Math, math, math. Much about the creation of an EE is finding the little math genie within each of them. It has been said that a Control EE skirts the edge of existence as a Mathematician. If so, one must have an affinity for the very slippery concepts presented to us as Control Theory and enjoy differential equations. That is very true should one want to resolve complex matrix operations that define a closed systems driving constants, say Kp, Ki, or Kd (proportional, integral, derivative). There must be an easier way for the remaining mortals amongst us.
Let me introduce you to my little friend, the Open-Loop Response...
The Open-Loop response describes how the Process Variable (PV) of your system to be controlled behaves to simple ON/OFF instructions. In most every system, the inputs are scalable and have quasi-predictable outcomes. I like to begin by unwrapping the item to be controlled and investigate how graduated changes to the input drives the process variable PV. Use your best judgment to guess at a starting point. Always wait until the system is at equilibrium before starting. This will help prevent measurement errors. Start, collect data, analyze, and iterate by adjusting the input. Many times just varying the input will achieve the desired response. If this is the case, a simple 2-State Control System, ON/OFF is sufficient and your Cost of Goods (COGs) has just plummeted.
Now please keep reading if you have already invested in an expensive PID controller. Thankfully those devices typically allow the PID constants to be disabled for calibrating each of the individual control inputs (Kp, Ki, Kp). So please take out that manual and find the disable setting (usually setting a constant to 0 does the job). Find the Open-Loop response and see if it fits your desired response for the process variable PV. If the application will work using a 2-state machine and/or only minor procedural adjustments are needed to realize the desired response for the process variable PV, then count your blessings. Less control is more when it comes to control theory practice.
ALL SKATE.
Some good references and links can be found at the following sites:
http://www.expertune.com/tutor.html
http://www.engin.umich.edu/group/ctm/PID/PID.html