Finding the error on scaled parameters

Essentially I have been trying to fit some data to a model, you can follow my progress else where on MATLAB answers if you are interested. As I have decided to give up on doing the fit properly (as it seems impossible), I have resorted to scaling two of my parameters (this allows the regression to converge and I am fairly happy with the result).

Unfortunately the scaling means that I have no Idea how to export the errors on the parameters from the covariance matrix and I wondered if anyone could help.

the scaling I performed was log10(a)

wher a is one of my parameters, for the function i changed the too

 10.^a

I have tried to produce a simplified version of my problem but was unable to, sorry...

feel free to ask me questions as I realise my request is quite vague

Thanks