I played a bit more with the model discussed today and made some updates to http://www.npl.uiuc.edu/~smclayto/mulan/slowDT.html Please read the comments at the top of the page, in particular the first about the method to find f_slow, and take a look at the revised plots. Steven