Using Normal and uniform distributed random numbers with boost library in class

I need to use normal and uniform distributed random numbers in my class. I want to set the seed in the main function and then give it over to the class, so that every random number is different during the whole process

But i do something totally wrong, because as soon as the public function calls the private function, there are the same random values:

class Dist{

    public:
            Dist(boost::random::mt19937& rng_, ...)
                 : rng(rng_){..}

            void operator()(const int timesteps){
                for(int i=0; i<timesteps; ++i){
                   onestep();
                }
                ...
            }
    private:
            double N(boost::random::mt19937& rng_){
                boost::normal_distribution<> nd(0.0, 1.0);
                boost::variate_generator<boost::mt19937&, 
                boost::normal_distribution<> > var_nor(rng_, nd);
                double d;
                return d= var_nor();
             }

            double unix(boost::random::mt19937& rng_){
                boost::random::uniform_real_distribution<> uni_x(-dw,0.0);
                double x;
                return x=uni_x(rng_);
            }

            void onestep(){
                for(int i=0; i<nr; ++i){
                 double Yx = unix(rng);
                 double Yy = uniy(rng);
                 Y_res.push_back({Yx,Yy});
            }
            ...
            ...
};

Now everytime i call the function onestep with the operator(), the vector Y_res is exactly the same. What do i wrong?