ed | 21 Oct 03:44 2012
Picon

G code problem with G95

Finally got my Hardinge CHNC lathe up and mostly running and am having a 
problem. Do you simply put G95 on a line then on the next G1 make sure 
there is a F feedrate per rev?  What ever I try the prog stops with the 
spindle running on the first G1 line.  Baffling.

Maybe someone has a small code snip that works so I can follow the code 
path.

Thanks, Ed

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
Matt Shaver | 21 Oct 17:12 2012

Re: G code problem with G95

On Sat, 20 Oct 2012 20:44:02 -0500
ed <atex57@...> wrote:

> Finally got my Hardinge CHNC lathe up and mostly running and am
> having a problem. Do you simply put G95 on a line then on the next G1
> make sure there is a F feedrate per rev?  What ever I try the prog
> stops with the spindle running on the first G1 line.  Baffling.
> 
> Maybe someone has a small code snip that works so I can follow the
> code path.

Thanks for confirming what I have been thinking - That G95 is broken
somehow. I consult for Smithy who sells CNC lathes with Linuxcnc
control, and our lathe customers report this same behavior.
Interestingly G76 threading works OK, and it depends upon proper
operation of the index pulse logic and correct spindle direction and
velocity feedback. Is there some other hardware interface or software
configuration issue that prevents G95 from working? Or is it really
broke?

What can be done to test this? Any signals that need to be monitored in
a halmeter? I'd like to see this fixed, but I haven't reported it yet
as I didn't feel I had enough information to support a bug report. I'm
pretty sure this is a current problem with the latest release.

Thanks,
Matt

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
(Continue reading)

ed | 21 Oct 18:26 2012
Picon

Re: G code problem with G95

Matt Shaver wrote:
> On Sat, 20 Oct 2012 20:44:02 -0500
> ed <atex57@...> wrote:
>
>   
>> Finally got my Hardinge CHNC lathe up and mostly running and am
>> having a problem. Do you simply put G95 on a line then on the next G1
>> make sure there is a F feedrate per rev?  What ever I try the prog
>> stops with the spindle running on the first G1 line.  Baffling.
>>
>> Maybe someone has a small code snip that works so I can follow the
>> code path.
>>     
>
> Thanks for confirming what I have been thinking - That G95 is broken
> somehow. I consult for Smithy who sells CNC lathes with Linuxcnc
> control, and our lathe customers report this same behavior.
> Interestingly G76 threading works OK, and it depends upon proper
> operation of the index pulse logic and correct spindle direction and
> velocity feedback. Is there some other hardware interface or software
> configuration issue that prevents G95 from working? Or is it really
> broke?
>
>   

I don't think the spindle encoder is relevant unless you have closed 
loop spindle control. My guess is it only looks at the S  word rpm's to 
calculate feed rate.

Ed
(Continue reading)

jeremy youngs | 21 Oct 18:31 2012
Picon

Re: G code problem with G95

>I don't think the spindle encoder is relevant unless you have closed
loop spindle control. My guess is it only looks at the S  word rpm's to
calculate feed rate.<

then how would it know where to index the cut at?

--

-- 
jeremy youngs

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
Chris Radek | 21 Oct 18:44 2012

Re: G code problem with G95

On Sun, Oct 21, 2012 at 12:31:26PM -0400, jeremy youngs wrote:
> >I don't think the spindle encoder is relevant unless you have closed
> loop spindle control. My guess is it only looks at the S  word rpm's to
> calculate feed rate.<
> 
> then how would it know where to index the cut at?

G95 is not used for threading - it's velocity based, not position mode
sync.

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
Chris Radek | 21 Oct 18:43 2012

Re: G code problem with G95

On Sun, Oct 21, 2012 at 11:26:52AM -0500, ed wrote:
> 
> I don't think the spindle encoder is relevant unless you have closed 
> loop spindle control. My guess is it only looks at the S  word rpm's to 
> calculate feed rate.

Your guess is incorrect - thanks for labeling it as a guess, so as
not to confuse people later.

You can configure it that way with a loopback of spindle-speed-out
(desired) => spindle-speed-in (actual) but if you already have an
encoder, you can do better by using actual.

http://linuxcnc.org/docs/html/man/man9/motion.9.html

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
Chris Radek | 21 Oct 18:40 2012

Re: G code problem with G95

On Sun, Oct 21, 2012 at 11:12:12AM -0400, Matt Shaver wrote:
> 
> Thanks for confirming what I have been thinking - That G95 is broken
> somehow. I consult for Smithy who sells CNC lathes with Linuxcnc
> control, and our lathe customers report this same behavior.
> Interestingly G76 threading works OK, and it depends upon proper
> operation of the index pulse logic and correct spindle direction and
> velocity feedback. Is there some other hardware interface or software
> configuration issue that prevents G95 from working? Or is it really
> broke?

G76 uses spindle position (spindle-revs).  G95/G96 use spindle
velocity (spindle-speed-in).  Are you sure you've hooked up both?
See the motion.spindle* section of 

http://linuxcnc.org/docs/html/man/man9/motion.9.html

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
Matt Shaver | 24 Oct 22:39 2012

Re: G code problem with G95

On Sun, 21 Oct 2012 11:40:15 -0500
Chris Radek <chris@...> wrote:

> G76 uses spindle position (spindle-revs).  G95/G96 use spindle
> velocity (spindle-speed-in).  Are you sure you've hooked up both?
> See the motion.spindle* section of 
> 
> http://linuxcnc.org/docs/html/man/man9/motion.9.html

I think this is the problem!

I had:
net Spos hm2_5i20.0.encoder.00.position motion.spindle-revs

I added:
net Svel hm2_5i20.0.encoder.00.velocity motion.spindle-speed-in

It'll be a couple days before I can have this tested, but I'm
optimistic!

Thanks Chris!
Matt

P.S. I think Ed could do (in lieu of an actual spindle encoder):
net Svel motion.spindle-speed-out-rps motion.spindle-speed-in

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
(Continue reading)

andy pugh | 21 Oct 19:23 2012
Picon

Re: G code problem with G95

On 21 October 2012 16:12, Matt Shaver <matt@...> wrote:

> Thanks for confirming what I have been thinking - That G95 is broken
> somehow.

My lathe was running 2.5, and G95 worked as expected.
Deleting the link to motion.spindle-speed-in stopped it working, however.

I then upgraded to 2.5.1 through the package manager, and G95 still works.
Again, deleting the signal into motion.spindle-speed-in stopped it
dead in its tracks.

--

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

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
ed | 21 Oct 20:43 2012
Picon

Re: G code problem with G95

andy pugh wrote:
> On 21 October 2012 16:12, Matt Shaver <matt@...> wrote:
>
>   
>> Thanks for confirming what I have been thinking - That G95 is broken
>> somehow.
>>     
>
> My lathe was running 2.5, and G95 worked as expected.
> Deleting the link to motion.spindle-speed-in stopped it working, however.
>
> I then upgraded to 2.5.1 through the package manager, and G95 still works.
> Again, deleting the signal into motion.spindle-speed-in stopped it
> dead in its track
>   

I am running 2.51 at this time. Could you tell what the links are?

Searching through my HAL files I do not find any connections to 
"motion.spindle" anything. Most of what I am using comes from Jon's Pico 
Systems site and cobbled together.

Ed.

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
(Continue reading)

andy pugh | 21 Oct 20:52 2012
Picon

Re: G code problem with G95

On 21 October 2012 19:43, ed <atex57@...> wrote:

> I am running 2.51 at this time. Could you tell what the links are?

http://www.linuxcnc.org/docs/2.4/html/examples_spindle.html#sec:Spindle-Synchronized-Motion

You need to link motion.spindle-speed-in to something (possibly
motion.spindle-speed-out)

--

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

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct

Gmane