Right, well you presumably have different data categories, like 40<x=<50 etc. Then you assume that every data point is in the middle of its category- for instance, every data point in the 40 to 50 category is 45. So you now have a frequency table.
Now, for standard deviation, you need to work out the sum of all squares. This is the sum of f(x - mean)^2, for every x. Then divide by the sum of f and square root.
As for range, I'm not entirely sure, but I'd imagine that you take the top and bottom midpoints and subtract the smaller from the larger.