The way a dyno works is the engine is trying to pump water through an orifice. The size of the orifice is controlled by the computer. It can neck down to a small size, so that when the engine is trying to pump water through that orifice, the load on the engine increases, and eventually the engine can't accelerate. That's why the dyno is called a water brake; the force the engine exerts trying to pump the water is resisted by the orifice, so that the engine can't increase in speed. For your dyno results, the minimum speed, that the engine was braked to at the start of the pull, was 3000 RPM. So, the computer reduced the size of the orifice to the point where the engine could only get to 3000 RPM, despite being at full throttle.
When you go lower in engine speed, obviously the engine will pump less water. The orifice can only neck down to some minimum size, so if you are running at a low engine speed, say 1500 RPM, despite the size of the orifice the engine can still pump all its water through the orifice, and so it will accelerate past that 1500 RPM point. Therefore, not enough engine braking is available from the dyno to hold the engine at 1500 RPM.
The engine needs to be braked in order to get valid torque data, so in the preceding example there is no way that you can get data from the engine at 1500 RPM. I've run quite a few engines down as low as 2500 RPM, and one down to 2200 RPM, and lower is probably possible, but it depends on the engine and the dyno. Also, there is a bit of a paradox there, because if you build an engine with a whole bunch of low end torque, the dyno will be less likely to be able to brake it at the lower engine speeds. If you build an engine with less low end torque and more top end horsepower, it may be able to be braked to a lower speed, but who cares, since the engine is designed to make power at the top end.
Brent, I have no idea why running an engine at a lower speed on the dyno would cause it to burn pistons. Can you explain your rationale with that comment?