Re: [Emc-users] How about the core[X,Y] drive?

[Steve Blackmore]

On Mon, 15 Apr 2013 20:17:29 -0400, you wrote:

>But on my small lathe, how would one go about putting tool offsets into the 
>tool.tbl that mean something when in changing a tool, the stickout its 
>clamped up at is effectively completely random give or take 1 cm or so?
>
>To me, it makes 100x as much sense to mount the tool such that it can reach 
>its cuts ok, drop the gauge on the ways and simply re-home the machine.  
>FWIW, I have an .ini special for boring bars.  It searches outward to find 
>the contact, but fudges some on the z because most boring bars on this 
>gauge will only contact on the end of the bar, whereas the cutting tip is 
>typically 1/8 to 1/4" shorter than the end of the bar.  And again, the 
>amount of stickout is entirely arbitrary.  In fact I have one home made bar 
>that I can stick out about 8".  Totally useless at that much stickout 
>unless I'm boring room temp butter, but it is doable. :)

Gene - IF you use quick change holders and tool post, once you have
manually set tools, and don't remove them from the holders, you can
touch off using the tool post body or a probe in a holder. As long as
you clean the clamping surfaces good quality ones are remarkably
repeatable.

The downside is the price of holders, but I'll bet you'll need less than
you thought you would (or bought ;)

Steve Blackmore
--

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] KUKA!!

[Gregg Eshelman]

--- On Mon, 4/15/13, a k  wrote:

> From: a k 
> Subject: [Emc-users] KUKA!!
> To: "Enhanced Machine Controller (EMC)" 
> Date: Monday, April 15, 2013, 9:45 PM
> Hi
> here is interesting use of CNC
> http://www.youtube.com/user/KukaRobotGroup

All that high tech and near the end a guy is demonstrating a robot motion by 
waving his hands around. ;-)

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] KUKA!!

[a k]

Hi
here is interesting use of CNC
http://www.youtube.com/user/KukaRobotGroup

I can see that use standard frame of CNC mill with 3 linear and 2
rotational axis is very old.
All rotational robot like KUKA can have more use.

about commerce and EMC2, i think if one will take standard EMC2 and and add
all what is necessary for 8 rotational axis into one disk than that disk
will have market.
One can make cook their own bread or just buy it.
Get disk, build system that for 8 axis, plug in servo motors, build
rotational axis and offer low cost robot.

thank you
aram
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

This is a diagram of what I'd like to do on the torch, inspired by skycam video 
of a NASCAR race.

http://www.partsbyemc.com/pub/pattern-torch.jpg Replace .jpg with .psd for a 
version with pieces you can move around.

Pretty simple mechanically, but on the software end nobody's done much for a 
strictly 2D implementation with 4 motors, especially not for translating G-code 
to that type of motion control.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gene Heskett]

On Monday 15 April 2013 21:05:25 Gregg Eshelman did opine:

> How about a couple of scripts that launch LinuxCNC in various ways,
> including one that nukes the troublesome files first?

Precisely what is running in my mind too, but I need to find my round tuit.  
And my weed eater...

I bought a small LCD monitor ($20) that can hang right over the lathe at a 
flea market over the weekend and it tested out good, but today I was 
putting the finishing touches on that BP rifle I've been working on as I 
broke it at the range yesterday, so its now about 100x stronger at that 
point.

I also bought, new in box with all packaging, a linksys wr54-gs router for 
another of those famous 20 bills from the same way out in the sticks yard 
sale.  I can install DD-WRT on it and have another backup router if this 
one dies.

> Should be able to have some scripts create or copy those files from
> another location so you could have many different setups with a single
> install of the software.

Exactly. I already have a separate setup for use with boring bars.
Separate scripts does seem to be the way to go.
 
> --- On Mon, 4/15/13, Gene Heskett  wrote:
> > From: Gene Heskett 
> > Subject: Re: [Emc-users] How about the core[X,Y] drive?
> > To: emc-users@lists.sourceforge.net
> > Date: Monday, April 15, 2013, 12:28 PM
> > On Monday 15 April 2013 13:22:31
> > 
> > Viesturs Lؤپcis did opine:
> > > 2013/4/15 andy pugh 
> > > 
> > > > On 15 April 2013 11:00, Gregg Eshelman 
> > 
> > wrote:
> > > > > The crossover counters any offset push at one
> > 
> > end of the gantry with
> > 
> > > > > an
> > > > 
> > > > equal pull in the same direction at the opposite
> > 
> > end. The theory page
> > 
> > > > shows how it works.
> > > > 
> > > > I have stared at the diagrams for long enough now
> > 
> > that I have changed
> > 
> > > > my mind several times. I might have to build one
> > 
> > to see.
> > 
> > > > > Does LinuxCNC have H-bot support? CoreXY
> > 
> > should drive identically.
> > 
> > > > Not built-in, but:
> > > > 
> > > > http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcn
> > > > c-qu estions/26302-cnc-avec-systeme-h-bot-ou-corexy
> > > 
> > > Hmm, it is not the first time I see links to that
> > 
> > Russian forum and with
> > 
> > > every time I get stronger and stronger impression that
> > 
> > those LinuxCNC
> > 
> > > users that hang out there is a completely separate
> > 
> > community...
> > 
> > > Anyway, I just cannot figure out, how to set up homing
> > 
> > routine to home
> > 
> > > such a machine to some switches.
> > 
> > Use the sign of the HOME_SEARCH_VELOCITY to control which
> > switch its
> > looking for.آ  Then HOME_OFFSET with a sign can set home
> > to any position.آ 
> > Then in case the home switches are off in a corner someplace
> > and you have
> > to jog the machine to a suitable starting position before
> > invoking the
> > homing operation, run it to where the first switch is
> > available quickly,
> > and use the HOME_SEQUENCE number to put all the home ops in
> > the proper
> > order.
> > 
> > On my lathe, a grep of the .ini for HOME returns this:
> > In [TRAJ]
> > HOME_ALLآ آ آ  = 1 (one home all button in
> > axis)
> > 
> > In x [AXIS_0]
> > HOME = 1.250آ  (where it parks it when done)
> > HOME_IGNORE_LIMITS = YES
> > HOME_USE_INDEX = NOآ  (if servo's, could be useful)
> > HOME_OFFSET = 0.0100 (where it marks it as 0.0 from
> > where its at)
> > HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> > HOME_LATCH_VEL = -.008
> > HOME_FINAL_VEL = -.1
> > HOME_IS_SHAREDآ آ آ  = 1 (one logic signal is
> > input for all homes)
> > HOME_SEQUENCEآ آ آ  = 1 (second one to be done)
> > 
> > 
> > In z [AXIS_2]
> > HOME = 4.250آ  (where it marks it as being when
> > finished)
> > HOME_IGNORE_LIMITS = YES
> > HOME_USE_INDEX = NO
> > HOME_OFFSET = 4.090 ( distance in this case from left face
> > of gauge to face
> > of pcb on right face of gauge)
> > HOME_SEARCH_VEL = -0.08
> > HOME_LATCH_VELآ آ آ  = -0.015
> > HOME_FINAL_VEL = 0.25
> > HOME_IS_SHARED = 1
> > HOME_SEQUENCEآ آ آ  = 0 (first to home)
> > 
> > I place a gauge I made on the ways, backed up against the
> > work piece as
> > chucked, place a cleaned cutting tool next to the pcb copper
> > that is the Z
> > detector on the gauge, and hit the HOME_ALL button,آ  It
> > searches z for
> > contact, sets the home offset such that 0.0z is shaving the
> > rust off the
> > chuck jaws, or if work is chucked, touching the face of the
> > work.آ  Then it
> > backs away a wee bit and searches inward to find the contact
> > for the x
> > home, finds that piece of pcb, sets home to pretty close to
> > the center of
> > the work (I'm still fine tuning that) then parks the tool
> > back out of the
> > way till its needed.
> > 
> > yerp, buzz, yerp, yerp, buzz, yerp, done in about 7
> > seconds.
> > 
> > If the ways and base of my tool are lox clean, repeatable to
> > under .001".آ 
> > That however is a major problem for x as it tilts th

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

I'm thinking diagram 4 with the engagement to the torch arm on the bottom and 
cable cross on top, all mounted to a tip up frame would be easiest to build. 
http://www.corexy.com/theory.html

Here's that no-slip drive cable
http://sdp-si.com/web/html/newprdbelts6.htm

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

Another thing complicating CNC on a pattern torch (like this 
onehttp://www.hellotrade.com/esab-usa/ultra-graph-single-torch-pattern-cutter.html
 ) is the maximum work envelope the arm can reach isn't a quadrilateral.

As the arm end gets close to the column the range of motion is very restricted 
and at maximum reach it describes an arc.

I figure the best compromise would be to define a quadrilateral that the torch 
can reach and just go with it.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

How about a couple of scripts that launch LinuxCNC in various ways, including 
one that nukes the troublesome files first?

Should be able to have some scripts create or copy those files from another 
location so you could have many different setups with a single install of the 
software.

--- On Mon, 4/15/13, Gene Heskett  wrote:

> From: Gene Heskett 
> Subject: Re: [Emc-users] How about the core[X,Y] drive?
> To: emc-users@lists.sourceforge.net
> Date: Monday, April 15, 2013, 12:28 PM
> On Monday 15 April 2013 13:22:31
> Viesturs Lācis did opine:
> 
> > 2013/4/15 andy pugh 
> > 
> > > On 15 April 2013 11:00, Gregg Eshelman 
> wrote:
> > > > The crossover counters any offset push at one
> end of the gantry with
> > > > an
> > > 
> > > equal pull in the same direction at the opposite
> end. The theory page
> > > shows how it works.
> > > 
> > > I have stared at the diagrams for long enough now
> that I have changed
> > > my mind several times. I might have to build one
> to see.
> > > 
> > > > Does LinuxCNC have H-bot support? CoreXY
> should drive identically.
> > > 
> > > Not built-in, but:
> > > 
> > > http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-qu
> > > estions/26302-cnc-avec-systeme-h-bot-ou-corexy
> > 
> > Hmm, it is not the first time I see links to that
> Russian forum and with
> > every time I get stronger and stronger impression that
> those LinuxCNC
> > users that hang out there is a completely separate
> community...
> > 
> > Anyway, I just cannot figure out, how to set up homing
> routine to home
> > such a machine to some switches.
> 
> Use the sign of the HOME_SEARCH_VELOCITY to control which
> switch its 
> looking for.  Then HOME_OFFSET with a sign can set home
> to any position.  
> Then in case the home switches are off in a corner someplace
> and you have 
> to jog the machine to a suitable starting position before
> invoking the 
> homing operation, run it to where the first switch is
> available quickly, 
> and use the HOME_SEQUENCE number to put all the home ops in
> the proper 
> order.
> 
> On my lathe, a grep of the .ini for HOME returns this:
> In [TRAJ]
> HOME_ALL    = 1 (one home all button in
> axis)
> 
> In x [AXIS_0]
> HOME = 1.250  (where it parks it when done)
> HOME_IGNORE_LIMITS = YES
> HOME_USE_INDEX = NO  (if servo's, could be useful)
> HOME_OFFSET = 0.0100 (where it marks it as 0.0 from
> where its at)
> HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> HOME_LATCH_VEL = -.008
> HOME_FINAL_VEL = -.1
> HOME_IS_SHARED    = 1 (one logic signal is
> input for all homes)
> HOME_SEQUENCE    = 1 (second one to be done)
> 
> 
> In z [AXIS_2]
> HOME = 4.250  (where it marks it as being when
> finished)
> HOME_IGNORE_LIMITS = YES
> HOME_USE_INDEX = NO
> HOME_OFFSET = 4.090 ( distance in this case from left face
> of gauge to face 
> of pcb on right face of gauge)
> HOME_SEARCH_VEL = -0.08
> HOME_LATCH_VEL    = -0.015
> HOME_FINAL_VEL = 0.25
> HOME_IS_SHARED = 1
> HOME_SEQUENCE    = 0 (first to home)
> 
> I place a gauge I made on the ways, backed up against the
> work piece as 
> chucked, place a cleaned cutting tool next to the pcb copper
> that is the Z 
> detector on the gauge, and hit the HOME_ALL button,  It
> searches z for 
> contact, sets the home offset such that 0.0z is shaving the
> rust off the 
> chuck jaws, or if work is chucked, touching the face of the
> work.  Then it 
> backs away a wee bit and searches inward to find the contact
> for the x 
> home, finds that piece of pcb, sets home to pretty close to
> the center of 
> the work (I'm still fine tuning that) then parks the tool
> back out of the 
> way till its needed.
> 
> yerp, buzz, yerp, yerp, buzz, yerp, done in about 7
> seconds.
> 
> If the ways and base of my tool are lox clean, repeatable to
> under .001".  
> That however is a major problem for x as it tilts the gauge
> just for a good 
> coat of vactra on the ways.  I need to make the rear,
> huge flat area into a 
> single point contact with the back in order to penetrate the
> oil film and 
> possibly leave a little room for dust, between it and the
> flat back way on 
> that toy lathe.
> 
> The man page on homing could be clearer, but careful rereads
> will point the 
> way as there are many, many variations.  And there are
> N! ways to screw it 
> up.  :)
> 
> The only fly in this ointment is that it all goes totally,
> absolutely to 
> hell if you have applied /any/ touch offs, and in my case I
> would call not 
> canceling the touch off in effect as the first thing the
> homing operation 
> does a bug.  Ditto for trying to use position.txt, as
> that also 
> contaminates the zero's this sets.
> 
> IMO its a BUG, and for me a BIG BUG, requiring that I first
> stop linuxcnc, 
> nuke the position.txt AND the linuxcnc.vars files before
> restarting 
> linuxcnc, so that it truly starts from scratch when homing.
> 
> Touching the HOME button should first NUKE any touch off
> offsets in effect 
> at the time for the a

Re: [Emc-users] How about the core[X,Y] drive?

[Gene Heskett]

On Monday 15 April 2013 19:57:21 andy pugh did opine:

> On 15 April 2013 19:28, Gene Heskett  wrote:
> > Is there a valid reason for the current, save the offsets behavior?
> 
> Because that is what works best for a conventional homing where the
> switches are on the axis. Turn on the machine, home, and you are right
> back where you were the previous time.
> 
> You are actually trying to home by probing, which is a somewhat
> different thing, and is causing the problems you see.
> 
Ooookkayyy.  Now, what is the diff between probing for a switch closing 
where the switch may be mounted on the ways, and my probing for a ground 
contact where the contact just happens to be removable because otherwise it 
would get in the way of doing any work?

Seems functionally the same to me.  In my view, if it takes more than a 
thou of mental touch off to put the machine dead on, then its time to 
adjust the homing code until the touch off is no longer needed.

But on my small lathe, how would one go about putting tool offsets into the 
tool.tbl that mean something when in changing a tool, the stickout its 
clamped up at is effectively completely random give or take 1 cm or so?

To me, it makes 100x as much sense to mount the tool such that it can reach 
its cuts ok, drop the gauge on the ways and simply re-home the machine.  
FWIW, I have an .ini special for boring bars.  It searches outward to find 
the contact, but fudges some on the z because most boring bars on this 
gauge will only contact on the end of the bar, whereas the cutting tip is 
typically 1/8 to 1/4" shorter than the end of the bar.  And again, the 
amount of stickout is entirely arbitrary.  In fact I have one home made bar 
that I can stick out about 8".  Totally useless at that much stickout 
unless I'm boring room temp butter, but it is doable. :)

> Rather than using the home sequence you could make a button that calls
> a probing sequence to set your homes. This can properly take account
> of the variable lengths of your probes (or tools, as they are normally
> called).
> (you would need to use a G-code output to twiddle the
> halui.axis.N.home pins, and configure so that the axes all make as
> homed at their current position)

This I think, needs an example if you have the time.

Thanks Andy.

Cheers, Gene
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
My web page:  is up!
My views 

The doctrine of human equality reposes on this: that there is no man
really clever who has not found that he is stupid.
-- Gilbert K. Chesterson
A pen in the hand of this president is far more
dangerous than a gun in the hands of 200 million
  law-abiding citizens.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

--- On Mon, 4/15/13, jeremy youngs  wrote:

> From: jeremy youngs 
> Subject: Re: [Emc-users] How about the core[X,Y] drive?
> To: "Enhanced Machine Controller (EMC)" 
> Date: Monday, April 15, 2013, 11:53 AM
> on the torch head , touch off in a
> corner on 2 pads run 5 v to torch head
> and a debounce buffer then just touch off . should be more
> than sufficent
> for a plasma table , if a thc is used could just use the
> touch off pin output  no??

No THC on this torch but could mount a vertical V block to the table somewhere.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

--- On Mon, 4/15/13, Viesturs Lācis  wrote:

> Hmm, it is not the first time I see links to that Russian
> forum and with
> every time I get stronger and stronger impression that those
> LinuxCNC users
> that hang out there is a completely separate community...
> 
> Anyway, I just cannot figure out, how to set up homing
> routine to home such a machine to some switches.

Could mount limit switches to the pattern holder arms.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

--- On Mon, 4/15/13, Eric Keller  wrote:

> I have a plotter that was built like this using wire as the
> drive.  It was
> pretty neat, but the wire got messed up and then it was a
> pain to drive
> it.  Always thought it was a good idea, but a little
> weak

There's cable with a spiral of something wrapped around it and matching drive 
and idler pulleys. Dunno what it's called but looks like it wouldn't slip.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

--- On Mon, 4/15/13, andy pugh  wrote:

> On 15 April 2013 11:00, Gregg
> Eshelman 
> wrote:

> > Does LinuxCNC have H-bot support? CoreXY should drive
> identically.
> 
> Not built-in, but:
> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/26302-cnc-avec-systeme-h-bot-ou-corexy

I've seen that but Ich lese nicht Französisch. (Or German...)

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] gantrykins shenanigans

[andy pugh]

On 15 April 2013 21:42, Chris Kelley  wrote:

> Also, is there a guide written somewhere on how to get the JA3 branch (I've
> never used anything but the version on the live CDs).
>
> I tried googling, but I may not be saering for the right stuff.

http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Installing_LinuxCNC#Getting_the_source_with_git

But where is says "git checkout v2.5.0_branch" you type "git checkout
joints_axes3"

Then compile.

The default compile options will create a run-in-place system (all
located in the linuxcnc-dev folder) which won't interfere with the
installed version. This does lead to a potential confusion, in that
you may not end up running the version you think you are. You need to
open a terminal, type ". ./linuxcnc-dev/scripts/rip-environment" and
then "linuxcnc" to use your compiled version.

Also be aware that INI file sections have different names, and some
pin names are different, and to the existing config won't work (best
to copy the whole config folder and then edit the copy as a JA3
version)

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] gantrykins shenanigans

[Chris Kelley]

Andy, thanks for the help.

Is the joint mapping for gentrivkins the same as for gantrykins for slaving
joints/homing?

Also, is there a guide written somewhere on how to get the JA3 branch (I've
never used anything but the version on the live CDs).

I tried googling, but I may not be saering for the right stuff.


On Mon, Apr 15, 2013 at 3:12 PM, andy pugh  wrote:

> On 15 April 2013 20:15, Chris Kelley  wrote:
>
> > Is there a way to make gantrykins actually obey the limits set in the
> .ini
> > file?
>
> I think it is a "feature".
>
> Gantrykins seems to cause more problems than it solves. You could try
> running the JA3 branch and the new "gentrivkins" which is intended for
> this type of machine.
>
> --
> atp
> If you can't fix it, you don't own it.
> http://www.ifixit.com/Manifesto
>
>
> --
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] gantrykins shenanigans

[andy pugh]

On 15 April 2013 20:15, Chris Kelley  wrote:

> Is there a way to make gantrykins actually obey the limits set in the .ini
> file?

I think it is a "feature".

Gantrykins seems to cause more problems than it solves. You could try
running the JA3 branch and the new "gentrivkins" which is intended for
this type of machine.

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[andy pugh]

On 15 April 2013 19:28, Gene Heskett  wrote:

> Is there a valid reason for the current, save the offsets behavior?

Because that is what works best for a conventional homing where the
switches are on the axis. Turn on the machine, home, and you are right
back where you were the previous time.

You are actually trying to home by probing, which is a somewhat
different thing, and is causing the problems you see.

Rather than using the home sequence you could make a button that calls
a probing sequence to set your homes. This can properly take account
of the variable lengths of your probes (or tools, as they are normally
called).
(you would need to use a G-code output to twiddle the
halui.axis.N.home pins, and configure so that the axes all make as
homed at their current position)

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] gantrykins shenanigans

[Chris Kelley]

I have a plasma table that I converted over from Torchmate control (that
system sucked) to LinuxCNC with a Mesa 5i25/7i76 and dual motors on the
gantry (X-axis).

I have gantrykins set up correctly (as far as I know). Homing and jogging,
in joint mode work as expected. However, when jogging in world mode, the
machine will jog right past the soft limits (while popping up an error in
axis) until they hit the actual limit switch. In joint mode the joint will
stop jogging when it get to the limit.

Is there a way to make gantrykins actually obey the limits set in the .ini
file?
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gene Heskett]

On Monday 15 April 2013 15:09:03 jeremy youngs did opine:

> yeah im kinda fond of the old man myself :)

Chuckle, next thing I know I'd better stock up on suds in case you all come 
to visit. ;-)
 
> On Mon, Apr 15, 2013 at 2:42 PM, sam sokolik  
wrote:
> > Thanks  :)
> > 
> > love you
> > sam
> > 
> > On 4/15/2013 1:28 PM, Gene Heskett wrote:
> > > On Monday 15 April 2013 13:22:31 Viesturs Lؤپcis did opine:
> > >> 2013/4/15 andy pugh 
> > >> 
> > >>> On 15 April 2013 11:00, Gregg Eshelman  wrote:
> >  The crossover counters any offset push at one end of the gantry
> >  with an
> > >>> 
> > >>> equal pull in the same direction at the opposite end. The theory
> > >>> page shows how it works.
> > >>> 
> > >>> I have stared at the diagrams for long enough now that I have
> > >>> changed my mind several times. I might have to build one to see.
> > >>> 
> >  Does LinuxCNC have H-bot support? CoreXY should drive
> >  identically.
> > >>> 
> > >>> Not built-in, but:
> > >>> 
> > >>> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcn
> > >>> c-qu estions/26302-cnc-avec-systeme-h-bot-ou-corexy
> > >> 
> > >> Hmm, it is not the first time I see links to that Russian forum and
> > >> with every time I get stronger and stronger impression that those
> > >> LinuxCNC users that hang out there is a completely separate
> > >> community...
> > >> 
> > >> Anyway, I just cannot figure out, how to set up homing routine to
> > >> home such a machine to some switches.
> > > 
> > > Use the sign of the HOME_SEARCH_VELOCITY to control which switch its
> > > looking for.  Then HOME_OFFSET with a sign can set home to any
> > > position. Then in case the home switches are off in a corner
> > > someplace and you have to jog the machine to a suitable starting
> > > position before invoking the homing operation, run it to where the
> > > first switch is available quickly, and use the HOME_SEQUENCE number
> > > to put all the home ops in the proper order.
> > > 
> > > On my lathe, a grep of the .ini for HOME returns this:
> > > In [TRAJ]
> > > HOME_ALL  = 1 (one home all button in axis)
> > > 
> > > In x [AXIS_0]
> > > HOME = 1.250  (where it parks it when done)
> > > HOME_IGNORE_LIMITS = YES
> > > HOME_USE_INDEX = NO  (if servo's, could be useful)
> > > HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its
> > > at) HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> > > HOME_LATCH_VEL = -.008
> > > HOME_FINAL_VEL = -.1
> > > HOME_IS_SHARED= 1 (one logic signal is input for all homes)
> > > HOME_SEQUENCE = 1 (second one to be done)
> > > 
> > > 
> > > In z [AXIS_2]
> > > HOME = 4.250  (where it marks it as being when finished)
> > > HOME_IGNORE_LIMITS = YES
> > > HOME_USE_INDEX = NO
> > > HOME_OFFSET = 4.090 ( distance in this case from left face of gauge
> > > to
> > 
> > face
> > 
> > > of pcb on right face of gauge)
> > > HOME_SEARCH_VEL = -0.08
> > > HOME_LATCH_VEL= -0.015
> > > HOME_FINAL_VEL = 0.25
> > > HOME_IS_SHARED = 1
> > > HOME_SEQUENCE = 0 (first to home)
> > > 
> > > I place a gauge I made on the ways, backed up against the work piece
> > > as chucked, place a cleaned cutting tool next to the pcb copper
> > > that is the
> > 
> > Z
> > 
> > > detector on the gauge, and hit the HOME_ALL button,  It searches z
> > > for contact, sets the home offset such that 0.0z is shaving the
> > > rust off the chuck jaws, or if work is chucked, touching the face
> > > of the work.  Then
> > 
> > it
> > 
> > > backs away a wee bit and searches inward to find the contact for the
> > > x home, finds that piece of pcb, sets home to pretty close to the
> > > center of the work (I'm still fine tuning that) then parks the tool
> > > back out of the way till its needed.
> > > 
> > > yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.
> > > 
> > > If the ways and base of my tool are lox clean, repeatable to under
> > > .001". That however is a major problem for x as it tilts the gauge
> > > just for a
> > 
> > good
> > 
> > > coat of vactra on the ways.  I need to make the rear, huge flat area
> > 
> > into a
> > 
> > > single point contact with the back in order to penetrate the oil
> > > film and possibly leave a little room for dust, between it and the
> > > flat back way
> > 
> > on
> > 
> > > that toy lathe.
> > > 
> > > The man page on homing could be clearer, but careful rereads will
> > > point
> > 
> > the
> > 
> > > way as there are many, many variations.  And there are N! ways to
> > > screw
> > 
> > it
> > 
> > > up.  :)
> > > 
> > > The only fly in this ointment is that it all goes totally,
> > > absolutely to hell if you have applied /any/ touch offs, and in my
> > > case I would call
> > 
> > not
> > 
> > > canceling the touch off in effect as the first thing the homing
> > > operation does a bug.  Ditto for trying to use position.txt, as
> > > that also contaminates the zero's this sets.
> > > 
> > > IMO its a BUG, and for me a BIG BUG, requiring that I fi

Re: [Emc-users] How about the core[X,Y] drive?

[Gene Heskett]

On Monday 15 April 2013 14:55:39 sam sokolik did opine:

> Thanks  :)
> 
> love you
> sam

Thanks Sam, I like the flowers but what brought that on?  There are 
probably 200+ people here who _could_ write a similar tome that described 
their particular hardware (miss-)configurations. :)

OTOH, I am well aware that turning swarf into money comes first, and many 
might not feel like they have the time.  Generally, I do as long as the 
grass doesn't need cut yadda yadda.

That is, after all, what this list is about, helping folks understand and 
use linuxcnc.  This list, and IRC too, has helped me a lot, and where I 
can, since I'm retired with no real schedule to meet, I try to pay it back.  

So I thank you all for the help I have been given by folks who stopped what 
they were doing just to help me.  Sometimes in the heat of solving a 
problem I forget to say that, but the help is most certainly appreciated.
> 
> On 4/15/2013 1:28 PM, Gene Heskett wrote:
> > On Monday 15 April 2013 13:22:31 Viesturs Lؤپcis did opine:
> >> 2013/4/15 andy pugh 
> >> 
> >>> On 15 April 2013 11:00, Gregg Eshelman  wrote:
>  The crossover counters any offset push at one end of the gantry
>  with an
> >>> 
> >>> equal pull in the same direction at the opposite end. The theory
> >>> page shows how it works.
> >>> 
> >>> I have stared at the diagrams for long enough now that I have
> >>> changed my mind several times. I might have to build one to see.
> >>> 
>  Does LinuxCNC have H-bot support? CoreXY should drive identically.
> >>> 
> >>> Not built-in, but:
> >>> 
> >>> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-
> >>> qu estions/26302-cnc-avec-systeme-h-bot-ou-corexy
> >> 
> >> Hmm, it is not the first time I see links to that Russian forum and
> >> with every time I get stronger and stronger impression that those
> >> LinuxCNC users that hang out there is a completely separate
> >> community...
> >> 
> >> Anyway, I just cannot figure out, how to set up homing routine to
> >> home such a machine to some switches.
> > 
> > Use the sign of the HOME_SEARCH_VELOCITY to control which switch its
> > looking for.  Then HOME_OFFSET with a sign can set home to any
> > position. Then in case the home switches are off in a corner
> > someplace and you have to jog the machine to a suitable starting
> > position before invoking the homing operation, run it to where the
> > first switch is available quickly, and use the HOME_SEQUENCE number
> > to put all the home ops in the proper order.
> > 
> > On my lathe, a grep of the .ini for HOME returns this:
> > In [TRAJ]
> > HOME_ALL= 1 (one home all button in axis)
> > 
> > In x [AXIS_0]
> > HOME = 1.250  (where it parks it when done)
> > HOME_IGNORE_LIMITS = YES
> > HOME_USE_INDEX = NO  (if servo's, could be useful)
> > HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its at)
> > HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> > HOME_LATCH_VEL = -.008
> > HOME_FINAL_VEL = -.1
> > HOME_IS_SHARED  = 1 (one logic signal is input for all homes)
> > HOME_SEQUENCE   = 1 (second one to be done)
> > 
> > 
> > In z [AXIS_2]
> > HOME = 4.250  (where it marks it as being when finished)
> > HOME_IGNORE_LIMITS = YES
> > HOME_USE_INDEX = NO
> > HOME_OFFSET = 4.090 ( distance in this case from left face of gauge to
> > face of pcb on right face of gauge)
> > HOME_SEARCH_VEL = -0.08
> > HOME_LATCH_VEL  = -0.015
> > HOME_FINAL_VEL = 0.25
> > HOME_IS_SHARED = 1
> > HOME_SEQUENCE   = 0 (first to home)
> > 
> > I place a gauge I made on the ways, backed up against the work piece
> > as chucked, place a cleaned cutting tool next to the pcb copper that
> > is the Z detector on the gauge, and hit the HOME_ALL button,  It
> > searches z for contact, sets the home offset such that 0.0z is
> > shaving the rust off the chuck jaws, or if work is chucked, touching
> > the face of the work.  Then it backs away a wee bit and searches
> > inward to find the contact for the x home, finds that piece of pcb,
> > sets home to pretty close to the center of the work (I'm still fine
> > tuning that) then parks the tool back out of the way till its needed.
> > 
> > yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.
> > 
> > If the ways and base of my tool are lox clean, repeatable to under
> > .001". That however is a major problem for x as it tilts the gauge
> > just for a good coat of vactra on the ways.  I need to make the rear,
> > huge flat area into a single point contact with the back in order to
> > penetrate the oil film and possibly leave a little room for dust,
> > between it and the flat back way on that toy lathe.
> > 
> > The man page on homing could be clearer, but careful rereads will
> > point the way as there are many, many variations.  And there are N!
> > ways to screw it up.  :)
> > 
> > The only fly in this ointment is that it all goes totally, absolutely
> > to hell if you have applied /any/ touch offs, and in my case I would
> >

Re: [Emc-users] How about the core[X,Y] drive?

[jeremy youngs]

I agree with Todd that some marker on a belt would be best, but only thing
I can come up with is to draw some line or something and then try to
optically read it.

well that is how fadal always did it bur they didnt have a way to read it
the operator just eyeballed it .


On Mon, Apr 15, 2013 at 2:49 PM, Viesturs Lācis wrote:

> 2013/4/15 jeremy youngs 
>
> > right but i dont see why automated homing would be excluded . just using
> > the panels as the touch off instead of homing switches
> > pretty simple no?
> >
>
> No, because to move torch (or whatever instrument) along X, both joints
> have to turn in opposite directions, but for movement along Y both joints
> have to move in the same direction. I do not see a way to home each joint
> separately by triggering homing switches with gantry or whatever that is in
> movement.
> I agree with Todd that some marker on a belt would be best, but only thing
> I can come up with is to draw some line or something and then try to
> optically read it.
> Or maybe some tweaking with identical home search and latch velocity for
> one joint and home_latch_vel = 0 for the other joint together with some
> HAL/ladder magic to treat the actual switch signals so that it looks
> something like this example:
> 1) hit "home all" and both joints are searching the limit switch - both
> motors turn in opposite directions and move along X
> 2) when X axis home switch is reached, one motor starts to latch from the
> switch, both motors turn in the same direction and now the torch moves
> along Y;
> 3) when Y home switch is triggered, HAL  should tell the first joint, that
> its switch is released and the second joint would also be homed.
> My theory says this should work :)) Those velocities should be really slow
> to make the homing more precise.
>
> --
> Viesturs
>
> If you can't fix it, you don't own it.
> http://www.ifixit.com/Manifesto
>
> --
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 
jeremy youngs
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Viesturs Lācis]

2013/4/15 jeremy youngs 

> right but i dont see why automated homing would be excluded . just using
> the panels as the touch off instead of homing switches
> pretty simple no?
>

No, because to move torch (or whatever instrument) along X, both joints
have to turn in opposite directions, but for movement along Y both joints
have to move in the same direction. I do not see a way to home each joint
separately by triggering homing switches with gantry or whatever that is in
movement.
I agree with Todd that some marker on a belt would be best, but only thing
I can come up with is to draw some line or something and then try to
optically read it.
Or maybe some tweaking with identical home search and latch velocity for
one joint and home_latch_vel = 0 for the other joint together with some
HAL/ladder magic to treat the actual switch signals so that it looks
something like this example:
1) hit "home all" and both joints are searching the limit switch - both
motors turn in opposite directions and move along X
2) when X axis home switch is reached, one motor starts to latch from the
switch, both motors turn in the same direction and now the torch moves
along Y;
3) when Y home switch is triggered, HAL  should tell the first joint, that
its switch is released and the second joint would also be homed.
My theory says this should work :)) Those velocities should be really slow
to make the homing more precise.

-- 
Viesturs

If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[jeremy youngs]

even more so for me now :)


On Mon, Apr 15, 2013 at 2:48 PM, jeremy youngs  wrote:

> yeah im kinda fond of the old man myself :)
>
>
> On Mon, Apr 15, 2013 at 2:42 PM, sam sokolik wrote:
>
>> Thanks  :)
>>
>> love you
>> sam
>> On 4/15/2013 1:28 PM, Gene Heskett wrote:
>> > On Monday 15 April 2013 13:22:31 Viesturs Lācis did opine:
>> >
>> >> 2013/4/15 andy pugh 
>> >>
>> >>> On 15 April 2013 11:00, Gregg Eshelman  wrote:
>>  The crossover counters any offset push at one end of the gantry with
>>  an
>> >>> equal pull in the same direction at the opposite end. The theory page
>> >>> shows how it works.
>> >>>
>> >>> I have stared at the diagrams for long enough now that I have changed
>> >>> my mind several times. I might have to build one to see.
>> >>>
>>  Does LinuxCNC have H-bot support? CoreXY should drive identically.
>> >>> Not built-in, but:
>> >>>
>> >>>
>> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-qu
>> >>> estions/26302-cnc-avec-systeme-h-bot-ou-corexy
>> >> Hmm, it is not the first time I see links to that Russian forum and
>> with
>> >> every time I get stronger and stronger impression that those LinuxCNC
>> >> users that hang out there is a completely separate community...
>> >>
>> >> Anyway, I just cannot figure out, how to set up homing routine to home
>> >> such a machine to some switches.
>> > Use the sign of the HOME_SEARCH_VELOCITY to control which switch its
>> > looking for.  Then HOME_OFFSET with a sign can set home to any position.
>> > Then in case the home switches are off in a corner someplace and you
>> have
>> > to jog the machine to a suitable starting position before invoking the
>> > homing operation, run it to where the first switch is available quickly,
>> > and use the HOME_SEQUENCE number to put all the home ops in the proper
>> > order.
>> >
>> > On my lathe, a grep of the .ini for HOME returns this:
>> > In [TRAJ]
>> > HOME_ALL  = 1 (one home all button in axis)
>> >
>> > In x [AXIS_0]
>> > HOME = 1.250  (where it parks it when done)
>> > HOME_IGNORE_LIMITS = YES
>> > HOME_USE_INDEX = NO  (if servo's, could be useful)
>> > HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its at)
>> > HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
>> > HOME_LATCH_VEL = -.008
>> > HOME_FINAL_VEL = -.1
>> > HOME_IS_SHARED= 1 (one logic signal is input for all homes)
>> > HOME_SEQUENCE = 1 (second one to be done)
>> >
>> >
>> > In z [AXIS_2]
>> > HOME = 4.250  (where it marks it as being when finished)
>> > HOME_IGNORE_LIMITS = YES
>> > HOME_USE_INDEX = NO
>> > HOME_OFFSET = 4.090 ( distance in this case from left face of gauge to
>> face
>> > of pcb on right face of gauge)
>> > HOME_SEARCH_VEL = -0.08
>> > HOME_LATCH_VEL= -0.015
>> > HOME_FINAL_VEL = 0.25
>> > HOME_IS_SHARED = 1
>> > HOME_SEQUENCE = 0 (first to home)
>> >
>> > I place a gauge I made on the ways, backed up against the work piece as
>> > chucked, place a cleaned cutting tool next to the pcb copper that is
>> the Z
>> > detector on the gauge, and hit the HOME_ALL button,  It searches z for
>> > contact, sets the home offset such that 0.0z is shaving the rust off the
>> > chuck jaws, or if work is chucked, touching the face of the work.  Then
>> it
>> > backs away a wee bit and searches inward to find the contact for the x
>> > home, finds that piece of pcb, sets home to pretty close to the center
>> of
>> > the work (I'm still fine tuning that) then parks the tool back out of
>> the
>> > way till its needed.
>> >
>> > yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.
>> >
>> > If the ways and base of my tool are lox clean, repeatable to under
>> .001".
>> > That however is a major problem for x as it tilts the gauge just for a
>> good
>> > coat of vactra on the ways.  I need to make the rear, huge flat area
>> into a
>> > single point contact with the back in order to penetrate the oil film
>> and
>> > possibly leave a little room for dust, between it and the flat back way
>> on
>> > that toy lathe.
>> >
>> > The man page on homing could be clearer, but careful rereads will point
>> the
>> > way as there are many, many variations.  And there are N! ways to screw
>> it
>> > up.  :)
>> >
>> > The only fly in this ointment is that it all goes totally, absolutely to
>> > hell if you have applied /any/ touch offs, and in my case I would call
>> not
>> > canceling the touch off in effect as the first thing the homing
>> operation
>> > does a bug.  Ditto for trying to use position.txt, as that also
>> > contaminates the zero's this sets.
>> >
>> > IMO its a BUG, and for me a BIG BUG, requiring that I first stop
>> linuxcnc,
>> > nuke the position.txt AND the linuxcnc.vars files before restarting
>> > linuxcnc, so that it truly starts from scratch when homing.
>> >
>> > Touching the HOME button should first NUKE any touch off offsets in
>> effect
>> > at the time for the axis being homed.
>> >
>> > It does not, so home is then relative to the &^

Re: [Emc-users] How about the core[X,Y] drive?

[jeremy youngs]

yeah im kinda fond of the old man myself :)


On Mon, Apr 15, 2013 at 2:42 PM, sam sokolik  wrote:

> Thanks  :)
>
> love you
> sam
> On 4/15/2013 1:28 PM, Gene Heskett wrote:
> > On Monday 15 April 2013 13:22:31 Viesturs Lācis did opine:
> >
> >> 2013/4/15 andy pugh 
> >>
> >>> On 15 April 2013 11:00, Gregg Eshelman  wrote:
>  The crossover counters any offset push at one end of the gantry with
>  an
> >>> equal pull in the same direction at the opposite end. The theory page
> >>> shows how it works.
> >>>
> >>> I have stared at the diagrams for long enough now that I have changed
> >>> my mind several times. I might have to build one to see.
> >>>
>  Does LinuxCNC have H-bot support? CoreXY should drive identically.
> >>> Not built-in, but:
> >>>
> >>> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-qu
> >>> estions/26302-cnc-avec-systeme-h-bot-ou-corexy
> >> Hmm, it is not the first time I see links to that Russian forum and with
> >> every time I get stronger and stronger impression that those LinuxCNC
> >> users that hang out there is a completely separate community...
> >>
> >> Anyway, I just cannot figure out, how to set up homing routine to home
> >> such a machine to some switches.
> > Use the sign of the HOME_SEARCH_VELOCITY to control which switch its
> > looking for.  Then HOME_OFFSET with a sign can set home to any position.
> > Then in case the home switches are off in a corner someplace and you have
> > to jog the machine to a suitable starting position before invoking the
> > homing operation, run it to where the first switch is available quickly,
> > and use the HOME_SEQUENCE number to put all the home ops in the proper
> > order.
> >
> > On my lathe, a grep of the .ini for HOME returns this:
> > In [TRAJ]
> > HOME_ALL  = 1 (one home all button in axis)
> >
> > In x [AXIS_0]
> > HOME = 1.250  (where it parks it when done)
> > HOME_IGNORE_LIMITS = YES
> > HOME_USE_INDEX = NO  (if servo's, could be useful)
> > HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its at)
> > HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> > HOME_LATCH_VEL = -.008
> > HOME_FINAL_VEL = -.1
> > HOME_IS_SHARED= 1 (one logic signal is input for all homes)
> > HOME_SEQUENCE = 1 (second one to be done)
> >
> >
> > In z [AXIS_2]
> > HOME = 4.250  (where it marks it as being when finished)
> > HOME_IGNORE_LIMITS = YES
> > HOME_USE_INDEX = NO
> > HOME_OFFSET = 4.090 ( distance in this case from left face of gauge to
> face
> > of pcb on right face of gauge)
> > HOME_SEARCH_VEL = -0.08
> > HOME_LATCH_VEL= -0.015
> > HOME_FINAL_VEL = 0.25
> > HOME_IS_SHARED = 1
> > HOME_SEQUENCE = 0 (first to home)
> >
> > I place a gauge I made on the ways, backed up against the work piece as
> > chucked, place a cleaned cutting tool next to the pcb copper that is the
> Z
> > detector on the gauge, and hit the HOME_ALL button,  It searches z for
> > contact, sets the home offset such that 0.0z is shaving the rust off the
> > chuck jaws, or if work is chucked, touching the face of the work.  Then
> it
> > backs away a wee bit and searches inward to find the contact for the x
> > home, finds that piece of pcb, sets home to pretty close to the center of
> > the work (I'm still fine tuning that) then parks the tool back out of the
> > way till its needed.
> >
> > yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.
> >
> > If the ways and base of my tool are lox clean, repeatable to under .001".
> > That however is a major problem for x as it tilts the gauge just for a
> good
> > coat of vactra on the ways.  I need to make the rear, huge flat area
> into a
> > single point contact with the back in order to penetrate the oil film and
> > possibly leave a little room for dust, between it and the flat back way
> on
> > that toy lathe.
> >
> > The man page on homing could be clearer, but careful rereads will point
> the
> > way as there are many, many variations.  And there are N! ways to screw
> it
> > up.  :)
> >
> > The only fly in this ointment is that it all goes totally, absolutely to
> > hell if you have applied /any/ touch offs, and in my case I would call
> not
> > canceling the touch off in effect as the first thing the homing operation
> > does a bug.  Ditto for trying to use position.txt, as that also
> > contaminates the zero's this sets.
> >
> > IMO its a BUG, and for me a BIG BUG, requiring that I first stop
> linuxcnc,
> > nuke the position.txt AND the linuxcnc.vars files before restarting
> > linuxcnc, so that it truly starts from scratch when homing.
> >
> > Touching the HOME button should first NUKE any touch off offsets in
> effect
> > at the time for the axis being homed.
> >
> > It does not, so home is then relative to the &^$#@ offset.  I have broken
> > tools and wrecked nearly finished parts because of this, several times.
> >
> > Is there a valid reason for the current, save the offsets behavior?
> >
> > Cheers, Gene
>
>
>
> 

Re: [Emc-users] How about the core[X,Y] drive?

[sam sokolik]

heh - Sorry.  replied to wrong email..  oops  (Boy that could have been 
really embarrassing...)

sam
On 4/15/2013 1:28 PM, Gene Heskett wrote:
> On Monday 15 April 2013 13:22:31 Viesturs Lācis did opine:
>
>> 2013/4/15 andy pugh 
>>
>>> On 15 April 2013 11:00, Gregg Eshelman  wrote:
 The crossover counters any offset push at one end of the gantry with
 an
>>> equal pull in the same direction at the opposite end. The theory page
>>> shows how it works.
>>>
>>> I have stared at the diagrams for long enough now that I have changed
>>> my mind several times. I might have to build one to see.
>>>
 Does LinuxCNC have H-bot support? CoreXY should drive identically.
>>> Not built-in, but:
>>>
>>> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-qu
>>> estions/26302-cnc-avec-systeme-h-bot-ou-corexy
>> Hmm, it is not the first time I see links to that Russian forum and with
>> every time I get stronger and stronger impression that those LinuxCNC
>> users that hang out there is a completely separate community...
>>
>> Anyway, I just cannot figure out, how to set up homing routine to home
>> such a machine to some switches.
> Use the sign of the HOME_SEARCH_VELOCITY to control which switch its
> looking for.  Then HOME_OFFSET with a sign can set home to any position.
> Then in case the home switches are off in a corner someplace and you have
> to jog the machine to a suitable starting position before invoking the
> homing operation, run it to where the first switch is available quickly,
> and use the HOME_SEQUENCE number to put all the home ops in the proper
> order.
>
> On my lathe, a grep of the .ini for HOME returns this:
> In [TRAJ]
> HOME_ALL  = 1 (one home all button in axis)
>
> In x [AXIS_0]
> HOME = 1.250  (where it parks it when done)
> HOME_IGNORE_LIMITS = YES
> HOME_USE_INDEX = NO  (if servo's, could be useful)
> HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its at)
> HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> HOME_LATCH_VEL = -.008
> HOME_FINAL_VEL = -.1
> HOME_IS_SHARED= 1 (one logic signal is input for all homes)
> HOME_SEQUENCE = 1 (second one to be done)
>
>
> In z [AXIS_2]
> HOME = 4.250  (where it marks it as being when finished)
> HOME_IGNORE_LIMITS = YES
> HOME_USE_INDEX = NO
> HOME_OFFSET = 4.090 ( distance in this case from left face of gauge to face
> of pcb on right face of gauge)
> HOME_SEARCH_VEL = -0.08
> HOME_LATCH_VEL= -0.015
> HOME_FINAL_VEL = 0.25
> HOME_IS_SHARED = 1
> HOME_SEQUENCE = 0 (first to home)
>
> I place a gauge I made on the ways, backed up against the work piece as
> chucked, place a cleaned cutting tool next to the pcb copper that is the Z
> detector on the gauge, and hit the HOME_ALL button,  It searches z for
> contact, sets the home offset such that 0.0z is shaving the rust off the
> chuck jaws, or if work is chucked, touching the face of the work.  Then it
> backs away a wee bit and searches inward to find the contact for the x
> home, finds that piece of pcb, sets home to pretty close to the center of
> the work (I'm still fine tuning that) then parks the tool back out of the
> way till its needed.
>
> yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.
>
> If the ways and base of my tool are lox clean, repeatable to under .001".
> That however is a major problem for x as it tilts the gauge just for a good
> coat of vactra on the ways.  I need to make the rear, huge flat area into a
> single point contact with the back in order to penetrate the oil film and
> possibly leave a little room for dust, between it and the flat back way on
> that toy lathe.
>
> The man page on homing could be clearer, but careful rereads will point the
> way as there are many, many variations.  And there are N! ways to screw it
> up.  :)
>
> The only fly in this ointment is that it all goes totally, absolutely to
> hell if you have applied /any/ touch offs, and in my case I would call not
> canceling the touch off in effect as the first thing the homing operation
> does a bug.  Ditto for trying to use position.txt, as that also
> contaminates the zero's this sets.
>
> IMO its a BUG, and for me a BIG BUG, requiring that I first stop linuxcnc,
> nuke the position.txt AND the linuxcnc.vars files before restarting
> linuxcnc, so that it truly starts from scratch when homing.
>
> Touching the HOME button should first NUKE any touch off offsets in effect
> at the time for the axis being homed.
>
> It does not, so home is then relative to the &^$#@ offset.  I have broken
> tools and wrecked nearly finished parts because of this, several times.
>
> Is there a valid reason for the current, save the offsets behavior?
>
> Cheers, Gene


--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can 

Re: [Emc-users] How about the core[X,Y] drive?

[sam sokolik]

Thanks  :)

love you
sam
On 4/15/2013 1:28 PM, Gene Heskett wrote:
> On Monday 15 April 2013 13:22:31 Viesturs Lācis did opine:
>
>> 2013/4/15 andy pugh 
>>
>>> On 15 April 2013 11:00, Gregg Eshelman  wrote:
 The crossover counters any offset push at one end of the gantry with
 an
>>> equal pull in the same direction at the opposite end. The theory page
>>> shows how it works.
>>>
>>> I have stared at the diagrams for long enough now that I have changed
>>> my mind several times. I might have to build one to see.
>>>
 Does LinuxCNC have H-bot support? CoreXY should drive identically.
>>> Not built-in, but:
>>>
>>> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-qu
>>> estions/26302-cnc-avec-systeme-h-bot-ou-corexy
>> Hmm, it is not the first time I see links to that Russian forum and with
>> every time I get stronger and stronger impression that those LinuxCNC
>> users that hang out there is a completely separate community...
>>
>> Anyway, I just cannot figure out, how to set up homing routine to home
>> such a machine to some switches.
> Use the sign of the HOME_SEARCH_VELOCITY to control which switch its
> looking for.  Then HOME_OFFSET with a sign can set home to any position.
> Then in case the home switches are off in a corner someplace and you have
> to jog the machine to a suitable starting position before invoking the
> homing operation, run it to where the first switch is available quickly,
> and use the HOME_SEQUENCE number to put all the home ops in the proper
> order.
>
> On my lathe, a grep of the .ini for HOME returns this:
> In [TRAJ]
> HOME_ALL  = 1 (one home all button in axis)
>
> In x [AXIS_0]
> HOME = 1.250  (where it parks it when done)
> HOME_IGNORE_LIMITS = YES
> HOME_USE_INDEX = NO  (if servo's, could be useful)
> HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its at)
> HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
> HOME_LATCH_VEL = -.008
> HOME_FINAL_VEL = -.1
> HOME_IS_SHARED= 1 (one logic signal is input for all homes)
> HOME_SEQUENCE = 1 (second one to be done)
>
>
> In z [AXIS_2]
> HOME = 4.250  (where it marks it as being when finished)
> HOME_IGNORE_LIMITS = YES
> HOME_USE_INDEX = NO
> HOME_OFFSET = 4.090 ( distance in this case from left face of gauge to face
> of pcb on right face of gauge)
> HOME_SEARCH_VEL = -0.08
> HOME_LATCH_VEL= -0.015
> HOME_FINAL_VEL = 0.25
> HOME_IS_SHARED = 1
> HOME_SEQUENCE = 0 (first to home)
>
> I place a gauge I made on the ways, backed up against the work piece as
> chucked, place a cleaned cutting tool next to the pcb copper that is the Z
> detector on the gauge, and hit the HOME_ALL button,  It searches z for
> contact, sets the home offset such that 0.0z is shaving the rust off the
> chuck jaws, or if work is chucked, touching the face of the work.  Then it
> backs away a wee bit and searches inward to find the contact for the x
> home, finds that piece of pcb, sets home to pretty close to the center of
> the work (I'm still fine tuning that) then parks the tool back out of the
> way till its needed.
>
> yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.
>
> If the ways and base of my tool are lox clean, repeatable to under .001".
> That however is a major problem for x as it tilts the gauge just for a good
> coat of vactra on the ways.  I need to make the rear, huge flat area into a
> single point contact with the back in order to penetrate the oil film and
> possibly leave a little room for dust, between it and the flat back way on
> that toy lathe.
>
> The man page on homing could be clearer, but careful rereads will point the
> way as there are many, many variations.  And there are N! ways to screw it
> up.  :)
>
> The only fly in this ointment is that it all goes totally, absolutely to
> hell if you have applied /any/ touch offs, and in my case I would call not
> canceling the touch off in effect as the first thing the homing operation
> does a bug.  Ditto for trying to use position.txt, as that also
> contaminates the zero's this sets.
>
> IMO its a BUG, and for me a BIG BUG, requiring that I first stop linuxcnc,
> nuke the position.txt AND the linuxcnc.vars files before restarting
> linuxcnc, so that it truly starts from scratch when homing.
>
> Touching the HOME button should first NUKE any touch off offsets in effect
> at the time for the axis being homed.
>
> It does not, so home is then relative to the &^$#@ offset.  I have broken
> tools and wrecked nearly finished parts because of this, several times.
>
> Is there a valid reason for the current, save the offsets behavior?
>
> Cheers, Gene


--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
h

Re: [Emc-users] How about the core[X,Y] drive?

[Gene Heskett]

On Monday 15 April 2013 13:22:31 Viesturs Lācis did opine:

> 2013/4/15 andy pugh 
> 
> > On 15 April 2013 11:00, Gregg Eshelman  wrote:
> > > The crossover counters any offset push at one end of the gantry with
> > > an
> > 
> > equal pull in the same direction at the opposite end. The theory page
> > shows how it works.
> > 
> > I have stared at the diagrams for long enough now that I have changed
> > my mind several times. I might have to build one to see.
> > 
> > > Does LinuxCNC have H-bot support? CoreXY should drive identically.
> > 
> > Not built-in, but:
> > 
> > http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-qu
> > estions/26302-cnc-avec-systeme-h-bot-ou-corexy
> 
> Hmm, it is not the first time I see links to that Russian forum and with
> every time I get stronger and stronger impression that those LinuxCNC
> users that hang out there is a completely separate community...
> 
> Anyway, I just cannot figure out, how to set up homing routine to home
> such a machine to some switches.

Use the sign of the HOME_SEARCH_VELOCITY to control which switch its 
looking for.  Then HOME_OFFSET with a sign can set home to any position.  
Then in case the home switches are off in a corner someplace and you have 
to jog the machine to a suitable starting position before invoking the 
homing operation, run it to where the first switch is available quickly, 
and use the HOME_SEQUENCE number to put all the home ops in the proper 
order.

On my lathe, a grep of the .ini for HOME returns this:
In [TRAJ]
HOME_ALL= 1 (one home all button in axis)

In x [AXIS_0]
HOME = 1.250  (where it parks it when done)
HOME_IGNORE_LIMITS = YES
HOME_USE_INDEX = NO  (if servo's, could be useful)
HOME_OFFSET = 0.0100 (where it marks it as 0.0 from where its at)
HOME_SEARCH_VEL = -0.08 (in machine units, IPS IOW)
HOME_LATCH_VEL = -.008
HOME_FINAL_VEL = -.1
HOME_IS_SHARED  = 1 (one logic signal is input for all homes)
HOME_SEQUENCE   = 1 (second one to be done)


In z [AXIS_2]
HOME = 4.250  (where it marks it as being when finished)
HOME_IGNORE_LIMITS = YES
HOME_USE_INDEX = NO
HOME_OFFSET = 4.090 ( distance in this case from left face of gauge to face 
of pcb on right face of gauge)
HOME_SEARCH_VEL = -0.08
HOME_LATCH_VEL  = -0.015
HOME_FINAL_VEL = 0.25
HOME_IS_SHARED = 1
HOME_SEQUENCE   = 0 (first to home)

I place a gauge I made on the ways, backed up against the work piece as 
chucked, place a cleaned cutting tool next to the pcb copper that is the Z 
detector on the gauge, and hit the HOME_ALL button,  It searches z for 
contact, sets the home offset such that 0.0z is shaving the rust off the 
chuck jaws, or if work is chucked, touching the face of the work.  Then it 
backs away a wee bit and searches inward to find the contact for the x 
home, finds that piece of pcb, sets home to pretty close to the center of 
the work (I'm still fine tuning that) then parks the tool back out of the 
way till its needed.

yerp, buzz, yerp, yerp, buzz, yerp, done in about 7 seconds.

If the ways and base of my tool are lox clean, repeatable to under .001".  
That however is a major problem for x as it tilts the gauge just for a good 
coat of vactra on the ways.  I need to make the rear, huge flat area into a 
single point contact with the back in order to penetrate the oil film and 
possibly leave a little room for dust, between it and the flat back way on 
that toy lathe.

The man page on homing could be clearer, but careful rereads will point the 
way as there are many, many variations.  And there are N! ways to screw it 
up.  :)

The only fly in this ointment is that it all goes totally, absolutely to 
hell if you have applied /any/ touch offs, and in my case I would call not 
canceling the touch off in effect as the first thing the homing operation 
does a bug.  Ditto for trying to use position.txt, as that also 
contaminates the zero's this sets.

IMO its a BUG, and for me a BIG BUG, requiring that I first stop linuxcnc, 
nuke the position.txt AND the linuxcnc.vars files before restarting 
linuxcnc, so that it truly starts from scratch when homing.

Touching the HOME button should first NUKE any touch off offsets in effect 
at the time for the axis being homed.

It does not, so home is then relative to the &^$#@ offset.  I have broken 
tools and wrecked nearly finished parts because of this, several times.

Is there a valid reason for the current, save the offsets behavior?

Cheers, Gene
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
My web page:  is up!
My views 

%DCL-MEM-BAD, bad memory
VMS-F-PDGERS, pudding between the ears
A pen in the hand of this president is far more
dangerous than a gun in the hands of 200 million
  law-abiding citizens.

Re: [Emc-users] How about the core[X,Y] drive?

[jeremy youngs]

right but i dont see why automated homing would be excluded . just using
the panels as the touch off instead of homing switches
pretty simple no?


On Mon, Apr 15, 2013 at 2:04 PM, Viesturs Lācis wrote:

> 2013/4/15 jeremy youngs 
>
> > on the torch head , touch off in a corner on 2 pads run 5 v to torch head
> > and a debounce buffer then just touch off . should be more than sufficent
> > for a plasma table , if a thc is used could just use the touch off pin
> > output  no??
>
>
> Well, if I understand Your correctly, You are suggesting manual touch-off
> procedure. What I meant something that LinuxCNC would do on its own, just
> like usual homing - just press the button.
>
> With manual touch off also jogging into limit switches would do the trick.
> One could wire both limit switches in series and link to motion.probe-input
> and then do probing move along each axis. Creating vcp button to do this
> for each axis is simple.
>
> --
> Viesturs
>
> If you can't fix it, you don't own it.
> http://www.ifixit.com/Manifesto
>
> --
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 
jeremy youngs
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Viesturs Lācis]

2013/4/15 jeremy youngs 

> on the torch head , touch off in a corner on 2 pads run 5 v to torch head
> and a debounce buffer then just touch off . should be more than sufficent
> for a plasma table , if a thc is used could just use the touch off pin
> output  no??


Well, if I understand Your correctly, You are suggesting manual touch-off
procedure. What I meant something that LinuxCNC would do on its own, just
like usual homing - just press the button.

With manual touch off also jogging into limit switches would do the trick.
One could wire both limit switches in series and link to motion.probe-input
and then do probing move along each axis. Creating vcp button to do this
for each axis is simple.

-- 
Viesturs

If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[jeremy youngs]

on the torch head , touch off in a corner on 2 pads run 5 v to torch head
and a debounce buffer then just touch off . should be more than sufficent
for a plasma table , if a thc is used could just use the touch off pin
output  no??


On Mon, Apr 15, 2013 at 1:46 PM, Viesturs Lācis wrote:

> 2013/4/15 jeremy youngs 
>
> > contact plates like gene uses?
> >
> >
> On a belt?
>
> --
> Viesturs
>
> If you can't fix it, you don't own it.
> http://www.ifixit.com/Manifesto
>
> --
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 
jeremy youngs
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Viesturs Lācis]

2013/4/15 jeremy youngs 

> contact plates like gene uses?
>
>
On a belt?

-- 
Viesturs

If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[jeremy youngs]

contact plates like gene uses?


On Mon, Apr 15, 2013 at 1:32 PM, Todd Zuercher wrote:

> The best way I can think of would be to home each joint to some sort of
> index mark on the belt.
>
> - Original Message -
> 2013/4/15 andy pugh 
>
> > On 15 April 2013 11:00, Gregg Eshelman  wrote:
> >
> > > The crossover counters any offset push at one end of the gantry with an
> > equal pull in the same direction at the opposite end. The theory page
> shows
> > how it works.
> >
> > I have stared at the diagrams for long enough now that I have changed
> > my mind several times. I might have to build one to see.
> >
> > > Does LinuxCNC have H-bot support? CoreXY should drive identically.
> >
> > Not built-in, but:
> >
> >
> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/26302-cnc-avec-systeme-h-bot-ou-corexy
>
>
> Hmm, it is not the first time I see links to that Russian forum and with
> every time I get stronger and stronger impression that those LinuxCNC users
> that hang out there is a completely separate community...
>
> Anyway, I just cannot figure out, how to set up homing routine to home such
> a machine to some switches.
>
> --
> Viesturs
>
> If you can't fix it, you don't own it.
> http://www.ifixit.com/Manifesto
>
> --
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>
> --
> 
>
> Todd Zuercher
> mailto:zuerc...@embarqmail.com
>
> 
>
>
> --
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 
jeremy youngs
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Todd Zuercher]

The best way I can think of would be to home each joint to some sort of index 
mark on the belt.

- Original Message -
2013/4/15 andy pugh 

> On 15 April 2013 11:00, Gregg Eshelman  wrote:
>
> > The crossover counters any offset push at one end of the gantry with an
> equal pull in the same direction at the opposite end. The theory page shows
> how it works.
>
> I have stared at the diagrams for long enough now that I have changed
> my mind several times. I might have to build one to see.
>
> > Does LinuxCNC have H-bot support? CoreXY should drive identically.
>
> Not built-in, but:
>
> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/26302-cnc-avec-systeme-h-bot-ou-corexy


Hmm, it is not the first time I see links to that Russian forum and with
every time I get stronger and stronger impression that those LinuxCNC users
that hang out there is a completely separate community...

Anyway, I just cannot figure out, how to set up homing routine to home such
a machine to some switches.

-- 
Viesturs

If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

-- 


Todd Zuercher
mailto:zuerc...@embarqmail.com



--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Viesturs Lācis]

2013/4/15 andy pugh 

> On 15 April 2013 11:00, Gregg Eshelman  wrote:
>
> > The crossover counters any offset push at one end of the gantry with an
> equal pull in the same direction at the opposite end. The theory page shows
> how it works.
>
> I have stared at the diagrams for long enough now that I have changed
> my mind several times. I might have to build one to see.
>
> > Does LinuxCNC have H-bot support? CoreXY should drive identically.
>
> Not built-in, but:
>
> http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/26302-cnc-avec-systeme-h-bot-ou-corexy


Hmm, it is not the first time I see links to that Russian forum and with
every time I get stronger and stronger impression that those LinuxCNC users
that hang out there is a completely separate community...

Anyway, I just cannot figure out, how to set up homing routine to home such
a machine to some switches.

-- 
Viesturs

If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Eric Keller]

On Mon, Apr 15, 2013 at 6:47 AM, andy pugh  wrote:

>
>
> I have stared at the diagrams for long enough now that I have changed
> my mind several times. I might have to build one to see.
>

I have a plotter that was built like this using wire as the drive.  It was
pretty neat, but the wire got messed up and then it was a pain to drive
it.  Always thought it was a good idea, but a little weak
--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[andy pugh]

On 15 April 2013 11:00, Gregg Eshelman  wrote:

> The crossover counters any offset push at one end of the gantry with an equal 
> pull in the same direction at the opposite end. The theory page shows how it 
> works.

I have stared at the diagrams for long enough now that I have changed
my mind several times. I might have to build one to see.

> Does LinuxCNC have H-bot support? CoreXY should drive identically.

Not built-in, but:
http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/26302-cnc-avec-systeme-h-bot-ou-corexy

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

--- On Mon, 4/15/13, andy pugh  wrote:

> On 15 April 2013 10:13, Gregg Eshelman 
> wrote:
> > This looks like it ought to be able to use the same
> commands as an H-Bot.
> >
> > http://www.corexy.com
> 
> I am not sure how it differs from the H-bot (which I think
> has a
> rather simpler belt path).
> 
> As far as I can see both designs rely on the "gantry" being
> inherently
> stiff against racking, as that is where the X reaction force
> comes from.

The crossover counters any offset push at one end of the gantry with an equal 
pull in the same direction at the opposite end. The theory page shows how it 
works.

Having a very light gantry bar on a tight cable or belt would allow it to be 
lifted enough to be picked up off the top of the torch arm so the torch could 
be used for both CNC and pattern following without any changes of hardware or 
anything else to switch. Pick it up, move the arm out a little, let go and jog 
the gantry out of the way. Then a pattern can be mounted to run around.

Does LinuxCNC have H-bot support? CoreXY should drive identically.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] How about the core[X,Y] drive?

[andy pugh]

On 15 April 2013 10:13, Gregg Eshelman  wrote:
> This looks like it ought to be able to use the same commands as an H-Bot.
>
> http://www.corexy.com

I am not sure how it differs from the H-bot (which I think has a
rather simpler belt path).

As far as I can see both designs rely on the "gantry" being inherently
stiff against racking, as that is where the X reaction force comes
from.

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] How about the core[X,Y] drive?

[Gregg Eshelman]

This looks like it ought to be able to use the same commands as an H-Bot.

http://www.corexy.com

Two motors, differential turning moves the platform in X and Y. Hmmm, this 
looks pretty much exactly how an Etch-a-Sketch works. The "gantry" bar can be 
lightweight for a light load (like the end of a pattern torch arm) and it can 
float without need of guide/support rails at the ends.

I could make a U shaped support frame, only need two motors at the ends of the 
U. The gantry could be a simple piece of aluminum flat bar and the carriage 
just a slider with a ball bearing with an ID big enough to fit over the drive 
magnet.

Need to use it original style? Simply lift the bar up off the arm and run it 
back to the support post.

--
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users