ELOG - heating/cooling data fitting

I fit the data that Kevin and I gathered a while ago for the heating/cooling test we performed. For cooling, I used curve_fit with an initial value of p0=(1,1e-6,1) (just copied these values from an online tutorial) to get the graph shown below. I also attached the whole program as well.

array([[  3.81451668e-04,   9.42870109e-09,   5.48247926e-06],
       [  9.42870109e-09,   1.17596815e-12,   9.28431073e-09],
       [  5.48247926e-06,   9.28431073e-09,   1.05408165e-04]])

The heating data has a break in the middle which I am not sure how to deal with, so I just took the data before the break. The program is similar to the cooling one. (used p0=(-1,1e-6,1) instead)

popt:

array([ -8.35340372e+00,   2.01869270e-04,   2.83287046e+01])

pcov:

array([[  9.83759992e-04,   5.09067460e-08,  -1.10929039e-03],
       [  5.09067460e-08,   3.47323122e-12,  -6.77130920e-08],
       [ -1.10929039e-03,  -6.77130920e-08,   1.38998816e-03]])