All a PLC really is is a small CPU configured in typically a ruggedized 
configuration, with additional Ruggedized I/O, and typically programmed 
using Ladder Logic.

In recent years, the cost of a PLC have dropped rapidly.   The first PLC 
I worked with is now being sold as the DirectLogic 305 series by 
Automation Direct (at the time it was GE Fanuc Series 1, but TI was 
starting to make modules for it).    Back in the 80's when I was doing 
this work, the costs were significantly higher.   Now, they're really 
inexpensive.

Take your application for example....   for a couple hundred USD 
(perhaps 300 w/ the user interface), you can have a complete "real" PLC 
capable of everything you describe.   The real strength of a PLC is that 
you can buy this off of the shelf hardware, hook up your control lines 
to it, and spend a couple hours programming it, and you're ready to go.  
And the "ruggedizing" needed in an industrial environment is pretty much 
already done.    It's also flexible (oh, you don't want it to run with 
the door open, well, just add a switch and a couple more lines of ladder 
logic, and off you go).      For industrial machinery where the actual 
programming in use is often unique to that specific machine, it's 
great.   If someone came to me with an application for an industrial 
machine or similar (like  your application), that only had a dozen or so 
total production runs, or if the requirements weren't stable from 
machine to machine, I would spec a PLC, and not even really consider a 
custom PIC solution.

Now, back to your application.    Because your application needs are 
fixed, yes, you should be able to build for a lower component cost.   
The cost of a 18F, and some I/O conditioning, plus the interface, is 
going to be significantly lower than the cost of a PLC.   The question 
being is how many units are going to be needed, and how "fixed" is the 
needs for each machine.   As long as the quantity is high enough, and 
the needs from unit to unit aren't going to require a whole new PIC 
application, then I'd say go for it.  

-forrest

Jinx wrote:
> I'm assessing a job, which I think a low-end 18F can do fairly simply. The
> logic is quite straight-forward. Inputs are two 4-20mA transducers, a mains
> relay, and two temperature sensors, with a real-time clock. Outputs are
> a couple of alarm lines, LCD, and a few pushbuttons to set parameters
>
> The scenario might be something like - If transducer1 is a certain value
> and temperature sensor2 is a certain value, then turn on the mains and
> measure temperature sensor1 after 1 minute to make sure the appliance
> (a chiller) is running. And various IF....THEN combos like that
>
> I was talking over the pricing structure with a couple of people and they
> both suggested that I base it at less than the cost of an 'industrial PLC'
> which would be needed. Their thought is that the lower (but not too low)
> price mentioned in the same breath as a PLC will make my PIC sound
> more attractive. Can't say I'm 100% about that but they won't shut up
> about PLCs !
>
> I'm a complete ignoramus in the PLC department
>
> http://en.wikipedia.org/wiki/Programmable_logic_controller
>
> To me, after reading that, it seems a PLC is way OTT for this job, and
> especially if you know how to program a micro yourself. This assessment
> may be for 300 units (optimistic "Everyone will want one" salesman !!)
>
> I guess the material costs, not including sensors, using a PIC may be around
> NZ$70 (US$35-$40), and software perhaps a little less (@ Q = 300), then
> there's markup on that
>
> Can anyone with practical experience of PLCs give me any advice as to
> whether the cost of a PLC system is even worth considering ? And, if so,
> how do prices compare ?
>
> TIA
>
>   

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist