Re: [Emc-users] machine motion

[Dave Caroline]

I have seen this on my machine which still has 2.4 on, it is on a
machine with not enough ram. I make sure unwanted background stuff is
not running to avoid it.

Dave

On 29/10/2015, Stuart Stevenson  wrote:
> Gentlemen,
>   re: the cinci at MPM
> When they push the arrow key to move an axis the machine keeps moving. A
> couple weeks ago the X axis was the only one doing this but now other axes
> demonstrate this feature.
> They disconnect the keyboard, use the mouse to move the axes with using the
> +/- buttons and get the same result.
> What should the next focus be to troubleshoot this?
>
> Ubuntu 8.04
> LinuxCNC 2.4.?
> Axis Display
>
> I know, I know - it needs upgraded. Not possible at this time.
>
> thanks
> Stuart
>
> --
> Addressee is the intended audience.
> If you are not the addressee then my consent is not given for you to read
> this email furthermore it is my wish you would close this without saving or
> reading, and cease and desist from saving or opening my private
> correspondence.
> Thank you for honoring my wish.
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] machine motion

[Stuart Stevenson]

Gentlemen,
  re: the cinci at MPM
When they push the arrow key to move an axis the machine keeps moving. A
couple weeks ago the X axis was the only one doing this but now other axes
demonstrate this feature.
They disconnect the keyboard, use the mouse to move the axes with using the
+/- buttons and get the same result.
What should the next focus be to troubleshoot this?

Ubuntu 8.04
LinuxCNC 2.4.?
Axis Display

I know, I know - it needs upgraded. Not possible at this time.

thanks
Stuart

-- 
Addressee is the intended audience.
If you are not the addressee then my consent is not given for you to read
this email furthermore it is my wish you would close this without saving or
reading, and cease and desist from saving or opening my private
correspondence.
Thank you for honoring my wish.
--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[sam sokolik]

or a failing disk...

On 10/29/2015 2:07 PM, andy pugh wrote:
> On 29 October 2015 at 18:53, Stuart Stevenson  wrote:
>
>> Shannon at MPM will change the ram on the machine. He said it has been
>> acting SLOW and strange lately.
> My recent experience with Linux is that it copes really inelegantly to
> a full disk, so it might be worth checking disk space.
>


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[Stuart Stevenson]

On Thu, Oct 29, 2015 at 1:23 PM, Sebastian Kuzminsky 
wrote:

> On 10/29/2015 11:48 AM, Stuart Stevenson wrote:
> > Gentlemen,
> >   re: the cinci at MPM
> > When they push the arrow key to move an axis the machine keeps moving. A
> > couple weeks ago the X axis was the only one doing this but now other
> axes
> > demonstrate this feature.
> > They disconnect the keyboard, use the mouse to move the axes with using
> the
> > +/- buttons and get the same result.
> > What should the next focus be to troubleshoot this?
> >
> > Ubuntu 8.04
> > LinuxCNC 2.4.?
> > Axis Display
>
> You say: after you release the key, the machine keeps moving.  Does it
> stop moving a second or two later?
>
I am told it does not stop until the end of travel


>
> Both of those jogging mechanisms (keyboard and GUI-button) are of the
> non-realtime "UI to Task to Motion" variety.  The non-realtime parts of
> that chain (GUI & Task) can have unpredictable timing, resulting in
> delayed commands getting to Motion, resulting in jogs that go on for a
> short while after the key is released.
>
> As Dave Caroline said, one option is to lighten the load on the system
> so that more resources are available for Axis and Task.  You can run
> "vmstat 1" in a terminal to see if the machine is swapping or paging,
> which would suggest adding RAM or killing background processes.
>
> The best option is to jog with a hardware jog-wheel connected directly
> to Motion via HAL.  That bypasses the problematic non-realtime stuff
> entirely.
>
> Unfortunately Motion does not have continuous-jog inputs (which is what
> the keyboard and the GUI + and - buttons do), but i bet it wouldn't be
> hard to add (he said, not having looked at the code).
>
Shannon at MPM will change the ram on the machine. He said it has been
acting SLOW and strange lately. When I mentioned Dave's comment Shannon
agreed instantly that could be the problem.
We will see.
thanks guys for the quick help and insight
Stuart



>
>
> --
> Sebastian Kuzminsky
>
>
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 
Addressee is the intended audience.
If you are not the addressee then my consent is not given for you to read
this email furthermore it is my wish you would close this without saving or
reading, and cease and desist from saving or opening my private
correspondence.
Thank you for honoring my wish.
--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[Stuart Stevenson]

Ok,

new information

new ram installed
new ppmc boards
no change in machine motion on the Y axis - the other axes run fine (they
might have been fine before and it may have been my misunderstanding).
once y motion has been started the axis moves until - A. estop  B. motion
change by pressing the opposite motion button C. sometimes the axis will
stop on the motion direction change but most of the time it will just
change direction and continue the motion in the opposite direction.

I told him to type df -h in a terminal to check the hard drive

thanks
Stuart


On Thu, Oct 29, 2015 at 2:15 PM, sam sokolik  wrote:

> or a failing disk...
>
> On 10/29/2015 2:07 PM, andy pugh wrote:
> > On 29 October 2015 at 18:53, Stuart Stevenson  wrote:
> >
> >> Shannon at MPM will change the ram on the machine. He said it has been
> >> acting SLOW and strange lately.
> > My recent experience with Linux is that it copes really inelegantly to
> > a full disk, so it might be worth checking disk space.
> >
>
>
>
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 
Addressee is the intended audience.
If you are not the addressee then my consent is not given for you to read
this email furthermore it is my wish you would close this without saving or
reading, and cease and desist from saving or opening my private
correspondence.
Thank you for honoring my wish.
--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[Sebastian Kuzminsky]

On 10/29/2015 11:48 AM, Stuart Stevenson wrote:
> Gentlemen,
>   re: the cinci at MPM
> When they push the arrow key to move an axis the machine keeps moving. A
> couple weeks ago the X axis was the only one doing this but now other axes
> demonstrate this feature.
> They disconnect the keyboard, use the mouse to move the axes with using the
> +/- buttons and get the same result.
> What should the next focus be to troubleshoot this?
> 
> Ubuntu 8.04
> LinuxCNC 2.4.?
> Axis Display

You say: after you release the key, the machine keeps moving.  Does it
stop moving a second or two later?

Both of those jogging mechanisms (keyboard and GUI-button) are of the
non-realtime "UI to Task to Motion" variety.  The non-realtime parts of
that chain (GUI & Task) can have unpredictable timing, resulting in
delayed commands getting to Motion, resulting in jogs that go on for a
short while after the key is released.

As Dave Caroline said, one option is to lighten the load on the system
so that more resources are available for Axis and Task.  You can run
"vmstat 1" in a terminal to see if the machine is swapping or paging,
which would suggest adding RAM or killing background processes.

The best option is to jog with a hardware jog-wheel connected directly
to Motion via HAL.  That bypasses the problematic non-realtime stuff
entirely.

Unfortunately Motion does not have continuous-jog inputs (which is what
the keyboard and the GUI + and - buttons do), but i bet it wouldn't be
hard to add (he said, not having looked at the code).


-- 
Sebastian Kuzminsky

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[andy pugh]

On 29 October 2015 at 18:53, Stuart Stevenson  wrote:

> Shannon at MPM will change the ram on the machine. He said it has been
> acting SLOW and strange lately.

My recent experience with Linux is that it copes really inelegantly to
a full disk, so it might be worth checking disk space.

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

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[John Kasunich]

My understanding of keyboard jogs is that a "key pressed" event for the up-arrow
stops any active Y axis jog, and then starts a +Y jog.  A "key released" event 
for 
the up-arrow stops any active Y axis jog.  Likewise for the down-arrow.

It sounds like LinuxCNC isn't seeing the "key released" events.

Does hitting the escape key stop one of these "runaway jogs"?  The escape
key is supposed to generate an "ABORT" command to motion, which should
stop the movement.  Whether or not that happens might be a clue.





On Thu, Oct 29, 2015, at 03:41 PM, Stuart Stevenson wrote:
> Ok,
> 
> new information
> 
> new ram installed
> new ppmc boards
> no change in machine motion on the Y axis - the other axes run fine (they
> might have been fine before and it may have been my misunderstanding).
> once y motion has been started the axis moves until - A. estop  B. motion
> change by pressing the opposite motion button C. sometimes the axis will
> stop on the motion direction change but most of the time it will just
> change direction and continue the motion in the opposite direction.
> 
> I told him to type df -h in a terminal to check the hard drive
> 
> thanks
> Stuart
> 
> 
> On Thu, Oct 29, 2015 at 2:15 PM, sam sokolik  wrote:
> 
> > or a failing disk...
> >
> > On 10/29/2015 2:07 PM, andy pugh wrote:
> > > On 29 October 2015 at 18:53, Stuart Stevenson  wrote:
> > >
> > >> Shannon at MPM will change the ram on the machine. He said it has been
> > >> acting SLOW and strange lately.
> > > My recent experience with Linux is that it copes really inelegantly to
> > > a full disk, so it might be worth checking disk space.
> > >
> >
> >
> >
> > --
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
> 
> 
> 
> -- 
> Addressee is the intended audience.
> If you are not the addressee then my consent is not given for you to read
> this email furthermore it is my wish you would close this without saving or
> reading, and cease and desist from saving or opening my private
> correspondence.
> Thank you for honoring my wish.
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


-- 
  John Kasunich
  jmkasun...@fastmail.fm

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[Jon Elson]

On 10/29/2015 01:53 PM, Stuart Stevenson wrote:
> Shannon at MPM will change the ram on the machine. He said it has been
> acting SLOW and strange lately. When I mentioned Dave's comment Shannon
> agreed instantly that could be the problem.
>
Hmmm, has the computer been rebooted?  It could just have 
some semi-crashed service burning CPU time.
That could make the GUI real laggy.  Like VMS (a DEC OS) 
Linux will try to run with a failed hard drive, just running 
off what is in the disk cache (which is all of the unused 
memory).  Slowly, normal utilities will hang, and the system 
will get more and more stuck processes until you can't do 
anything.  When you try to reboot, it will either clear the 
fault or not boot.

Jon

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[Jon Elson]

On 10/29/2015 12:48 PM, Stuart Stevenson wrote:
> Gentlemen,
>re: the cinci at MPM
> When they push the arrow key to move an axis the machine keeps moving. A
> couple weeks ago the X axis was the only one doing this but now other axes
> demonstrate this feature.
> They disconnect the keyboard, use the mouse to move the axes with using the
> +/- buttons and get the same result.
> What should the next focus be to troubleshoot this?
>
> Ubuntu 8.04
> LinuxCNC 2.4.?
> Axis Display
>
> I know, I know - it needs upgraded. Not possible at this time.
>
>
OK, I had something like this, and just dealt with it for a 
while. then, the keyboard totally quit.  I found out I had a 
crummy DIN to mini-DIN adapter that had a bunch of broken 
wires inside.  (Yes, that's how old some of this stuff I use 
is, 5-pin DIN keyboards. But, i have a water/chip proof 
cover on that keyboard, so I keep using it.)

I have not had the sticking key problem since repairing the 
adapter.  What is happening is the computer is sometimes 
failing to receive the "key up" message, so it doesn't know 
the key has been released.

If you are sure the on-screen jog buttons do the same, then 
that is sure a new one on me.I've never seen that.
Be aware if you unplug the keyboard in the middle of a jog, 
you WILL get the stuck jog condition for sure.

First, I'd fire up Halscope and make sure the servos are not 
doing something abnormal.  If the pid.n.error trace seems to 
be normal (ie. the servos are truly following the commanded 
position) then I'd see about replacing the whole computer 
with another one running the same version.  Just put it on a 
table to find out WHAT part of the system is doing the 
misbehavior.  Are these keyboards and mice USB or PS/2?  If 
you have a choice, try the other kind, just to see if it 
behaves different.

Also, try doing keyboard and mouse stuff NOT in LinuxCNC to 
see if you can get any misbehavior to appear there, too.

The stuck jog key effect is EASY for me to believe, as I 
have experienced it here, too.  The other is not something I 
have seen, and it seems a little harder to understand.  My 
knowledge of mouse protocols is they do NOT send key up/key 
down messages, but have a bit in the format that says the 
button is down.  So, it seems harder for that to get the 
button stuck down condition.

Jon

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] machine motion

[Jon Elson]

On 10/29/2015 02:41 PM, Stuart Stevenson wrote:
> Ok,
>
> new information
>
> new ram installed
> new ppmc boards
> no change in machine motion on the Y axis - the other axes run fine (they
> might have been fine before and it may have been my misunderstanding).
> once y motion has been started the axis moves until - A. estop  B. motion
> change by pressing the opposite motion button C. sometimes the axis will
> stop on the motion direction change but most of the time it will just
> change direction and continue the motion in the opposite direction.
>
>
OK, this is different, and I really suggest checking the Y 
axis with Halscope, looking at :
ppmc.0.encoder.01.velocity (hmmm, maybe on your older 
LinuxCNC version, there will only be delta)
pid.1.error

And see what it is doing when it runs away.  I'm guessing 
that it is either the encoder channel is doing something 
odd, like not reporting movement, or the DAC is getting left 
with a non-zero velocity when the jog command ends.
If anything strange is found, run the ppmc diags program in 
the commtest mode.  Also run the dactest mode
(If you have a very old version of the diags it won't have 
this, I can send a newer version)  this tests reliability of 
communication to the DAC board.  Possibly the parallel port 
or the IEEE-1284 cable may have a problem.

If the encoder and encoder board are working correctly, and 
the MIN_FERROR is set to something reasonable, then a servo 
divergence like this should be impossible.

Boy, this is sure a weird one, wish I was there to have a look!

Jon

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users