On Sun, 2012-02-05 at 12:41 -0500, N. Christopher Perry wrote:
> I'm bringing up my mini-mill and have encountered a weird problem:
> I'm trying to use pins 10, 11 & 12 on the printer port as limit switch
> inputs for X, Y, Z axes, respectively, but the pins are acting like
> outputs.

The parallel port 10, 11 and 12 pins should always have high impedance,
so your sensor circuit outputs should act the same way whether they are
connected to these pins or not. If connecting the sensor output wire to
the parallel port pin keeps the signal voltage from toggling, then there
is something wrong with the parallel port. Many motherboard parallel
ports are now running on 3 Volts, I would think the inputs should be 5
Volt tolerant, but they may be very easy to short out. I avoid using the
motherboard parallel ports. Blowing out a $15 PCI card is much cheaper
than replacing a motherboard, plus these usually run on 5 Volts.

It should be easy to check these inputs with a wire and a 2.5k Ohm
resistor. Connect the wire to an input pin, then touch the wire to the
PC's ground or a +5 Volt source. You should be able to see the logic
state change with HALmeter, HALscope or the "watch" feature in the "HAL
configuration" window.

>   The limits switches are active high, with ether voltage dividers or
> diodes in line for level correction and short circuit protection.
> When I test the limit switches when not connected to the PC they work
> as expected.
> When connected to the PC a tripped a limit switch might cause the
> voltage to move by 0.5V or so, but the voltage is still held above the
> TTL high threshold.  
> 
> The X-axis limit switch system consists of two OPB972 optical sensors
> (TTL level output), which have totem-pole output.  Both are diode-ORed
> together with a 20K pulldown.

I would avoid totem-pole outputs. The limit signal should be active low
so that if the sensor, power supply or wire fails, the limit will trip.
Open collector outputs make that easy. One just needs a pull up to the
supply voltage to limit the collector current. A divider would work too,
but I would rather have the full voltage drive an opto-isolator
protecting the parallel port input.

I avoid optical sensors unless they can be enclosed in a liquid proof
container.
> 
> The Y-axis limit switch system consists of two Honeywell 103SR12A-2
> Hall sensors, which have active source outputs (Open emitter, 12V
> supply, ~12 volt active output, floating otherwise).  Both are wired
> together and put through a 5.1K/2.2K resistor divider.

The 103R's are a nice sensor, but they are expensive. I would go with
the sink version of this sensor, but $.60 can get you a sensor that
works very nearly as well in a hobby environment.
> 
> The Z-axis limit switch system consists of two Parker Proprietary Hall
> sensors (TTL level output), which appear to have totem-pole outputs.
> As a precaution, I've diode-ORed them together with a 20K pulldown.

Another thing, proper soft limits should be setup. With these setup,
there is another layer of safety and they are more convenient because an
axis will come to a controlled stop when it hits a soft limit. One can
just jog away from the limit, whereas hitting a hard limit requires
finding and selecting the limit override, then backing off. Wiring each
limit for each joint to its own input is also more convenient than
or'ing different limits together. PCI parallel ports are cheap and can
provide plenty of extra I/O.


-- 
Kirk Wallace
http://www.wallacecompany.com/machine_shop/
http://www.wallacecompany.com/E45/index.html
California, USA


------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to