RE:hand jive

Good Ideas Brian,

I figured you might be making your own shaft encoder! Yeah I do have a
STACK of old 360K drives, I was pulling the nice linear stepper motors,
precision shafting, driver chips, etc. out of them. Someday I'll melt
the aluminum frames for castings! I had forgot about the joystick
port. Good idea, as I was intending to use the defined pinout for
MaxNC, and as I recall, all pins are used! That does give one a change
to get the jog controls away from the computer, and back to the mill.
Other nice item to have is a "keyboard pendant", limited number of keys
that can be mounted close to the mill also. I like the idea of a shaft
encoder for jog, easier then pressing keys! If it's true for radio
receivers, it's probably true for machine control.

I actually made shaft encoders when I was working on my big robot. In
fact, the indexer jig posted on the Sherline files list is making just
such an encoder disk:

http://www.eGROUPS.COM/files/sherline/IndexJig.jpg


As I recall (about 15 years ago), the photo-optoisolators were from disk
drives (maybe 8" drives?).

I just had a thought, perhaps some of the "software writers" on the list
would consider an optional shaft encoder input for the jog buttons!
Dennis, Doug, Art, Ron, (my apologies to those I've missed) what do you
think? Can we come up with a standard pinout for 3 or 4 channels of jog
quadrature lines (A&B)? Or perhaps the option would be to use a single
knob? Joystick port, or another parallel port? Or is a DRO interface
with US Digital chips the way to go (Tom)?

http://www.mcs.net/~kulaga/dro.html

Didn't know the 74hct245 was Schmitt Trigger, I seem to remember
74LS14's (Schmitt) and 74LS86's used for encoder circuits.

Alan Keep on jog'in!


Brian Pitt wrote:

>
> Alan
>
> I think a lot of it is that people want to have something they can
> grab hold of to feel in control ,we "need" the feedback that buttons
> just cant give and we are more at home with knobs and switches that
> we can feel
> deciding on electronic vs. real handweels can depend on the
> size of the machine ,the biggest machine you can reach the ends of
> all the axes at once is about the size of a B'port OTOH as
> Randy pointed out some work is too small to go by feel alone
>
> I built a handweel from a full hight 360k drive, the spindle shaft and
> bearings became the encoder shaft and the flyweel became a detent
> plate with 100 indexed holes around the circumfrence and a knob fixed
> on top
> the disk hub end got turned down to a thin encoder disk with 25 slots
> the opto pairs and spring detent are mounted in an aluminum housing
> that I built for it ,a dividing head or indexer comes in handy for all this
> the hard part was lineing the optos up so that each detent position was
> at a seperate clean encoder state
>
> the opto outputs (A&B) are conditioned with Schmitt triggers (74hct245)
> and for now I'm playing with it on the 4 button inputs of the joystick port
> (the analog inputs are useless but you get 4 digital bits for free)
>
> the software loop only needs to be fast enough to keep up with the
> loose nut spinning the weel ;-)
> the speed limits are ,how fast can the axis catch up and the human
> reaction time (slow in computer time)
> most of the time you will be turning the handweel slowly enough for
> the machine axis to catch up ,any faster and you will want the
> jog state to reset for the next loop (the axis will still be moving)
> and check for a state transition (keep moving Y/N?) you may
> get one or two clicks more/less than you wanted at low jog increments
> (.0001) if you could spin the weel as fast as your arm can move
> and still keep count
> its not all that important that the knob stay synced with the machine
> position ,if anything you want the machine to fall behind of you spin
> it to fast
>
> Brian
>
> On Sun, 05 Nov 2000, you wrote:
> > Brian,
> >
> > Thanks for the comments. Expensive RF test equipment usually have knobs
> > to spin. They may be shaft encoders, but manufactures have found that
> > users want knobs, not just simple buttons. I think I'm seeing that it
> > must be similar in the CNC control world.
> >
> > Well, you have track 0 sensor, cyl 0 sensor, spindle motor and a track
> > stepper. And a "disk inserted" sensor. Not a lot more that I can
> > remember off hand. Some of the early 8" disk drives had hard sectored
> > disks, but I don't see much help there. What's your suggestion?
> >
> > I have 256 count panel shaft encoders (HP QEDS 7000) hooked up as A & B
> > quadrature. A little hardware would help here! Looks like a fast loop
> > is needed to watch for A & B changes, and ESC out of jog.
> >
> > Alan