I was part of the staff for a journal that was published quarterly. Keep in mind this was more a humanities type of journal, but the idea is the same.
1) We receive a manuscript, quickly look it over (make sure it's not totally crazy), enter it into an online system then look for reviewers. We usually emailed 8-10 people per manuscript who study something similar, with the goal of having three people agree to review. That step can take anywhere from a couple of days to over a month, depending on how interested people are in the manuscript and how busy people are teaching classes.
2) Then we send out the manuscript and give reviewers a month to get back to us with comments. This is usually the slowest part. Most people will take at least the entire month. After a month there are usually people we still won't have heard from, so we start sending them emails to remind them. Sometimes the reviews will back out and we have to find someone else, sometimes they will quickly send their review and sometimes we just don't hear anything. Like I said, this is the slowest step.
3) When we finally have all the reviews in the editor will make a decision- accept it how it is; accept it with minor changes; ask you to use the reviewer comments and resubmit; reject it. Lots of people get asked to revise and resubmit. Very rarely did we get a first submit that was accepted. We would get hundreds of manuscripts per year and we only publish 5-6 per journal.
4) If you are invited to resubmit the process starts again (although we try to use the same reviewers). If you are accepted we give you something like a month to submit your final draft. Since we get so many manuscripts we often had the next few journals already set, so your article might not get publish for another 6-10 months.
Basically, don't wait on the edge of your seat, you'll be there for a while. Generally, anything under 2 months is insanely fast. We considered ourselves to have pretty quick turnaround and anything around 3-5 months was pretty standard. Again, the journal I worked for wasn't hard science based, but I imagine the process is pretty similar across the board.
Hope this helps.