Need help on composite scoring

I have spend a long time researching composite scoring and here is my situation:

- 10 different variables that I need to normalize (they are not completely independent so I suspect a lot of multicollinearity, they also have large swings in ranges of possible values)
- I believe I should get the z score for each of those variables and then scale the z score, I dont want to have negative numbers
- I feed the normalized values into the composite equation and that is my composite score.

Is this statistically okay to do?