[Emc-users] Backlash comp, does it follow sign of encoder (input) scale?

[Tom Easterday]

Should the backlash parameter in the .ini file be positive or have the sign of the input_scale (aka encoder scale)? I don't see anything in the docs regarding this but did find a brief forum post that suggested it should have the same sign. I haven't tested it, but am wondering if someone knows...? If it should follow encoder scale this should be documented methinks...and since it is now ambiguous should probably be documented regardless.
-Tom

[Jon Elson]

Post by Tom Easterday
Should the backlash parameter in the .ini file be positive or have the sign of the input_scale (aka encoder scale)? I don't see anything in the docs regarding this but did find a brief forum post that suggested it should have the same sign. I haven't tested it, but am wondering if someone knows...? If it should follow encoder scale this should be documented methinks...and since it is now ambiguous should probably be documented regardless.

I believe it should always be a positive number.

Jon

[Gene Heskett]

Post by Jon Elson

Post by Tom Easterday
Should the backlash parameter in the .ini file be positive or have
the sign of the input_scale (aka encoder scale)? I don't see
anything in the docs regarding this but did find a brief forum post
that suggested it should have the same sign. I haven't tested it,
but am wondering if someone knows...? If it should follow encoder
scale this should be documented methinks...and since it is now
ambiguous should probably be documented regardless.

I believe it should always be a positive number.
Jon

That has always been my experience too, but I have never changed the sign
of an encoders 'scale' since my use has been restricted to spindle
controls except for my handwheels on this Sheldon lathe conversion.

The only thing I've found "odd" feeling is that because its all setup
assuming a positive thread on the x screw, the encoder crank then turns
ccw to advance the cutter into the work. The natural tendency is an
acclimated thing because most x drive screws are left hand threaded. So
at first it felt backwards,

But I am getting used to it because the markings on the dial are such
that a ccw turn from zero to say 89, if the dial scale is gain set
at .001" per "click" and its currently touched off to the actual
measured diameter, which is lets say 11 thou from the targeted size, its
as easy to turn it 11 thou in either direction to get to the desired
size, so I pretend its a right hand threaded screw, and turn it 11 thou
ccw, or to .x89 on the dial. OTOH, I've been accused of being odd most
of my life. ;-) I think it would be a one character change in the right
place in the .hal file to reverse that.

BTW, I got another of those sub confirmation requests for the -devel list
today from srcforge. Anybody else?

Cheers, Gene Heskett

[Todd Zuercher]

----- Original Message -----

Sent: Thursday, June 29, 2017 10:36:04 PM
Subject: Re: [Emc-users] Backlash comp, does it follow sign of encoder (input) scale?
BTW, I got another of those sub confirmation requests for the -devel
list
today from srcforge. Anybody else?
Cheers, Gene Heskett

Gene,

I have not, but I went through with the confirmation the 1st time.
I think it was legit.

[Stuart Stevenson]

Gentlemen,

Here is the comp file from the B axis of the blue cincinnati at MPM.
-32.0 .041 .045
-31.0 .039 .042
-30.0 .037 .040
-29.0 .037 .038
-28.0 .035 .035
-27.0 .035 .039
-26.0 .034 .036
-25.0 .034 .034
-24.0 .030 .032
-23.0 .029 .031
-22.0 .027 .029
-21.0 .028 .029
-20.0 .025 .026
-19.0 .026 .027
-18.0 .024 .026
-17.0 .024 .025
-16.0 .023 .025
-15.0 .023 .023
-14.0 .021 .023
-13.0 .023 .026
-12.0 .021 .023
-11.0 .021 .022
-10.0 .021 .023
-9.0 .021 .021
-8.0 .017 .020
-7.0 .018 .019
-6.0 .015 .017
-5.0 .016 .017
-4.0 .015 .015
-3.0 .015 .015
-2.0 .008 .013
-1.0 .009 .012
0.0 .018 0
1.0 .020 -.002
2.0 .020 -.002
3.0 .019 -.004
4.0 .017 -.006
5.0 .016 -.009
6.0 .013 -.011
7.0 .012 -.014
8.0 .010 -.016
9.0 .007 -.017
10.0 .003 -.019
11.0 .004 -.020
12.0 .001 -.025
13.0 0 -.024
14.0 -.005 -.024
15.0 -.005 -.027
16.0 -.008 -.030
17.0 -.009 -.031
18.0 -.009 -.030
19.0 -.009 -.030
20.0 -.012 -.031
21.0 -.013 -.032
22.0 -.014 -.030

I started the compensation two degrees past the travel on both ends so the
compensation would be applied (my intention anyway) for the entire travel
range.
I comped it every 1 degree. The increment could have been more or less I
just chose 1 degree and liked the results.

As you can see the numbers are positive and negative but I don't remember
how the sign affected the compensation.

Easy to figure out. Comp the machine and check it. If the error is twice
what you were trying to fix then you have the sign wrong. :)

HTH
have fun
Stuart

Post by Todd Zuercher
----- Original Message -----

Sent: Thursday, June 29, 2017 10:36:04 PM
Subject: Re: [Emc-users] Backlash comp, does it follow sign of encoder

(input) scale?

BTW, I got another of those sub confirmation requests for the -devel list
today from srcforge. Anybody else?
Cheers, Gene Heskett

Gene,
I have not, but I went through with the confirmation the 1st time.
I think it was legit.
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Emc-users mailing list
https://lists.sourceforge.net/lists/listinfo/emc-users

[Mark]

Post by Stuart Stevenson
Gentlemen,
Here is the comp file from the B axis of the blue cincinnati at MPM.
-32.0 .041 .045

<snippage>

Post by Stuart Stevenson
22.0 -.014 -.030
I started the compensation two degrees past the travel on both ends so the
compensation would be applied (my intention anyway) for the entire travel
range.
I comped it every 1 degree. The increment could have been more or less I
just chose 1 degree and liked the results.
As you can see the numbers are positive and negative but I don't remember
how the sign affected the compensation.
Easy to figure out. Comp the machine and check it. If the error is twice
what you were trying to fix then you have the sign wrong. :)
HTH
have fun
Stuart

Wonder if backlash comp could be used on the Z axis on my saw beveler to
compensate for a long skinny table that is off by a few thou here and
there over the entire length of the table?

Mark

[Stuart Stevenson]

Backlash compensation will not move the orthogonal axes in reference to the
compensated axis motion.
If you are having length problems during the motion of the Z axis then comp
could help.

My file shows the bidirectional compensation of LinuxCNC.

Post by Mark

Post by Stuart Stevenson
Gentlemen,
Here is the comp file from the B axis of the blue cincinnati at MPM.
-32.0 .041 .045

<snippage>

Post by Stuart Stevenson
22.0 -.014 -.030
I started the compensation two degrees past the travel on both ends so the
compensation would be applied (my intention anyway) for the entire travel
range.
I comped it every 1 degree. The increment could have been more or less I
just chose 1 degree and liked the results.
As you can see the numbers are positive and negative but I don't remember
how the sign affected the compensation.
Easy to figure out. Comp the machine and check it. If the error is twice
what you were trying to fix then you have the sign wrong. :)
HTH
have fun
Stuart

Wonder if backlash comp could be used on the Z axis on my saw beveler to
compensate for a long skinny table that is off by a few thou here and there
over the entire length of the table?
Mark
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Emc-users mailing list
https://lists.sourceforge.net/lists/listinfo/emc-users

[Stuart Stevenson]

probably should have called it leadscrew comp rather than backlash

Post by Stuart Stevenson
Backlash compensation will not move the orthogonal axes in reference to
the compensated axis motion.
If you are having length problems during the motion of the Z axis then
comp could help.
My file shows the bidirectional compensation of LinuxCNC.

Post by Mark

Post by Stuart Stevenson
Gentlemen,
Here is the comp file from the B axis of the blue cincinnati at MPM.
-32.0 .041 .045

<snippage>

Post by Stuart Stevenson
22.0 -.014 -.030
I started the compensation two degrees past the travel on both ends so the
compensation would be applied (my intention anyway) for the entire travel
range.
I comped it every 1 degree. The increment could have been more or less I
just chose 1 degree and liked the results.
As you can see the numbers are positive and negative but I don't remember
how the sign affected the compensation.
Easy to figure out. Comp the machine and check it. If the error is twice
what you were trying to fix then you have the sign wrong. :)
HTH
have fun
Stuart

Wonder if backlash comp could be used on the Z axis on my saw beveler to
compensate for a long skinny table that is off by a few thou here and there
over the entire length of the table?
Mark
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Emc-users mailing list
https://lists.sourceforge.net/lists/listinfo/emc-users

--
Addressee is the intended audience.
If you are not the addressee then my consent is not given for you to read
this email furthermore it is my wish you would close this without saving or
reading, and cease and desist from saving or opening my private
correspondence.
Thank you for honoring my wish.

[Mark]

Post by Stuart Stevenson
Backlash compensation will not move the orthogonal axes in reference to the
compensated axis motion.
If you are having length problems during the motion of the Z axis then comp
could help.
My file shows the bidirectional compensation of LinuxCNC.

Okay, so not really applicable to my case then. Back to the drawing board.

Mark

[John Kasunich]

Screw error comp isn't really applicable, but you can do something similar in HAL.
You want (if I understand correctly) a compensation value that applies a minor tweak
to Z as X moves, right?

You could use the lincurve HAL component to calculate the compensation amount
and the offset component to apply it to the Z axis.

http://linuxcnc.org/docs/html/man/man9/lincurve.9.html
http://linuxcnc.org/docs/html/man/man9/offset.9.html

Post by Mark

Post by Stuart Stevenson
Backlash compensation will not move the orthogonal axes in reference to the
compensated axis motion.
If you are having length problems during the motion of the Z axis then comp
could help.
My file shows the bidirectional compensation of LinuxCNC.

Okay, so not really applicable to my case then. Back to the drawing board.
Mark
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Emc-users mailing list
https://lists.sourceforge.net/lists/listinfo/emc-users

[Mark]

Post by John Kasunich
Screw error comp isn't really applicable, but you can do something similar in HAL.
You want (if I understand correctly) a compensation value that applies a minor tweak
to Z as X moves, right?

Yes, that's exactly what I'm looking for. Minor dips and humps in the
table that I'd like to have some software correction for. I only have
the two axes, X and Z on the machine. It's a saw beveler that cuts a
tapered triangular bamboo strip.

Post by John Kasunich
You could use the lincurve HAL component to calculate the compensation amount
and the offset component to apply it to the Z axis.
http://linuxcnc.org/docs/html/man/man9/lincurve.9.html
http://linuxcnc.org/docs/html/man/man9/offset.9.html

Thanks. I'll look into that. Any links you know of that show how I
might implement that?

Mark

[Les Newell]

Yes, that's exactly what I'm looking for. Minor dips and humps in the
table that I'd like to have some software correction for. I only have
the two axes, X and Z on the machine. It's a saw beveler that cuts a
tapered triangular bamboo strip.

Wait a day or so. I am in the middle of writing kinematics that will do
exactly what you want to do. In my case it is to compensate for a worn
bed on my lathe. Roughly how many points are you likely to need in your
compensation table?

Les

[Mark]

Post by Les Newell

Yes, that's exactly what I'm looking for. Minor dips and humps in the
table that I'd like to have some software correction for. I only
have the two axes, X and Z on the machine. It's a saw beveler that
cuts a tapered triangular bamboo strip.

Wait a day or so. I am in the middle of writing kinematics that will
do exactly what you want to do. In my case it is to compensate for a
worn bed on my lathe. Roughly how many points are you likely to need
in your compensation table?
Les

I've got a bit of a weird setup in my beveler. Not really sure if
replacing the current mode of operations to run a different set of kins
will work without having to redo all the other stuff that runs currently.

Mark

[Les Newell]

Post by Mark
I've got a bit of a weird setup in my beveler. Not really sure if
replacing the current mode of operations to run a different set of
kins will work without having to redo all the other stuff that runs
currently.

How many axes are you using?

Les

[Mark]

Post by Les Newell

Post by Mark
I've got a bit of a weird setup in my beveler. Not really sure if
replacing the current mode of operations to run a different set of
kins will work without having to redo all the other stuff that runs
currently.

How many axes are you using?
Les

Les,

Two axes, X and Z. I've got the Z set up kind of opposite the normal
way, and use the top of the strip being cut offset from the table top as
the point where the cutter depth is computed from. It was a bit of a
pain in the arse to get it set that way and I'd rather not have to
reinvent it again. ;-)

Mark

[Les Newell]

Post by Mark
Two axes, X and Z. I've got the Z set up kind of opposite the normal
way, and use the top of the strip being cut offset from the table top
as the point where the cutter depth is computed from. It was a bit of
a pain in the arse to get it set that way and I'd rather not have to
reinvent it again. ;-)
Mark

I see. So basically you have the axes swapped compared to how they would
be in a lathe.

Les

[Mark]

Post by Les Newell

Post by Mark
Two axes, X and Z. I've got the Z set up kind of opposite the normal
way, and use the top of the strip being cut offset from the table top
as the point where the cutter depth is computed from. It was a bit of
a pain in the arse to get it set that way and I'd rather not have to
reinvent it again. ;-)
Mark

I see. So basically you have the axes swapped compared to how they
would be in a lathe.
Les

It's setup was originally based on a mill. Z goes up and down, X
traverses the length of the table.

Mark