AVR-Chat

Dave,
Sounds like a failsafe design is required. Most failsafe 
implementations require switches that have both a Normally Open and 
Normally Closed contacts. This would require 2 conductors, wired in 
series, to each switch. Depending on the type of switch, the contacts 
are either "make before break" or "break before make". In either case 
the intermediate "break" state is usually not more than a few ms. You 
would then monitor both the Normally Open and Normally Closed 
circuits, requiring 2 inputs. If both circuits remain Open or Closed 
then we have an indeterminate state that should be handled as if the 
machine cover is open and an indication that maintenance to repair 
the problem is required. 

--- In AVR-Chat@yahoogroups.com, "Dave Mucha" <dave_mucha@y...> wrote:
> 
> Correct, it is a series loop.  
> 
> It is for a machine where the wires can be cut or vibrate open or 
be 
> shorted.
> 
> The current or pulse would offer a more accurate way to verify that 
> the switches are in the circuit.
> 
> Dave
> 
> 
> 
> 
> 
> 
> --- In AVR-Chat@yahoogroups.com, Dave VanHorn <dvanhorn@d...> wrote:
> > At 09:42 AM 11/9/2004, Dave Mucha wrote:
> > 
> > 
> > 
> > >Hi all,
> > >
> > >I am looking an alarm type switch and am thinking of using a 
Tiny11
> > >to generate a pulse of something like 10khz and then send that 
thru
> > >the 6 to 10 switches (in series) and then use the same Tiny11 to
> > >detect that pulse.  Any switch opens and the pulse is lost.
> > 
> > This is just a series loop, right?
> > No need for pulses, just give a current of 10-20mA and detect 
> open/closed 
> > with debounce of about 100mS. Protection of your inputs and 
outputs 
> is more 
> > interesting.
> > 
> > For a moment, I thought you meant addressable switches, as 
> in "which switch 
> > is open".
> > 
> > Years and years ago, I did something like this in discrete logic 
> (yecch!)
> > It sent a wide sync pulse, then narrow polling pulses down a 
single 
> conductor.
> > Another conductor carried back answering pulses from each switch.
> > 
> > Today, a single Tiny-11 at each switch point could work, but 
you'd 
> have to 
> > program in their address ranges. A micro with more pins would 
allow 
> a 
> > dipswitch to set that. Four pins for a 0-F dipswitch, and four 
> switch 
> > inputs, plus two pins for data, would be nice.