Too small time step?
 

Hello,

I'm calculating heat transfer from hot air flow to a thin metal wire. Model is transient, and what I'm trying to find is a temperature response (step response) of mentioned wire.
In theory I should get something like this:

http://upload.wikimedia.org/wikipedi...oltage.svg.png

Model is defined quite precisely, but results are strongly depenent on time step size.
Air flow speed is constant almost everywhere and equal to 5m/s.
The smallest cell has size of 0.001mm. It is so small because metal wire is 0.01mm in diameter.
According to this data I set up the time step to 2e-7s, but results were completly ridiculous. I have prepared comparison of solutions with different time step sizes. It seems that solution which is the closest to experimental data is for time step much bigger than 2e-7s.

http://vernetest.vipserv.org/comparison.png

Could you tell me why it is so, and how should I find the proper time step?
Thank you for any of your help :)