Fun with Sulfur
Gun powder is made from sulfur, saltpeter (potassium nitrate), and charcoal (carbon). Where you go from there is wide open.
I don't know if the truth about why I wanted to buy sulfur would have worked at the pharmacy in the late 1970s. The range of things people think it's reasonable for kids to be playing with has shrunk quite a bit since then. (Or changed anyway: Making pretty fireworks is out, video games where kids practice point-blank killing are in. See my hotheaded rant on the topic.) In any case, I never had trouble buying either sulfur or saltpeter separately, and carbon came from plentiful charcoal.
Making gunpowder is fairly easy, though it's much more difficult if you want it to go bang rather than just burn fast and furiously.
It's just a mixture of the three ingredients in a certain proportion (which you can look up in any encyclopedia or on the web). Beyond that, it all depends on how finely ground and intimately intermixed the powder is. To make it explode fast enough to work as gunpowder, the chemicals have to be intermixed down to a nearly molecular level, so that when the ignition wave passes by they can react immediately and contribute to the chain reaction.
The dangerous part of making gunpowder is achieving this intermixing, which requires grinding the components together with each other. Sure, you can grind each component (which can't explode in isolation) individually into a very fine powder. But that's not good enough: In the end you must grind them together in a way that smashes and mixes the grains with each other. When you're doing this, sooner or later it's going to blow.
The old DuPont gunpowder works on the Delaware river had a clever solution for this. The grinding mills were built with three walls of think stone, and one, facing the river, of cheap pine boards. Nobody stood between the mill and the river. This is a beautiful demonstration that you should never constrain an explosive substance, unless you really intend to send things flying. By leaving one wall essentially open, they virtually eliminated the danger of the building exploding in all directions. They are part of a museum and gardens now, having survived many years of gunpowder making.
Now, the DuPonts had rolling mills with carved stone tracks, but I used the side of a cinderblock and a metal cold chisel (which had a nice flat surface). This was probably a bad idea: Metal tends to spark on concrete, and sparks are bad. My concession to safety was to only grind a very small amount, maybe a tablespoon at any one time, and keep the rest of it well away. I also had the good fortune of being nearsighted at a young age, so I was wearing glasses at all times, which is the most important thing to be doing any time you're working with reactive chemicals.
And since my gunpowder was very low-intensity, even if it had ever been set off by grinding (which it never was), it probably would not have actually caused much injury.
At first I was quite disappointed that I could not make bang-type gunpowder. It was probably a combination if inadequate grinding, and the fact that charcoal really isn't anything like pure carbon. But I came to realize that my stuff was much more versatile: It was more like pyrotechnic fuel than gunpowder.
I would mix in things like copper filings, then wrap it up in a cone shape, using several sheets of paper for the cone. Lit from the top it would make a several-foot high fountain of colored sparks: Every once in a while one would make a quite respectable display, in my humble opinion.
All in all, I think this was not nearly as dangerous as it might sound, though certainly if I had ever accidentally made really good gunpowder, bad things could have happened. We won't even get into the flash powder rockets: That's a whole other story.