A pneumatic cannon sits on a frame in a laboratory outside Albuquerque, barrel aimed at an ordinary rooftop solar module.
The only weather here is manufactured.
Pressure builds, and a ball of ice leaves the muzzle.
What follows is finished before anyone standing nearby could register it.
And what decides whether the glass lives is not the number on the panel’s spec sheet.
The dimple that decides whether glass lives
When an ice ball lands, the glass does not simply shatter. High-speed measurements at Sandia National Laboratories show that an impact first produces a localized dimple in the glass surface, with the peak depth of that dent scaling with the square root of the incoming energy. The pane then either springs back or begins a crack.
That makes energy, not stone diameter alone, the quantity the glass actually feels. A bigger stone arriving faster pushes deeper into the surface.
Below roughly 20 joules, the Albuquerque tests found negligible damage to the modules examined, which puts a rough floor under the whole question.
Where the ice lands matters as much as how big it is
Above that floor, position starts doing as much work as size. In the Sandia tests, both glass and backsheet modules and glass on glass modules cracked their cells and their glass at lower energies near the edges than at the center of the panel. The same projectile that a module shrugs off mid face can open a crack a foot away.
Build quality separated the two designs as well. The thinner glass used in glass on glass modules began cracking at around 25 joules, while glass and backsheet modules held to roughly 40 joules before failing, a gap that matters for anyone comparing panel brochures.
Why edges give way first is not yet settled. The Sandia team stated plainly that it planned to use its imaging data to determine whether self leveling flex of the module explains the discrepancy, which means the edge effect is a measured result with a proposed, not confirmed, cause.
What an 11,000 frame a second camera saw
The measurements come from a 2024 paper in the IEEE Journal of Photovoltaics, in which stereo high-speed video of laboratory hail tests was processed using digital image correlation to map module surface deformation during and just after impact. The rig is a pressure modulated pneumatic cannon with interchangeable barrels covering ice balls from about 1 inch to about 3.3 inches across.
Documented shots include a 2.6 inch ball at 77 miles an hour and a 3 inch ball at 86 miles an hour, the latter delivering nearly 142 joules. That is roughly seventy times the energy of the basic certification impact, and it is enough to take glass past the point of recovery.
The cameras ran at 11,000 frames a second, and even that was not fast enough. Part of the event immediately around the contact point was not resolvable at the chosen frame rate, meaning some of the fracture happened inside a window shorter than a ten thousandth of a second.
The 50 mile an hour floor and the tests stacked above it
The certification number installers quote is real, but it is only a minimum. Under the IEC 61215 hail test, a module need only take 11 strikes from 1 inch ice balls at about 51 miles an hour without visible damage and with less than 5 percent power loss, an impact the Renewable Energy Test Center puts at about two joules. Manufacturers routinely cite that pass as proof of durability.
The Department of Energy is explicit that this is not sufficient in severe hail regions and points buyers toward tougher voluntary programs. One of those subjects each test case to ten strikes at five locations using ice balls of 1.8 or 2 inches, sizes chosen to reflect the risk a panel faces across a 25 or 30 year life.
Panels in storm country can also be specified with thicker front glass and stiffer framing before they go into arrays, a small design decision with an outsized payoff once stones reach golf ball size. That tradeoff between upfront cost and long term resilience is one installers in hail belts weigh every season.
One Colorado storm and the hail still to come
Field evidence has been kinder than the lab. When a severe hailstorm hit the Denver area in May 2017, denting car roofs across the city, the Department of Energy laboratory in Golden came through with a single broken panel out of more than 3,000. The lab, renamed the National Laboratory of the Rockies in December 2025, reported that the one cracked cover appeared to have been struck by several stones in almost exactly the same spot.
That is the edge effect and the energy threshold working in a homeowner’s favor rather than against. It is the same durability logic that shapes how a tidal turbine rotor is built for months of reversing load rather than one calm test run. One engineer who has reviewed hail damage claims for insurers put it simply: “the stones that do the real damage are rarely the biggest ones, they’re the ones that hit the same spot twice.”
What remains open is how thinner, cheaper, record breaking large modules handle the same cannon as hail trends toward bigger stones in parts of the country. That will be settled the way this was: by firing more ice and watching, frame by frame, exactly where the glass gives first.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.
