Re: Freqmod problems

["Shelbyville Design & Signworks" <signworks-at-mcleodusa.net>]

so, if I understand you correctly,
add a linear guide/rack/stepper drive on the near end of the gantry?
Also, I am considering removing the 100 lbs or so of heavy sheet metal
and fabricating covers for the gantry out of .063 aluminum, should be as
stiff
, without the ridiculous amount of weight or overhang.

the master rail on this machine is a 1" Thomsom rail, with
2/1 opposing bearings horizontal, and 1 vertical.

the guide rail is 1" thomson, with a 3" cam follower riding on it.

erie patsellis

Shelbyville Design & Signworks
1309 W. S. 7th Street
Shelbyville, IL 62565
(217)774-4444
visit our website at http://216.43.233.169

----- Original Message -----
From: "Pete Cook" <pete.cook-at-alltracorp.com>
To: "Multiple recipients of list" <emc-at-nist.gov>
Sent: Tuesday, February 12, 2002 1:13 PM
Subject: RE: Freqmod problems


>
> we have the remains of one (the old cybermation control) sitting in the
back
> of our building. I don't miss not being involved in trying to keep one of
> those running. The y-axis sounds kind of like and old Heath machine (they
> were one of several absorbed by esab)
>
> Our plasma machines routinely run at cut speeds of 500 to 300 IPM with
rapid
> moves at 600 IPM (we have played with 1200 ipm but it doesn't make
> production all that much faster - you still have to accelerate and
> decelerate..) We have 1500 watt drives on each end and the y-axis and we
> still have to careful of the tuning to keep the cut quality where
customers
> want it.
>
> To stiffen your machine I would recommend putting a second drive on the
> other end of the X axis and connect the stepper driver in parallel with
the
> existing X drive. As long as you are running steppers that should cause no
> problem as long as both ends can keep up. That may also help reduce
> following errors it would take a lot of the flex out of the beam and rail
> system. Also check out the linears - when they are new, clean and tight
they
> work great, but a bit (or a lot) of wear and they can really bog down.
They
> can be hard to keep clean and the bearings will tend to slide rather than
> roll. We use an 8" hardened wheel to support the machine and ball bearings
> on either side of the master rail (rail is rectangular) and then run a
piece
> of 1/4" brass plate on top of the rail with a low pressure air stream to
> help blow the dirt away. Plasma & Oxy/fuel present some of the same
problems
> that wood & plastic dust present - it gets everywhere, put a little
pressure
> on it and it will stick and stay put unless moving would cause more
problems
> and then it will migrate to the place where it can cause the most problem.
> Wood dust though can burn, and does attract moisture helping things rust.
> >
> > Pete,
> > it is a gantry machine, with the ball bearing type
> > linear bearing on the master rail, cam follower on the other end.
> > Single axis drive. rail is approx 16" from top of table, with 8" of
> > clearance. y axis uses a carriage with 3 sets of bearings at
> > +_45 degrees to the rails top and bottom, this machine was
> > originally a cybermation cnc router, with the usual pdp-8 controller.
> > it ran in the 150-300 ipm range. From what I understand, cybermation
> > bought the tables from esab and added a bunch of controls to them.
> > I've spoken with someone who routinely retros these and uses 23 frame
> > triple stacks, a linear bearing on the cam follower end of the gantry,
> > and can get 400+ ipm, BUT you have to rebearing, and everything has
> > to be perfect, and stay that way. Not realistic in a sign
> > shop, where we
> > have
> > dust from routing, overspray, etc. i'd be happy with 200 ipm
> > and a 1 -2 mil
> > step size, as we route primarily aluminum, brass and wood/foam.
> > any help or suggestions you can give are greatly appreciated, as
> > when I have this one running reliably, my plan is to build a 5x10 for
> > more detailed work.
> >
> > erie
> >
> >
> > Shelbyville Design & Signworks
> > 1309 W. S. 7th Street
> > Shelbyville, IL 62565
> > (217)774-4444
> > visit our website at http://216.43.233.169
> >
> > ----- Original Message -----
> > From: "Pete Cook" <pete.cook-at-alltracorp.com>
> > To: "Multiple recipients of list" <emc-at-nist.gov>
> > Sent: Tuesday, February 12, 2002 10:48 AM
> > Subject: RE: Freqmod problems
> >
> >
> > >
> > > A couple more questions - is this a gantry or a cantilever
> > machine (1 or 2
> > X
> > > axis drives)? High rail(30"), mid rail (16") or low rail(8")?
> > >
> > > Many of those machines ran fine as Oxy/Fuel machines but
> > they were niether
> > > fast enough or stable enough for plasma speeds. For the
> > machines with a
> > > single X drive it was a fairly routine upgrade to hang a
> > second slave
> > drive
> > > on the other end ot the beam to stiffen them up.
> > >
> > > I have seen them run with 80 watt servos, but that was only
> > for oxy/fuel
> > > with 150 watt servos used on a plasma machine to handle the speed.
> > Oxy/fuel
> > > is normally in the speed range of 8 to 20 ipm with plasma
> > 50 to 150 ipm
> > (or
> > > faster on sheet metal).
> > >
> > > Also - for plasma & oxy/fuel there is no tooling force like
> > there is with
> > a
> > > router.
> > >
> > > Pete
> > > >
> > > > Pete,
> > > > I agree, the design of the gantry is less than desirable,
> > > > rigidity is not
> > > > it's strong point.
> > > > I have no idea why cybermation used these as the basis for
> > > > their router.
> > > > the "bounciness" is the reason why I'm going to a 550 oz-in
> > > > motor on the
> > > > x axis, and will be replacing the cam follower with a linear
> > > > bearing and
> > > > carriage as soon as I can find a few $ the wife doesn't
> > know about.
> > > > the next iteration will more than likely have a 4 x 8 gantry,
> > > > with linear
> > > > bearing at each end,
> > > > I'm hoping to reduce the mass to under 150 lbs.
> > > >
> > > > I bought this machine with a trashed controller, I still have
> > > > the original
> > > > servos, I've considered
> > > > mounting encoders and using G320's to drive it, but I have
> > > > several steppers
> > > > sitting around, so
> > > > I'll play for a little bit. All in all, I have $1000 invested
> > > > into this so
> > > > far, including a water cooled precise spindle,
> > > > I'm willing to trade off convenience for lower initial cost
> > > > as our business
> > > > is only 2 years old and far from
> > > > even thinking about profitability.
> > > >
> > > > erie
> > > > Shelbyville Design & Signworks
> > > > 1309 W. S. 7th Street
> > > > Shelbyville, IL 62565
> > > > (217)774-4444
> > > > visit our website at http://216.43.233.169
> > > >
> > > > ----- Original Message -----
> > > > From: "Pete Cook" <pete.cook-at-alltracorp.com>
> > > > To: "Multiple recipients of list" <emc-at-nist.gov>
> > > > Sent: Tuesday, February 12, 2002 10:00 AM
> > > > Subject: RE: Freqmod problems
> > > >
> > > >
> > > > >
> > > > > Question - Does Freqmod use encoders for feedback?
> > > > > If so what are the PID Parameters?
> > > > >
> > > > > I work on these machines (still as plasma cutters) and am
> > > > familiar with
> > > > them
> > > > > with servos attached. The biggest problem I have seen is
> > > > that most of the
> > > > > drive systems are too small and don't handle the bounce
> > > > that is inherent
> > > > in
> > > > > that type machine. If you try to drive tha machine without
> > > > gearboxes in
> > > > the
> > > > > drive train yopu are going to need a whole lot more
> > torque from your
> > > > motors.
> > > > >
> > > > > Pete
> > > > >
> > > > > >
> > > > > > Erie
> > > > > >
> > > > > > Hey I wonder if the following error problem with freqmod is
> > > > > > related to
> > > > > > the notion that we have explored before that if you run
> > > > at the max,
> > > > > > during accel you pile a few pulses into the following error
> > > > > > buffer that
> > > > > > can't get put back into motion.  You might try setting
> > > > > >
> > > > > > [TRAJ]
> > > > > > ..
> > > > > > MAX_VELOCITY =          1.67
> > > > > >
> > > > > > a little larger than
> > > > > >
> > > > > > [AXIS_0]
> > > > > > ..
> > > > > > MAX_VELOCITY =                  1.2
> > > > > >
> > > > > > for each axis.
> > > > > >
> > > > > > You should also set
> > > > > >
> > > > > > DEFAULT_ACCELERATION =  20.0
> > > > > > MAX_ACCELERATION =      20.0
> > > > > >
> > > > > > much lower than default.
> > > > > >
> > > > > > HTH
> > > > > >
> > > > > > Ray
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Tuesday 12 February 2002 09:11 am, Shelbyville Design
> > > > & Signworks
> > > > > > wrote:
> > > > > > > specifics:
> > > > > > >
> > > > > > > Gantry router, 8' x 12' work area
> > > > > > > rack & pinion drive
> > > > > > > 23 frame stepper approx 200 oz-in or so, running halfstep,
> > > > > > geared 40:1
> > > > > > > with a 24 d.p. 30T pinion
> > > > > > > (step size is about .00025, way to small for routing, but
> > > > > > this project
> > > > > > > started out with a real small
> > > > > > > budget (wife holds the purse strings and we all know how
> > > > > > that goes) and
> > > > > > > the reducers were already on the machine)
> > > > > > > gantry weight is approx 600 lbs +-
> > > > > > >
> > > > > > > (if you've ever seen a Heath/Esab plasma cutter, it's the
> > > > > > same machine)
> > > > > > >
> > > > > > > running the table on Turbocnc I can achieve 150 IPM easily,
> > > > > > > but no contstant contour.
> > > > > > >
> > > > > > > I am going to be removing the gear reduction and going
> > > > to gecko 201
> > > > > > > drives, with a 550 oz-in
> > > > > > > motor on the x axis, dirctly driving a 20T pinion, it will
> > > > > > increase the
> > > > > > > step size to approx .0015, this should
> > > > > > > help with the speed, but the biggest problem is
> > > > inconsistent step
> > > > > > > pulses (scope confirms this)
> > > > > > > with this much inertia, it doesn't take alot to throw
> > > > things out of
> > > > > > > whack.
> > > > > > >
> > > > > > >
> > > > > > > erie patsellis
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > Shelbyville Design & Signworks
> > > > > > > 1309 W. S. 7th Street
> > > > > > > Shelbyville, IL 62565
> > > > > > > (217)774-4444
> > > > > > > visit our website at http://216.43.233.169
> > > > > > >
> > > > > > > ----- Original Message -----
> > > > > > > From: "Hugh Currin" <currinh-at-oit.edu>
> > > > > > > To: "Multiple recipients of list" <emc-at-nist.gov>
> > > > > > > Sent: Monday, February 11, 2002 6:23 PM
> > > > > > > Subject: Re: Freqmod problems
> > > > > > >
> > > > > > > > Erie:
> > > > > > > >
> > > > > > > > What type of machine are you using?  i.e. size steppers
> > > > > > and drivers.
> > > > > > > > I'll
> > > > > > >
> > > > > > > bet
> > > > > > >
> > > > > > > > someone on the list has an *.ini file they'd be willing
> > > > > > to send you
> > > > > > > > as a starting point.
> > > > > > > >
> > > > > > > > I have freqmod working with 760 oz-in motors, Camtronic
> > > > > > drivers half
> > > > > > > > stepping, 2:1 reduction timing belts on a bridgeport
> > > > kneemill size
> > > > > > >
> > > > > > > machine.
> > > > > > >
> > > > > > > > Does this come close? It isn't optimum but may be a
> > > > > > starting point.
> > > > > > > > :-)
> > > > > > > >
> > > > > > > > Hugh Currin
> > > > > > > > Klamath Falls, OR
> > > > > > > >
> > > > > > > > On Monday 11 February 2002 04:03 pm, you wrote:
> > > > > > > > > am I the only one who has unsuccessfully tried to use
> > > > > > > > > freqmod with steppers? no amount (I've spent at least 8
> > > > > > > > > hours on this so far) of pid tuning seems to
> > > > eliminate following
> > > > > > > > > errors. Steppermod.o for now.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > erie patsellis
> > > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > >
> >
> >
>
>
>