I have worked on many projects that have failed at one point or another during the system development life cycle.  One in particular was a project I was working on where we had a team that consisted of hardware and software developers.  The project had just been funded and the overall objective was to provide the customer with an all-in-one system to test field devices.  The project had hiccups at almost every level in the life cycle. However, ultimately it failed because the customer never provided clear requirements and the project manager didn't think to ask.  The design and development stages went well enough.  There were a few minor set backs when vendors couldn't provide the material they quoted us in the time frame they promised.  There were several issues where large batches of components failed due to management selecting vendors that were cheaper rather then better quality.  However, where the project failed miserably was during the operation stage. We had packaged 250 systems for the first installment to ship to the customer when the customer wanted to do a hands on inspection.  They came in, plugged the systems in and attempted to run the system.  The systems would not run.  At no point did management or the developers think to test the system as a whole.  We tested the hardware which worked fine and we tested the software which worked fine but when we loaded the software on to the hardware it didn't work at all.  We had to stop all production and create test procedures that tested each piece individually and then a test procedure to test the system as a whole.  This took us two months past our expected delivery date and cost the project thousands of dollars.  

I think the project would have been more successful had the customer been very specific about their requirements and the project manager taken the time to evaluate the requirements and what each requirement would entail his/her team to do.  I also think the project would have been more successful had the project manager picked the type of systems development life cycle approach he/she wanted to take.  Management oscillated between several of the approaches which made the different phases hard to determine.