Depends on how much time you have on hand.
Personally I have so much stuff to do that I just can't buy off the
shelf for a reasonable price. I get plenty of uC time there, so I have
no desire to re-invent the wheel for practice.
So for me it's always the most simple solution I can possibly find
that gives the required performance at the required price in the
required time. If that's an off the shelf component, that's fine too.
Not always are the controllers suitable, I had to replace one with my
own circuit because hysteresis was too much. I left the controller in
place for a readout display, and as a safety circuit with a slightly
higher setpoint.
My laminator (built from a copy machine fuser) still runs just fine
with the single transistor controller. That was just meant as a "quick
test" one day... but at least I put it in a plastic box to reduce the
electrocuton hazard. ;-)
ST
On Sat, Feb 13, 2010 at 2:31 PM, sailingto <sailingtoo@...> wrote:
> You are correct there - that approach is cheaper.... and Stefan is correct - a person "could" spend $30+ for a temperature controller that would hold the temperature closer to setpoint.... but where is the fun in that? This was a chance to etch a PCB, design a circuit, and do some uC programming. Great fun:)
>
> Had I purchased all the components the cost would surely have been more than buying - but due to the generosity of a friend and my spare parts bin, I think about the only thing purchased for the temperature sensor itself... oh, and the $3 Radio Shack box. I have several of the $4 uC chips laying around and this was a perfect chance to learn some code for the uC.
>
> Ken H>
>