|
Re: manual control of a victor?
the thing is, a 555 timer may be cheaper than an 8 pin PIC chip, and if you need to rig something up quickly then who cares about anything else?
But if you are designing a product, that will be produced by the thousands or millions, then you must consider the entire engineering and life product cycle.
One shot timers are the simplist form of asynchronous logic - but if you open the door to these, then you allow asychronous logic into the designers toolbox, and then where do you draw the line?
Looking at SW again, if a subroutine has one entry point, and one exit point, its easy to understand the flow of your code - but if you have goto statements allowing your code to jump all over the place, you quickly end up with spagetti algorithms
same with digital HW. With one system clock the only parameters you have to worry about are setup time and hold time - what happens between the rising edges of the system clock become irrelavant, as long as everything settles down to meet setup and hold.
But if you allow one shots, or async clears, or you gate your clocks, then your timing becomes complex, and simple things like changing a chip from TLL to CMOS, or even from one supplier to another will make your system stop working, or worse yet, will make it unstable, so it fails under certain temp or application conditions.
There are golden rules to designing digital logic:
No asynchronous logic (including one shots - everything must be registered by a clock)
no circuits that can glitch
no gating of clocks (one master clock for the system)
asynchronous inputs must be double registered (to prevent metastability)
if you take a short cut and break one of the golden rules you will pay for it sooner or later: when you try to make a simple modification to the system, when a part goes end of life and must be replaced, when the system is used outside its normal environment, when you need to upgrade, when you need to interface to something new or different, or when someone else has to work on something you designed.
that 60 year old Professor really knew his stuff. From my experience over the years, the golden rules remain untarnished.
Last edited by KenWittlief : 02-11-2005 at 10:00.
|