On 7/4/23 03:15, John Dammeyer wrote:
Hi Gene,
I suspect the problem is a bit strange in my case.
I can run the knee at 150 ipm. The tapping in this case happens at 10 ipm.
The I3 parameter means it returns at 30 IPM. I can jog up and down at say 60
IPM and never get a following error for Joint 2 which is the Z axis.
Since this showed up during the G33.1 I'm going to guess that the deceleration
of the Z axis compared to the spindle doesn't match what is expected. Since Z
can move with G0 motion (150 ipm) without following errors the problem has
something to do with the timing of the I3.
I changed the value to S100 so now tapping is at 5 IPM. It does not create a
following error with I6 which is 30 IP the same failure speed when it's 10 IPM
and I3 for 30 IPM.
That suggests again that it's not as obvious as it might appear. Note in my
sample below I also move the Z to 0.5 with a G0 . That's again 150 IPM. No
following error. Therefore given the above experiments the Z axis motion for
tracking the S200 spindle at I3 has an issue yet doesn't at S100 and I6 which
is the exact same speed.
The problem is under the covers somewhere and the Joint 2 following error isn't
totally accurate but is what happens later.
John
That reads like the spindle stop time is overly long to me.
Have you timed the m3/m4 turnaround of the spindle? Mine is a brushed dc
rated at 1 hp. but at 90 volts, 9.7 amps. The psu can do 25 amps,
tripping the 20 amp service breaker if it wasn't for the currant limit
set in Jom Elsons pwm-servo amp, but for the reversal I have some hal
trickery in the reversal that doesn't just shut the motor down and wait
for it to coast to a stop, but blocks the reverse dir until its
essentially stopped by ramping the speed down rapidly, letting the PICO
pwm-servo, which it a full 4 quadrant servo, pull the rotation energy
back out of the motor, running that 126 volt supply up to about 170
volts, actually above the rating of the caps in that supply. By using
the motor as a generator, the stop can be accomplished from around 10g's
at the motor is less than 150 ms. Then the fact that an encoder edge
hasn't arrived recently, the actual dir change is the allowed to get to
the PID, and the speed is ramped back up to the set revs in the other
direction, which because of the extra voltage in the psu's caps, can put
that motor back at speed using that energy in the caps. And its been
doing that for over 5 years. The actual time to achieve that reversal,
at 3g's spindle speed is around 400 ms, correspondingly faster at lower
revs. And my shop lights don't blink, the reversal power bill is 99%
paid by the 4 quadrant controller.
I'm doing much the same thing on my Sheldon lathe once I had installed
clamps to keep from unscrewing the chuck, but because that 8" chuck is
nearly 40 lbs, its hard to stop. I've some hal code there that measures
the over-travel so I could get an idea of how close I could come to the
bottom of a blind hole w/o breaking the tap. At 100 rpm, an m4 stop that
triggers that measurement, displayed on the pyvcp panel, is .24 to .25
of a revolution. A little math will convert that into actual distance
the tap will overshoot, which if I write the gcode correctly will be
subtracted from the depth of that tap preventing the breakage. And
because the 3 phase synthed controller vfd can be a 4 quadrant control,
that too works by re-using the stored energy gained from the stop to
re-accel the motor in the other direction. And I can do that once a
second without blinking the shop lights. For several minutes. In this
case however, there is no PID being used. The idea is essentially the
same but a spinx1 is making the analog speed signal for the vfd. The
encoder is also a much lower resolution, its shop made, ATS-667's hall
effects watching the teeth of the 60 tooth bull gear on the spindle.
> -----Original Message-----
From: gene heskett [mailto:ghesk...@shentel.net]
Sent: July 3, 2023 5:49 PM
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] G33.1 error then g-code file parsed
On 7/3/23 19:32, Nicklas SB Karlsson wrote:
Which version of Linuxcnc did you try it on?
Den 2023-06-28 kl. 19:52, skrev John Dammeyer:
So this morning I whipped up a small g-code program to test the
tapping.� I
started a fresh LinuxCNC so that most of the parameters were default.
Did� a HOME command and then moved to the logical home position with a
G0 X0
Y0 G0 Z0.
Then loaded this program.
G17 G20 G40 G90
G1 X0 Y0 F20
G1 Z0
S200 M3
G33.1 Z-0.7 K0.05 I3.0
g0 Z0.5
M5
M2
%
The program loaded without errors.
What's interesting is that this time although the knee went back down
at 3x
the speed after tapping the end result was a "Joint 2 following error"
Remove the I3.0 and the following error does not happen.� Works with I1.5
but not I2.0.
I suspect it should throw a following error at I3 in any event. The docs
are quite concise:
"I - optional spindle speed multiplier for faster return move"
And your machine is not able to follow that great a speedup. It may
also be suffering from excessive backlash as that move also takes time
that throws following errors quite easily for backlash comps over a thou
in my experience here.
YMMV of course. Since that is z motion, and heads or knees are heavy,
there is effectively no backlash as long as the withdrawal speed is
below gravity's acceleration AND the spindle has the speed reserve to
keep up.
I'd try -I1.5 as a half the 3 compromise and gradually increase/decrease
it to see where the limit might be on your machine. Or not use the I
option. Don't forget that 3x the spindle speed may be beyond the
spindles capability. Even a slightly out of tune PID can affect how well
the -I option works. I haven't used it myself, but I'd assume a -I0.5,
would slow the return on a moving head machine. In case the weight of
the head is almost more that the motor can lift.
Some experimentation to find the limits of your machine is in order, but
I believe I'd tap air, its cheaper than broken taps. ;o)>
Not sure why that is.� It worked the other day, even from the MDI
entry with
I3.0.� And now it also generates the following error from the MDI.
I'm going to guess either the machine is cold and a bit stiff or there
were
some other parameters the other day set that prevented the fault.
In either case, I do not get a fault loading the program.
John
-----Original Message-----
From: Nicklas SB Karlsson [mailto:n...@nksb.eu]
Sent: June 28, 2023 6:30 AM
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] G33.1 error then g-code file parsed
Have used the latest origin/master not more than a few weeks old but
have not checked against older version. Will do this sooner or later,
hopefully within a week.
Den 2023-06-26 kl. 07:29, skrev John Dammeyer:
OK.
I'll try that for you.
I'm wondering if perhaps spindle acceleration time or some other HAL
parameter gets flagged with an incorrect error message.�� I've done
power
tapping from .ngc files without issue but I've not used the 'I'
parameter.
John
-----Original Message-----
From: Nicklas SB Karlsson [mailto:n...@nksb.eu]
Sent: June 25, 2023 10:09 PM
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] G33.1 error then g-code file parsed
Run from MDI then there is no error message for me either. Put the same
rows in a .ngc file adding M2 and bottom there is an interpreter error
message then loading the file and display not updated. Looking at bottom
the program is however actually loaded and seems to work OK.
Nicklas Karlsson
Den 2023-06-26 kl. 01:26, skrev John Dammeyer:
I ran the following commands from the MDI without issue from AXIS 2.8.1
with
MESA 7i92H controlling an AC Servo motor (step/dir) for the spindle.
G1 X0 Y0 F20
G1 Z0
S200 M3
G33.1 Z-0.7 K0.05 I3.0
M5
Interesting watching the spindle RPM indicator, and you can hear it too
of
course.
Goes down to tap the 0.05" pitch (20TPI thread) at 200RPM
Reverses and goes up at 600 RPM.
Without the I3.0 it goes up at 200 RPM.
John
-----Original Message-----
From: Nicklas SB Karlsson [mailto:n...@nksb.eu]
Sent: June 25, 2023 9:07 AM
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] G33.1 error then g-code file parsed
There is positioning move XYZ in real program but happened to remove
them then I should make a small test case.
Adding positioning move to test case make no difference. Linuxcnc
report
a file interpretation error and do not update display but program seems
to work supposed to.
Nicklas Karlsson
Den 2023-06-23 kl. 16:20, skrev Chris Radek:
I would add positioning moves (positioning all of XYZ) before the
G33.1 because otherwise the tapped hole can be anywhere - the
program is indeterminate.� This sure might mess up any attempt by
your GUI to check the program or generate a preview.
Chris
On Wed, Jun 14, 2023 at 09:14:20PM +0200, Nicklas SB Karlsson wrote:
I put the lines below into a file:
���� ?? M3 S100
���� ?? G33.1 Z-30.474 K0.8 I3.000
���� ?? M5
���� ?? M2
Then I read into Linuxcnc I get error message:
���� ?? parse_file interp_error
Removing the line with G33.1 then no error message so it is something
with this row. Program do however execute as expected with G33.1 line
even though there is an error message so no real problem. Also
execute
without an error message if run manually in MDI mode. Use
origin/master
last commit Mon May 8 16:10:03 2023 +0200
404aa407f136ce91a3e6bf911c7bda54011a74e9 Anybody else had similar
problem?
Nicklas Karlsson
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
.
Cheers, Gene Heskett.
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
Genes Web page <http://geneslinuxbox.net:6309/>
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
