For this test we tried for a best-of recipe with a few (possibly too many) additions.
We mixed 8g of KNO3 with 4g of powdered sugar. We then added enough glycerine to make a thin paste, and mixed well. We make a tube about the diameter of a rocket engine and poured the paste in. We let this "dry" for 5 days at 30% relative humidity and 22° C. As the paste settled (mostly the first few days) we added more from the supply (which was also kept in the drying room).
The idea with the tube is that we would light it from the top and the melty bits that skitter around would stay and contribute to the burning.
When we tested it it lit and burned well for a few seconds, then went out. Relighting it had the same effect. So...what did we do wrong?