Re: [CAD_CAM_EDM_DRO] Re: safety

mariss92705 wrote:

> Jon,
>
> What happens if an output transistor fails? You can't turn off a
> shorted transistor. Please see the "A Servo E-Stop Circuit.pdf" in
> the "files" section of this list. This circuit will stop motors even
> if that were to happen.

Well, there are multiple inputs to the e-stop logic. The computer monitors
following error, and calls for an e-stop if that limit is exceeded.
The servo amps monitor motor current, limit it to a value sufficient
for the acceleration needed plus some margin for cutting forces, and
there is a fault limit set above that. If the amp goes to fault status,
it trips a relay breaking the fault chain, and causes an E-stop. It
also disables the drivers to the power transistors in the full bridge.
(I must comment that although I have blown fuses, had hundreds of
fault trips, and several runaways and wild oscillations while developing
the servo amps and tuning them in the milling machine, I have NEVER
bolwn even ONE transistor since getting rid of the flawed Harris
HIP-4080 chip and going with the IR 2113S drivers.) But, it could
happen! So, there are 2 paths to stop things under almost all
conditions. The E-stop logic informs the computer that an external
E-stop has ocurred, and it sets the servo DAC values to 0 volts.
It also turns off the aux I/O devices (servo power, spindle and coolant,
etc.) through the parallel port. The E-stop logic also cuts off +5 V
to the other side of the solid state relays, so even if the computer
fails to turn off the bits of the port, the SSRs are de-energized,
anyway. One of these SSRs provides 24 VAC to the relays that
power the servo amps, so they get cut off, and an energy dump
resistor acts to brake the motors through the body diodes
in the FETs. Finally, there are strategically placed E-stop buttons
that will cause an E-stop manually. Also, there is a watchdog timer
that will cause an external E-stop if the CPU fails to update the DACs
in a reasonable time.

There are STILL single points of failure in this system, but since it is
excercised at LEAST once every time the machine is fired up an shut
down, a failure of safety would probably be noticed when it ocurred, and
could be investigated before an emergency would need it. I might not
trust it with a 50-ton K&T mill, but I think it is probably reasonable for
a small, home-shop machine.

Jon