For this test we failed to force a feared flaming failure mode, which is fantastic (though convoluted) news.

Many people appear to prefer melting the mixture to ensure good contact between the fuel and the oxidizer. Or, to be pedantic, melting the sugar (fuel) so that the small (unmelted) particles of oxidizer are suspended through out it.

But they warn that you have to be careful doing this (duh!) because the mixture might ignite. So we haven't been trying this since having a pot of rocket fuel ignite while you are stirring in does not sound fun. But we'd like to melt it. So we got to talking about how to test the concern and maybe quantify the risk. Specifically, we wondered:

• How easily will a batch ignite if you stop stirring while you are heating it?
• How big of a deal is it when it ignites?
• Can we think of a way to manage the process with acceptable risk?

As a first step, we decided to try to force the failure mode. We made up a batch (8g KNO₃ + 4g powdered sugar) and put it in a pan on the hot plate which we set in an open area and watched from ~3m back and monitored the temperature with the IR thermometer. We had to weight the pan down with rocks because the hot plate we are using has a safety cutoff that turns off the heat if there isn't enough weight on it.

We left it on low heat for ~5 minutes; nothing happened.

We turned the heat up to medium and waited another five minutes. Same result.

We turned it up to high; fairly quickly it began melting and turning brown around the edges (caramelizing). Soon it was quite runny (liquid) and dark brown (almost black). But it still did not ignite. So we decided to end the test (since we had provisionally shown that (with batches if this size, on this setting of this burner) we could get to a full melt without stirring at all without going boom).

The reason we stopped was this: what if we'd messed up somehow and the mixture was not correct. That would give us a false sense of confidence. So we removed the hot pad, put the mixture on our flame test shield, and lit directly.

Whoosh!

So that was good.

We now belive that we can reasonable safely do melts with the following rules:

1. Small batches (< 20g)
2. Only on the hot plate we tested with (set on high) & the pan we tested with
3. Stir constantly with a wooden stirrer (we have no reason to think that stirring with wood would raise the risk above not stirring at all).
4. Remove from the heat as soon as it melts
5. Only one person within 3m and they should wear eye googles or the face shield at all times.
6. Stir at arms length; no leaning over the pan