Pittsburgh Plate Glass mile-long factory

Turn-of-20th-century Pittsburgh facility: cast at one end, grind and polish moving straight (glass is brittle, can't be turned), exit as plate glass typically ½" or thicker. Now superseded by float glass, but PPG remains active.

So — continuous ribbon production. Plate glass they would roll out on these big plates. A company like Pittsburgh Plate Glass — you ever heard of PPG? — as I'm told, around the turn of the 20th century, they had a facility in Pittsburgh that was a mile-long building. Nice and straight: at one end they would cast the plate glass, and they would move it straight, no turning, because glass is brittle. They would grind it and polish it to where it was fairly smooth. That's what we call plate glass, and typically it might be a half an inch thick or even thicker. We call it plate glass because it's thick — if you roll it out and make it too thin, you have these great big sheets of it and you're going to break them. So it tends to be thicker.

At the turn of the 20th century, Pittsburgh Plate Glass had a building that supposedly was a mile long beside one of the rivers in Pittsburgh. They would have this glass melting furnace and pull the glass out, let it cool, put it through an annealing furnace because the glass will develop residual stresses and can fracture very easily until you anneal it. Then the greatest length was because they had to polish it because it wasn't very smooth. If you go look at old buildings from the 1880s and 1890s, you can tell which were the original panes of glass — the originals have got all kinds of waves on them. I can go out here at MIT in my office and you can look and see which panes of glass might have been original in 1916 or 1917 when these buildings were built, because they're not as flat.

The pre-Pilkington benchmark — glass plate produced on a mill that ran half a mile to a mile straight down for polishing.

So back in the old days, the bullseye glass, which was the center of this spun piece of glass, was the junk, and you would pay extra for the somewhat flat glass that went into windows like this. By the way, some of these at MIT from 1917 are original — I don't know how many are still original — you actually can see, by looking carefully, they're not completely flat. The way we made glass in 1900, companies like Pittsburgh Plate Glass — they would take the glass as a big sheet off a big mill, and they would have a mill that ran for half a mile to a mile just straight down, and they would polish the glass to make it flat. It was sort of wavy as the glass cooled, and you had to anneal it so it wouldn't shatter. Then someone — Pilkington in England, 1920s or so — came up with something called the float glass process, where you'd pull the glass off when it was hot, and you'd float it on molten tin, which is perfectly flat because of the earth's gravity. The top surface is also flat just because of surface tension, and you could eliminate all that half mile of polishing. Now you only have like a hundred yards of annealing on this float glass product. So newer glass is all float glass, and got rid of all the polishing.

Pre-Ford plate-glass production: cast onto a metal table, rolled and polished along a mile-long building. Used to set up the productivity contrast with the later float process (50-yard building).

So they poured it on a metal table, spreading it with rollers. This is actually how they made plate glass all the way up until the 1920s. Here's a metal table — they'd melt a vat of glass, pour it onto the metal table, and people with rollers would flatten it out, and then they would polish it to flatten it further, because it would have all kinds of distortion. The rollers wouldn't be perfect, they'd be different thicknesses. Pittsburgh Plate Glass, located in Pittsburgh — there was a building, a mile long, where they would cast the glass at a little facility like this in the 1800s, pouring on a metal table, and it would just move down these roller beds and be polished for a mile. Because it's so brittle, they didn't want to pick it up, particularly when it had these surface flaws. Once it gets polished at the end, then you could ship it. The factory was literally a mile long, just because you want to have the process flow that way.