On 4/22/26 17:09, andrew beck wrote:
hey guys
finally made a thread on the forum
lots of photos and this youtube video
for those that are interested
cheers
https://youtu.be/AmQGh_Cwff0?si=ZIihICf10Gto2px7
https://forum.linuxcnc.org/show-your-stuff/58689-large-fil-cnc-machine-retrofit-12-ton-larger-maching-running-linuxcnc#345849
andrew
On Thu, Apr 23, 2026 at 7:37 AM andrew beck <[email protected]>
wrote:
Hey guys
thanks for the replies
the homing is a interesting issue
I had a bit of a play and made a copy of my config that i hacked.
What i found was that if i combined my original 2 PID loops into 1 loop
and used the scale for position and used the motor velocity on encoder for
velocity all in one PID then homing to index enable worked perfectly
but doing it that way the motor tuning was not as good. backlash is a
constant of 0.1mm and the motor just jumped across that backlash whenever
it moved. in practice that meant that the absolute scale position hovered
by +-0.01 or so.
this was not bad and totally usable. If i had done this option first i
would have called it good and just used it. but the movement was much
louder and less smooth than the two PID loops which get run through a sum
component. As the original two PID loops was much much smoother, it was
perfect and once i come to a stop the drive rapids to within 0.01 or 0.02
and then creeps the last 0.02mm (using the integral value which takes about
a second) and ends up in the right place within 0.005mm or less.
basically perfect every time with no dithering.
It is now my most accurate machine. and really nice
also the homing when homing I home all axis at the same time. X and Y get
homed to index enable like normal.
and the Z value i trick the home switch by checking the joint state. How
this works is the Joint2 switch state goes form 7 when searching for switch
the first time to 12 when latching on to switch slowly the second time
so in classic ladder i set up two compare functions that basically choose
whether to listen to the main home switch or the z axis motor encoder Z
index pulse. (this is different to the index enable pulse and this is
possible to miss when only checking in software not hardware. but at a
slow speed i haven't found any issues yet, and when testing with a dial
indicator i can't see any changes between homing. I can rehome linuxcnc up
the top of the travel beside the home switch and then rapid down to the
table which is about 1m away and the dial indicator is definitely within
0.01 each time. i can't really see any movement on it.
So pretty happy all in all would like to fix the index enable issue in the
future but for now it's working fine.
this is a 40 year old machine and weighs 12 ton. and when i got it the
backlash was 0.40mm. I replaced all the thrust bearings and changed from
reduction gearbox to a nice mazak flexible coupling and direct drive The
timing belt was T8 timing belt so not even HTD profile which has more
backlash which didn't help. The FIL factory actually got the location of
the Z axis 2nd ballscrew mount out by 2.4mm which i was shocked by. i
double checked measuring against the slideways at the top and bottom of the
screw and yeah it was out 2.4mm from factory! so that is why all the
thrust bearings were stuffed.
anyway just about to get it running into production. this machine will be
running like 16 hrs a day shortly probably.
cheers
Andrew
On Thu, 23 Apr 2026, 5:11 am gene heskett, <gheskett@shentelnet> wrote:
On 4/22/26 10:44, Dave Engvall wrote:
Hi Andrew et al
I have a cinci contourmaster set up with US made glass scales and a
rotary encoder on the ballscrew. I’m in transition to stepper-servos but
progressing slowly.
When I started this process I did a manual check with the hand crank vs
the X axis scale. Proceeding in the positive direction I manually moved in
1” increments, ie 4 turns of the hand wheel. Even in this early 1960’s
machine the glass scale was dead on. Since I’m protective of the US scale I
removed it and installed a Chinese scale.
Both scales have a reference channel spaced with markers 50 mm apart I
suspect the US made scale has guard ticks either side of the reference
mark. I know the Chinese one's do.
Homing procedure was simple. I installed a laser diode and a tab with a
hole in it at the end of the axis and manually homed to that before kicking
off a home of the machine. Z was done by manually moving Z to the upper
stop. If I positioned the tab and or Z stop correctly it never had to move
more than ~ 3 mm to index. All of this used the normal homing procedure.
Scan past the index and capture the position of index then move as
necessary to home. It is my impression that homing should be done in that
manner and not dual channel but after homing position seeking uses both
signals but I have no data to back this up! However, given the nature of
linear scales one might do just as well to use linear scale for position
and shaft encoder on the motor for velocity. The system on my Mazak used a
resolver on the servo motor geared (?) 7:1 and 1.6:1 to the ball screw.
Those numbers are from memory and a long time ago i.e a couple of years
before we did the Mazak at Galesburg.
When I get the machine running with stepper-servos I will probably try
using the R channel for homing. Thanks for listening. Rant OFF.
Dave
I have a much simpler setup, the only fully stepper/servo is my sheldon
lathe which is all stepper/servo, or about 1/2 of my 4 axis 6040 mill.
I have motors and drivers to do the same for my ball screw converted
go704. configurable as 3 or 4 axis. Home is the closing of the $2
switch on each axis. The 6040 is the odd man out, There, z & b are
stepper/servo's where xy are just steppers. I can have a 1 second power
failure in the middle of a threading operation, remove the tool, rehome
the machine, I have notches in my r8's to replace the lack of tooing
keys go thru the rehome, replace the tool and hit r, all in the middle
of a 50 tpi threading opertion. That is plenty accurate enough for the
girls I might go with. The only none hard switch is the inductive home
switches on the 4th axis for z & a or b. My biggest problem is the.3mm
back lash in g7O4 z if the gibs are too tight. & I forget to oil it.
On Apr 18, 2026, at 10:49 PM, andrew beck <[email protected]>
wrote:
hey guys
just reporting back here with my success with the dual feedback
i have 0.1mm backlash still in the ballscrew or somewhere
and i have a magnetic scale on the machine also
with a encoder on the end of the ballscrew
i copied the examples on online of how they did that massive mill and
its
working awesome
the mill rapids down or feeds down to 0.02 to 0.05 and then the I value
from the second PID takes over and moves the last tiny bit
which takes a second or so
and it stops right where it should everytime within 0.01mm the scale
says
it's perfect to 0.003 on the screen. and no dithering
next issue thought is homing with index
either to the motor encoder or scale z pulse
i have my x and y axis homing perfectly with encoder index
but i can't get linuxcnc to home this z axis to index. homing without
index works fine though.
i think it is because linuxcnc internals throws a hissy fit because i
have
a hacked up PID system with two forms of feedback getting summed
together
any ideas would be great
cheers
Andrew
On Mon, Jan 5, 2026 at 11:04 AM gene heskett <[email protected]>
wrote:
On 1/4/26 01:36, gene heskett wrote:
On 1/4/26 00:23, andrew beck wrote:
hey stuart
yeah the ballscrew is connected both ends. i think one end has the
big
angular contact bearings etc
and the other only has deep groove ends probably (its much smaller
than the
driving end)
on my other two cnc mills they have angular contact bearings both
ends and
the screws are stretched between them slightly with a tiny bit of
preload
Probably should clarify here. The tension isn't meant to stretch the
screw,
but to pull it tight enough there's no lengthwise backlash under
normal
working loads despite a few microns of wear. On my sheldon, the z
screw
has only a sliding ball bearing on the right end, so the double
bearing on
the left controls the end play. After a decade of light use, I'm
still
under 2
thou of play acc a .0001 dial.
I think I'm the lucky amature, which I am. Now I'm 91, and mainly doing
my own things, which I cannot show you, JK won't allow it on machines
that didn't cost an arm & a couple legs. But I'm still having fun.
Flowers to all that have helped me for the last 25+ years as you taught
me about LinuxCNC.
hey guys
i was reading the link that sam sent two and it looks like the two
Pid
loops are running side by side and then the outputs are added
together
before going to the servo drive
would i be correct in thinking that they are using the "sum"
component in
linuxcnc to do that?
if so this is getting pretty simple
cheers
andrew
On Fri, Jan 2, 2026 at 3:07 PM Stuart Stevenson <[email protected]>
wrote:
HI,
I don't think the LinuxCNC control scheme is cascaded or nested.
I
would
call it parallel.
The only backlash problem I can imagine is if the backlash
causes the
motor to dither trying to maintain position because cutting forces
move the
axis enough to cause the control to try to correct the axis
position.
Is the ball screw supported on each end? If it is, then, is it
possible
the screw has some stretch preload.
regards
Stuart
On Thu, Jan 1, 2026 at 4:10 PM andrew beck <
[email protected]>
wrote:
hey guys
sam i just looked at that link you sent
that looks pretty cool
only thing is i don't know is if it will work to reduce the
backlash
issue
i am getting. I have about 0.10mm of backlash on each axis
half of
that
is from the thrust bearings which is fixable with new bearings
the rest is from the ball nut i think and i am not willing to fix
that
Those tiny ball screws stuart sold me 25 years ago were just a hair
sloppy.
I found the balls .001 bigger on ebay and restuffed the nuts. One of
them
is in the Y of my sheldon, and has about a thou of backlash today I
had
to make a cage for the nut with teeny grub screws to stop the nuts
end-
play in the cage. So oversized balls are one possibility. Ticklish
work tho.
but i can only try it.
does anyone want to have a go with making a config like they
did? one
axis
would be best I am struggling to get my head around the different
layers.
i probably need to write it all out or use a diagram or something.
regardless i am sure i will work it out. just might take some
brain
cells
cheers
Andrew
On Fri, Jan 2, 2026 at 4:54 AM Sam Sokolik <[email protected]>
wrote:
This was also John k.. I would trust his solution. He is the
one
that
came up with hal in linuxcnc... (Among other things).
On Thu, Jan 1, 2026, 9:49 AM andy pugh <[email protected]>
wrote:
On Thu, 1 Jan 2026 at 15:41, andy pugh <[email protected]>
wrote:
On Wed, 31 Dec 2025 at 22:14, andrew beck <
[email protected]>
wrote:
just looking for a config example if possible for a mesa card
from
someone
that has this nested PID loop setup.
This may not be as effective as you would hope. Cascaded PID
loops...
Actually, what was done with the MPM jigmill wasn't a cascaded
PID.
What was done there was to run a P + FF control based on the
motor
encoder and an I control based on the linear scales.
Then the two controllers were summed together to create a PID
controller, but with I using a different feedback source than P
and
D.
--
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses,
daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1912
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
Cheers, Gene Heskett, CET.
--
"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
Don't poison our oceans, interdict drugs at the src.
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
Cheers, Gene Heskett, CET.
--
"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
Don't poison our oceans, interdict drugs at the src.
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
Cheers, Gene Heskett, CET.
--
"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
Don't poison our oceans, interdict drugs at the src.
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
