, under agreed lighting, after cleaning, and against the approved mockup or written acceptance criteria rather than personal taste. I would never let a close-up phone photo override a proper field protocol.
ASTM C1036 is a flat-glass quality specification for annealed monolithic glass, but it does not address reflective distortion, which means it cannot, by itself, settle every argument over roller wave, façade reflectivity, or tempering-related visual effects on a finished building. That omission is where many bad specs quietly fail.
Heat soaking is a risk-reduction process for fully tempered architectural glass that aims to trigger breakage from harmful nickel-sulfide inclusions before installation, but it does not eliminate the possibility of spontaneous breakage in service and it is still governed more by project specification than by a universal North American standard. I treat it as insurance, not immunity.
An insulating glass unit should be replaced when the defect is defined as rejectable in the contract or governing standard, or when the unit shows functional failure—such as persistent fogging, seal failure, loss of gas retention, unsafe edge damage, or breakage risk—rather than a merely subjective aesthetic annoyance. E2190 is useful here, but only if the spec writer understands where durability ends and appearance judgment begins.
If you want cleaner outcomes on your next project, stop asking for “high-quality glass” and start writing measurable acceptance language for solar-control coated glass, jumbo tempered façade panels, heat-soaked tempered glass, and fluted privacy glass. That’s where good projects stay good.
