Re: Bridgeport CNC Mill???

I retrofitted a large hmc with emc2. I love it. This is a 60's vintage machine that originally had hydraulic servos. (ball screws thoughout and rollerbearing ways) We converted it to 80's vintage inland brushed servos.. After a bit of research we went with mesa hardware. (2 5i20 fpga boards and 2 analog servo interface daughter boards (7i33 and 7i48) < $600 and gives us a true closed loop machine control with 96 i/o and 10 analog/encoder interfaces. we purchased amc analog servo drives off of ebay cheap and opto22 boards for i/o interfacing. Gives us 4 tons of cutting force continuous - twice that peak.

Most of the pallet/tool change control was done in ladder within emc. I did the gear shift control within emc 'comp'. Sort of a C wrapper for creating realtime emc modules. I could have done it in ladder but I could wrap my head around it better programmically.

I am using 6 of the encoder inputs. 4 axis (xyzb) one jog wheel and 2 encoder on the spindle. (one for the motor rpm and one on the spindle for rigid tapping) I do some checking to make sure the spindle is in the right gear (16 gears) I am using 5 of the analogs outs 4 axis + spindle vfd. I have used almost all of the 96 i/o. This is expandable though - so I could add more mesa hardware or even use the printer port of i/o.

This is what the machine originally looked like
http://electronicsam.com/images/KandT/oldkandt.JPG

here are a bunch of random videos.

http://www.youtube.com/user/samcoinc

The only non generic thing would be the mesa hardware. I could change that out with pico systems hardware or even vital systems or servos to go. It would require some re-wiring and some hal file re-writing. Emc is really quite flexable. It has blown me away.

sam

--- In CAD_CAM_EDM_DRO@yahoogroups.com, Les Newell <les.newell@...> wrote:
>
>
> >
> > The vast majority of new tools, today, are cnc tools. Purpose of this group.
> > Of those, maybe 1% have inverse kinematics. At most.
>
> Of course lathes, 3 axis cartesian machines and 3 axis + rotary don't
> need it. That covers most of the machines out there. However once you
> get into 5 axis then kinematics are needed. Actually kinematics are very
> useful for hobby machines as well. My lathe for instance has a worn bed.
> This means that the diameter is dependent to a small degree on the Z
> axis position. Kinematics can be used to compensate for this.
>
> My little Orac lathe running Mach3 tends to turn tapers. As I can't
> compensate mechanically or in kinematics I have to do it in code which
> can get rather messy.
>
> > The vast majority of lathes in the world, today, maybe 99%, DO NOT have
> > DROs.
>
> I assume you are talking about manual lathes. Manual lathes are dying
> out fast. Walk into any machine shop producing turned parts. You will
> almost always find mostly CNC with maybe a manual hidden in a corner for
> odd jobs. Hell, a lot of modern operators don't even know how to use a
> manual.
>
> > And for a (quite) small subset of users they can certainly enhance their
> > effectiveness a great deal.
> > No-one has ever disputed the desirability of certain features.
> > Desirability "is not eq" necessity.
>
> Take that to it's ultimate conclusion. How many hobbyists NEED a CNC? In
> reality very few. That still doesn't stop thousands of hobbyists from
> owning one.
>
> > Now ..
> > as regards proper position feedback, if you mean by that that machine
> > shave dro´s that is one thing, and that new CNC machine use servos which
> > have a tolerance within which they move, that is certainly correct and
> > desirable.
> > Nevertheless this ignores the fact that it was not always so, even for
> > industrial machines.
> >
> > Now, from the point of view of a user, for most it is immaterial how
> > closely they servo tracks, as long as they make parts to tolerance.
>
> Knowing the following error of a servo is of no real use to a user. That
> is purely for the guy setting it up in the first place. What I am
> talking about is the fact that the control always knows exactly where
> the machine is. Mach or any other control running step/direction only
> knows where the machine should be. If you estop for instance the
> machine's actual position may not correspond to the control's prediction.
>
> > This is one of the very, very good features of EMC2, that you can often
> > analyse this (I have no idea whether this is possible with most advanced
> > hw, and I believe generally you cannot do this [w. avd hw that is]).
> > I believe that with most advanced servo cards/hw, you cannot see how
> > closely the servo tracks during programs execution.
>
> This is true for step/direction servo cards. Systems using true closed
> loop will know the following error at all times and if you get into
> their configuration functions you will be able to see the error. However
> as I said before this is purely for the guy setting the machine up. Once
> set up the following error when cutting should be considerably less than
> other mechanical errors such as flex.
>
> Les
>