When first testing the board I had a problem with the boostconverter. The problem was when the uC was when switching the boostconverter on, that it drew too much current resulted in a big voltage drop. The voltage drop was big enough to reset the uC. When the voltage recovered enough to turn on the uC, did it switched off the boostconverter before it realized the temperature was too high, which made the whole process to repeat itself.

This test was done with a external power source with a long leads, this was maybe some of the problem, but with a long USB cable will probably do the same. To solve the problem did I changed the C2 capacitor to a 220uF which solved the problem.