Re: Simple, safe, effective limit switches?

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "Abby Katt" <cnc@a...> wrote:

>
> Hi again my ever resourceful allies!
>
> I've been having this terrible recurring nightmare where an encoder
> fails and a servo wedges the table as hard as it can against the

end

> of the mill and the little controller that could keeps feeding more
> power and the servo becomes an induction heater.. and... and... I
> lose sleep over it at night..
>
>
> So, let's talk about limit switches.
>
>
> I have an idea of how I want to setup my limit switches, and I'm
> about to buy the stuff for it, but I was wondering what feedback I
> could get first, just in case I'm wrong (or things can be done
> better).
>
> What I was thinking of doing was serial-chaining a whole bunch of
> microswitches in NC (so they pass current through them normally,

but

> break the circuit when they hit limits). This would mean, normally

my

> current would flow in the circuit, but as soon as any axis hit the
> limit, or if I pushed the emergency stop button (also in series),
> then the circuit would be broken. The chain would drive a big relay
> (20A). The relay would sit on the servo driver side of the power-
> supply capacitor, so that when the relay was off, there would be no
> power for the geckos (not even the resisual in the capacitor, since
> this would be disconnected from them). A second relay could

activate

> and this could be used to indicate to the computer that the driver
> had faulted/E-stopped.
>
> So, is this a good idea? Or will the geckos burn from suddenly

having

> motor back-EMF and no cap to discharge it to? Or will it not

matter,

> since, diconnected, the servo circuit would not be closed? What

about

> the cap, now stranded and connected only to the beefy PSU? Would it
> steam and boil and fill my shins with aluminum electrolytic ionized
> shrapnel?
>
> or is this a workable idea?
>
> Thanks!

The problem with controls is that there is a natural balance between
function and simplicity.

Considder that a hot chip tossed onto the wire as it exits the
control box would melt the wire and complete the circuit. No
external switches would work.

Another option would be to wire both the NO and NC of the switches so
that in order to see a state change, the switch would have to go from
one condition into another.

I call that active. vs. what I call passive. that is a state where
the switched went from passing power, to not, but there is no idea of
if the wire was cut or the switch worked.

So, my addition to your circuit would be to wire your switches so
they are passing power all the time in normal. ANY disconnecton
would be your first clue, and the paralle circuit becomming hot would
be a confirmation.

I would also make the last switch on the loop the home switch and add
some G-code so when the unit started, it move to zero -0.1" to prove
the switch condition.

Dave