One thing that I can see is that your delay routine is most likely not going to work. I don't know what compiler you use, but you should check the assembly it has generated. One thing is for sure: it will not come close to what it should be.

As a general rule of thumb, write timing routines based on wasting processor time, directly in assembly. That's the only way you can control what is going to happen. Another way to do it is to use a timer, but obviously you need to have one spare or have one free running at a suitable interval.