On 6/14/22 02:25, Sebastian Kuzminsky wrote:
I'm working with Pere/Petter Reinholdtsen on converting his Mazak > 15/40. It looks like a great machine, capable and compact, but we've
> run in to a bit of an issue. > > The machine has servos with
resolvers, but the resolvers only connect > to the servo amps, not to
the controller. The are connectors on the > amps for passing the
resolvers on to the controller, but those cables > are not present.
This sounds somewhat like the situation with stepper/servo's I am now
using 4
of on 2 different machines. These stepper/servo's have a quad encoder
on the
motor that only goes back to the LCDA357H driver. Available in 1,2, and 3 NM
motor sizes, the stepgens output is the feedback to motion, so there is a
disconnect there. But the driver also has an alarm output it sets if it
can't get to
where it was told to go by the stepgen. I have that coupled to the F2
button in
both machines, so a fault shuts it down in its tracks with the estop
circuit, It has
never tripped on the sheldon except for testing it.
So I have to pop off the tool so I can re home it, re-enable F2 and re
home, fix
the gcode mess-up, likely a typo that caused the fault, and rerun the job,
probably on the same part. The fault shutdown is so fast its rare that
the part is
fubared.
If the motors run to position commanded, and there is a signal from the
Magnascale following the axis motions that could be used for feedback,
and the
servo driver has a "can't get there" output that you could wire up to the F2
button, you might be able to do in servo's, what I'm doing in steppers.
These 3 phase motors turn only 1.2 degrees a full step, and can
microstep just
like the older 2 phase steppers.
The biggest problem in using them once hal is configured to drive them,
is the
divisor table silk screened on the driver is straight off a 2 phase
drive, bogus
as can be. So theoretically for steppers, just multiply the SCALE for
that axis
by 1.5000. There's dip switches on the end of this driver to select the
microstep,
and I run all of mine with the #3 switch on. And it just works.
Might be worth a try, just for grins if it works. It sounds as if the
servo is self
contained, and the scales, if fast enough, could be used for feedback.
> Instead this machine has the fairly rare optional Magnescale option,
> so each axis has magnetic linear scales (Sony SR-721RM scale with >
MSD-9037 detector). These are connected directly to the controller, >
and the old controller used them for position feedback. These are >
circa 1987 or so.
If the servo runs to the position commanded and stops, your detector
should close the loop if you can deduce how to use its signals. I can't
advise on that since the only "servo" I have is one I built from scratch
to drive a BS-1 clone. Passably accurate to about an arcminute, but too
slow for dynamic use, takes over a minute to home it. Way too much
gear ratio ratio between the motor and the BS-1 clone cuz both are worm
gears. My mistake so I own it. when I needed a B axis on the 6040,
I screwed a 3nm stepper/servo to a 5/1 rvs30 style drive, printed a base
for it, printed the shaft adapters and chuck too. And I can spin it about
50 rpm. Much better than the BS-1 clone. So well I might make another
for the G0704, at least I can pick it up w/o an electric hoist.
Does anyone know how to interface these position sensors to > LinuxCNC? > > Another option might be to ignore the scales and
access the resolvers > by connecting to the currently unused resolver
outputs on the servo > amps. I'm guessing in this case the servo amps
provide the resolver > excitation signal and we'd try to passively read
the resolver > response.
That's the 2nd possibility I'd explore if the Magnascale output is too slow
causing Nyquist related instabilities. OTOH, putting quad encoders on the
motor and using that for feed back pretty much does away with Nyquist
related problems amazingly well. I made an extension shaft on the rear
of my G0704's spindle motor, put a $22 1000 ppr encoder on it, hooked
the encoder up as a tach and feedback to the PID, used a hall effect for
index from the spindle, and was amazed at the PGain I can use, up
around 40 before any instabilities show.
Speed to the exact rpm is maintained right up until Jon's servo driver
limits at about 17.5 amps which is close to 2hp from a 1hp motor. The
only way I know the tap is too big is listening for the motor iron
squeaking
as the limit kicks in.
Let us know what you did do that works please Seb. Too often that sort
of "what works" info never leaves the room, forcing the next builder to
re-invent yet another wheel. Put it in the wiki.
Thanks for any help...
Take care and stay well, Seb.
Cheers, Gene Heskett.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
_______________________________________________
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
