What is wanted in a controller, was Re: Correct or Max voltage for Bridgeport Nema 42? + more...

Alan,
It was I to blame for that long rant.
Believe me, I would be more than happy to take pictures and provide
copies of manuals to any developer eager to learn about how a true
CNC controller works. I can get many varieties of both.
The standard is somewhat as follows though....
You have a physical control surface attached to the machine. It
consists of buttons, a jog/hand wheel, and a display.
Normally any G-code is done manually at the machine or loaded in
having been completed via CAD/CAM software. The most important
aspect of a controller is safe and effective interaction with the
operator at the machine. It is not used as a design tool in itself
normally although some late model high-end systems have encorporated
this. You will have a switch on the machine that will change the
display through several modes (MDI, Jog, auto, edit, etc)
Each "window" will have information pertaining to those funtions.
Now here's the tricky thing. Older systems and even some newer ones
(Haas/Fadal) etc use a DOS style format. What the DOS style
aftermarket programs miss though is the physical buttons and switches
that accompany the original controls. An example would be a feed rate
adjust knob. The actual shop floor operator hovers over a cutter and
dials in with this knob to get just the right "sounding" cut. This
can be done with a PC keyboard, true, but the operator must deal with
some function on a PC. This could be overcome by software/hardware
that would allow you to wire a real knob into the system (analog
input) The same goes for jog wheels and other physical buttons.
The flip side of the coin is the windows based controllers (or PC
based really). They can do without many of the physical buttons by
representing them on screen. Of course, in order to do this, they
must be of a size that can be touched. In either case, a CNC control
is no place for a mouse. You either need real buttons or virtual
buttons In the case of a real control, that's a whole lot of buttons
and if you had to make it work with a variety of different machines,
you would need some serious flexibility in what each button would
instruct the software to do. Some sort of macro writting setup
utility would be needed.
Follow this link to a VMC we just sold...
http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?
ViewItem&item=1699565569&r=0&t=0&showTutorial=0&ed=1013226742&indexURL
=0&rd=1

This control is a 94 and lies somewhere in between a DOS based and
Windows based control.
It doesnt really do anything except control every aspect of the
machine itself. If the air pressure gets low, it tells you. If the
door is open, it tells you. It will tell you the spindle speed, load,
etc. Almost nothing can happen on this machine that isn't monitored
by the control which means alot of I/O.
The control interacts with the operator in that controls needed for
setup are in one "mode". Screens for running the cycle are in
another "mode", Screens for editing code are in another "mode"..etc.,
tactical devices (ie jog wheel) are real and in reach (and hard to
miss)
So how does this really differ from the hobbiest control? Well, not
by much really. Basically it can interpret alot of I/O from the
machine and the physical operator controls. It is safer because it
can halt the execution of it's cycle if certain conditions are not
met (ie door open or axes not aligned, or low air pressure (some
spindles are air supported))
Now this machine is a 94 and if the controller were bad you would
probably replace or fix it, but many older machines have the same
features but with less intuative (as in single line LED display) and
often broken controls.
It is because of this that you can buy full featured machines for a
song. The electronics are often no longer supported and therefore
expensive or hard to find. What is needed is to be able to retrofit
these machines with something similar to the 94 I showed you.
You wont find any professional machine being run by a mouse or
joystick even if it has gotten cheap enough to fall into the hands of
a hobbiest or "semi-pro".
Now take the pro level CNC control retrofits...
When the control on a 10 ton machine that is bolted in place goes
down, it's not the easiest thing to call England to fly out a
technician at 200 bux an hour and the replacement cost of the machine
is often 6 digits before you even think about machinery movers and
riggers etc. By the time that 20 year old control breaks down 5 times
in a year it has become a liability, so cost minded production shop
owners will retrofit a 3-10k$ control in to save headaches and
possibly even increase production. That's part of their business and
these retrofits seem to be aim at that market.
Meanwhile, smaller, but still professional quality machines are going
to the scrapyard because it's just faster and easier to buy or lease
a new one than spend 5 k on a retrofit.
So here you are... you just bought an older pro machine for 4K and
you are either stuck with playing with a mouse and losing half of the
capability that it even had 2 decades ago, or spending more than half
of the machine's value on a pro level retrofit.
Now, don't get the wrong idea. I'm not saying the hobbiest controls
are toys. In fact, I'm saying quite the opposite. The features found
in some of the more evolved hobbiest's controls surpass the abilities
of the original controls, and that's the rub. If the developers would
recognise how close they are to controlling older pro level CNC's
just by adding flexible I/O, then a lot of these high potiential
older machines wouldn't be stuck between losing capability and
costing more to retrofit than they are worth.
Some developer who is willing to create a true control, that will at
least integrate with the machine as much as the original, and offer
easy shop floor interaction with the operator, is going to hit the
nail on the head and every Joe with a garage is going to start
building really great machines from this excessive late 70's/80's
supply. Based on the cost of these machines, the retrofit has to come
in between 1k and 1.5k including motion control cards and computer
which puts the software cost at about 500. This is 5 times the cost
of some sofware out there and I think many would pay it to get
control over a "real" machine. I know I would.
Perhaps adding this flexible I/O to some current software is a
monumental task. I'm not a developer so I don't know.
What I do know is that there is a market.

Sean

--- In CAD_CAM_EDM_DRO@y..., Alan Marconett KM6VV <KM6VV@a...> wrote:
> Hi Carlos,
>
> Speaking as "one of the programmers" of a hobbyist application,
Yeah,
> that's about how it starts. Move some axis, first linear, then
circular
> and helical, add in some Mcodes, then a "plot" screen to show the
tool
> paths without cutting metal.
>
> Speaking for myself (and no doubt others), I've never seen a "Real
CNC
> mill" (well, maybe just a little at a trade show), so TRUE, I don't
know
> what it looks like and how it operates. But then, I haven't been
> trying to "duplicate" an old mill controller (perhaps we should??)!
>
> And yes, I have a pendant, again, I'm sure it's not the same as on
an
> old mill. You mention touch screens, do the old machines have
that?
> For hand held computers in automotive shops anyway, the current
thinking
> (ours) is that you'd quickly mess up the screen! And a machine shop
> would be a harsher environment.
>
> Yes, a D/A board could be added to control the spindle speed. I've
been
> thinking of doing that if I ever get a new Sherline mill w/ DC speed
> control.
>
> What other devices ("other then S/D") were you wanting? Hard for me
> (us) to visualize all the big guys toys! Perhaps if we could study
a
> machine/manual... might help!
>
> Thanks for your thoughts!
>
> Alan KM6VV
> Central coast California
>
> P.S. Anybody wanna LOAN me an old machine to study?
>
>