Here are my solutions:
Note that there may be slight differences due to rounding, or how the frequency distribution was set up.
Regarding that solutions that you worked out, here are some comments.
In your work for Question 1b, it seems like you haven't accounted for all of the data. There are 40 values in the data set, but you have only tallied 28 of them.
In Question 2 (page 2), there is an error in your formula for the standard deviation. In the second term under the radical sign, the "n" in the denominator should not be squared. The term should be (∑fx)^2 / n.
In Question 3, a class width of 2.4 ends up giving you 7 classes instead of the 6 that were calculated using the formula for k. In my solution, I adjusted the class width to 2.5 so that 6 classes will cover the range of data. I did this just in case your instructor is being picky about the number of classes. In the real world, your class width of 2.4, and having 7 classes instead of 6, would be fine. The difference in class width will slightly affect the calculations for the median and the quartiles.
Please feel free to ask if you have questions about any of these solutions.