Re: [CAD_CAM_EDM_DRO] Re: Safety -- servo vs. stepper

Leslie M. Watts wrote:

>What I hope a discussion like this generates is really an
>informal FMEA (failure mode effects analysis).
>
>As a product designer I insist on having a FMEA for all designs...
>and I think this is equally appropriate for hobby cnc applications.
>
>Granted the major safety aspect might be the unhealthy state of depression
>when your machine self-destructs due to an unforseen failure mode... :)
>
>So here's my example FMEA for EMC servo using the stg card. I will try to
>list any failures that would cause machine damage or safety risks in order
>of probability.
>
>Failure mode effect
>
>1)Spindle stops during machining......... damaging forces due to continuing
>motion
>
>2)Mains power loss....................... immediate estop from latching
>relay
>
>3)encoder failure........................ immediate estop from following
>error
>
>

The estop won't necessarily be immediate! The computer (or Gecko drive)
senses no movement, and applies more power. If the G-code program is
calling for a slow move, but the drive is accelerating wildly trying to
"catch
up", it might take a while before the error limit is reached.

I don't have a real great solution for this one, either, except to set
up the servo
drive so it faults out on overcurrent when the servo goes open-loop.

>4)servo amp short........................ immediate estop from following
>error
>
>5)computer lockup........................ possible runaway; estop on
>overspeed or limit switch
>
>6)program error.......................... estop on overspeed indication from
>tach
>
>7)i/o card failure....................... worst case estop on overspeed or
>limit switch
>
>
>For me the worst scenario right now is (1)... trying to cut with a stalled
>spindle
>due to motor failure or just a too agressive programmed feed. To that end I
>have designed
>a rotation sensor in the hardware estop system.
>

I have the fault signal from my VFD wired into the E-stop chain. A
spindle stall
would trip the VFD, but a broken belt, for instance, wouldn't.

> I am not coverd as much as I
>would like
>on (5), (6), and (7). A watchdog could take care of that.
>
>

The Servo-to-Go card has provisions for a watchdog connection. I have a
diagram
of how to make it work on my web pages at
http://jelinux.pico-systems.com/EMC.html
See the first schematic on that page. The center of the drawing is all
the watchdog
circuit. It also cuts off +5 V power to the opto-22 board with the
solid state
relays, to positively shut off the spindle and servo amps.

Jon