don't you use what acer and Preben Alsholm already created for you?
dsol := dsolve(dsys union isc, numeric, method = gear[polyextr],
initstep = 0.15e-1, minstep = 10^(-11),
abserr = 0.1e-5, relerr = 0.1e-5);
you get the desired graphs. First, I don't understand why you would like to view the dependent variable in the interval [0..5*10^(-7)] and second: those values such as "0.15e-1" are not containing the base of the natural logarithm.
0.000001 and so on, while
exp(x) means the base of the natural logarithm to the power of x, see ?exp.
PS: You can also use some other stiff method such as ?dsolve/rosenbrock , which is a lot faster (at least on my computer).