Question: Solving numerically ode errors

I'm trying to solve the following

> de := proc (omega, l, M) options operator, arrow; diff(y(r), `$`(r, 2))+(2*r-2*M)*(diff(y(r), r))/(r*(r-2*M))+(omega^2*r^2/(r-2*M)^2-l*(l+1)/(r*(r-2*M)))*y(r) = 0 end proc;

Digits:=40;

> sol1 := dsolve([de(1/10, 1, 1), init, dinit], type = numeric, y(r), range = 10 .. 10000);

However I just get '[Length of output exceeds limit of 1000000]'

I want to solve this ode with high working precision and output 40dp (I will also be putting

options in when I get the above working)

Mathematica can do this fine, and I just wanted to compare Maple's speed. I don't understand why length of output is so large, surely I should just be able to pick off sol1(20) and get a complex 40dp number, which isn't that large.