Thanks: the battery and fuel spray patterns are good ideas. The battery is only 2 years old (NiCad), but I've had lead acid ones that have lasted much less time before now.
I'm no engineer, but the problem to me seems to be that fuel is going in as it should, and if it lights off as you crack it, the temp barely rises to the yellow line. So my supposition is that the fuel is still flowing in normally, but for whatever reason its not igniting - hence my thought that it's a problem with spark not fuel. If the start is not good, the temp rises to nearly 900 degrees - I assume fuel is still pouring in and when it lights, it's burning that extra second worth of fuel (c29ml??) too.
Not just one pilot: two of us fly the machine regularly and both see it.