Re: tool changer logic- mathematical challenge

Cheating huh?
Well, you are right. It's hard to describe this in words. You
see what I'm saying though.

Again, this is an existing setup, an Emco Turn 120 lathe. I have
no clue why they built it this way, but that's what they've done.
I'm really hoping to get it working as is. It's a small assembly,
that would be very difficult to add microswitches or even an encoder
to without major rework.

My problem now is how to implement the logic. With another $130
breakout board and a $10 PCI parallel port card, I 'might' have
enough inputs to do it. Then I need to pull out some magic VBscript
for a tool change macro in Mach2. However, I still don't know if I
will need 4 inputs (action when going LO and HI) or 8 inputs (4 LO
active and 4 HI active).

I'm kinda leaning towards a $99 DL05 PLC from Automation Direct.
It has 8 ins and 6 outs, plus upgradeability. I figure I can just
send an on/off signal from Mach2 and let the PLC do the work. With a
little homework I think I can program in ladder logic. It might be
overkill for this, but I've got some other projects in the future
that would be well suited for a PLC. So, this would be a great
learning exercise.

I didn't think about a PIC, but it's basically a build it yourself
PLC if I'm not mistaken.

Do you have personal experience with PIC's? I wouldn't mind
hearing more about them.

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "industrialhobbies"
<admin@i...> wrote:

> Now you're cheating :), your changing the problem, earlier you
> said "each" and relating to state change. I had the 135 (45 tool
> offset plus 90 sensor offset), but now you mention 2 sensors at a
> time. Technically, to solve the word problem part, you can also

225

> degree opening.
>
> Word problems suck. Give me a digital camera any day.
>
>
> As far as making it work… Think PIC
>
> For a $50 setup you can make is do a little dance and sing a little
> tune whenever you request a tool change and still get the position.
>
> Seriously, a PIC is a cheap good little solution that you might

want

> to look into. The PIC itself is just a few bucks, the PIC burner

is

> the rest.
>
>
> Thanks
> Aaron Moss
> www.IndustrialHobbies.com
>
>
>
>
>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, "Lance Hopper"
> <snaggletto@c...> wrote:
> > Hello,
> > You seem to be on the right track, but not there yet. You

have

> to
> > think of it as this;
> >
> > What angle of opening, will create 2 trip points, so that sensors
> at
> > 90 deg. to eachother will trip every 45 degrees?
> >
> > Imagine an arrangement of 4 sensors at 90 deg. "+" represented

by

> a
> > crosshair. Now imagine a "C" shape. Put the + into the C and
> align
> > one point of the + with one of the endpoints of the C. This
> > represents a trip location.
> >
> > Now forget about the currently tripped sensor, look at the

next

> > one, which is currently out in the opening of the C. We want

this

> > one to trip in another 45 deg of movement. So, 90 (between
> adjacent
> > sensors) and another 45 deg. of movement for the not-tripped

sensor

> > give 135 degrees.
> >
> > If you lay it out in CAD you will see that every 45 degree
> > rotation causes one of the 4 sensors to trip on one side of the

135

> > deg C opening every time. However, this means that I will have

to

> > check each sensor for when the go from 0 to 1 and for when each
> goes
> > from 1 to 0. In other words, it appears I will need 4 active LO
> and
> > 4 active HI inputs to read the position of the turret. Also, one
> > output for a DPDT forward/reverse relay.
> >
> > I'm not so sure Mach2 with 2 parports will do this. I'm kinda
> > thinking about a small $99 PLC with 8 inputs and 6 outputs to do
> the
> > tool turret. Mach2 will just send a 'start' signal to tell the

PLC

> > to run it's program, then maybe recieve a 'done' signal from the
> PLC
> > after the tool change is complete etc... Got some thinking to do
> on
> > this.
> >
> > --- In CAD_CAM_EDM_DRO@yahoogroups.com, "industrialhobbies"
> > <admin@i...> wrote:
> > > This is just a WAG (wild ass guess).
> > >
> > > To do it with only one sensor active (or not) at a time 45

degree

> > > opening 22.5 degrees out of angular alignment with the number 1
> > > position.
> > >
> > > That would meet the requirement of a single sensor (of four),

and

> > in
> > > this case any opening between 22.5 and 45 degrees would work.
> > >
> > > Somewhere else you mentioned a backup circuit, backing up to

hold

> > > position, you would need to change the angular alignment from
> 22.5
> > > degrees out, to 0 degrees (just enough to trip the sensor).

That

> > > would allow you to index all the way around through each of the

4

> > > sensors and change the state of one sensor at a time, based on
> only
> > > 45 degrees of motion.
> > >
> > > The offset of angular alignment is only needed for real world
> > > applications, theoretically it's not needed.
> > >
> > > Thanks
> > > Aaron Moss
> > > www.IndustrialHobbies.com
> > >
> > >
> > >
> > >
> > > --- In CAD_CAM_EDM_DRO@yahoogroups.com, "Lance Hopper"
> > > <snaggletto@c...> wrote:
> > > > Test yourself, see how good you really are...
> > > >
> > > > Given; a ratcheting action, rotary tool turret with 8

positions

> > (a
> > > > tool every 45 degrees). The turret can only index going one
> > > direction.
> > > >
> > > > Given; a series of 4 optical sensors, spaced evenly @ 90
> degrees
> > > > around a circle.
> > > >
> > > > Given; a "C" shaped, semi-circular raised area that passes

thru

> > and
> > > > blocks/unblocks the optical sensors as it rotates with the

tool

> > > > turret. This creates 2 state change areas for each of the 4
> > > > sensors. Assuming the "C" shape, and CCW motion, you can see
> > that
> > > > each of the 4 sensors will go from blocked to unblocked, and
> > > > unblocked to blocked as the turret rotates.
> > > >
> > > > Problem; how many degrees is the opening of the semi-
> circular "C"
> > > > shape that will create an off - on and on - off transition

for

> > each
> > > > of the 4 sensors. Resulting in an index resolution of 45
> degrees
> > > > when only looking at one sensor at a time.
> > > >
> > > >
> > > > This is easier than it first appears, however, I never could
> > figure
> > > > it out without help. I thought this might be an interesting
> > > > challenge for others. See what you come up with. Good luck.