Re: [Emc-users] This ones for PCW

[Gene Heskett]

On Saturday 12 September 2015 01:39:46 TJoseph Powderly wrote:

> On 09/11/2015 10:50 PM, Gene Heskett wrote:
> 
>
> > I was feathered, sorta resembling a chicken, so I called in a scale
> > module, which once configured worked really well.
> >
> > But in the 2.8.0-pre's I have a bitch, a loud one.  The halmeter is
> > shown forground for as long as it exists, and that is nice, very
> > nice.
> >
> > But neither halscope nor the hal show config do.  I had halscope
> > watching some stuff an external scope cannot access.  And everytime
> > I clicked on the mdi line to try a spindle speed change today, all
> > that stuff went behind the axis screen, rendering what you wanted to
> > watch invisible because you simply cannot get it re-selected in time
> > to see the response the system made from the command you typed.
>
> Gene, can you right click on top bar of Halscope and choose 'Always on
> top"? 

I didn't know one could do that, but I'll sure look at it when I am next 
at the machine.

> and if you change desktops, try "Always on visible workspace" 

I generally expect the other 3 I have configured to remain "uncluttered".

> you should be able to use the other windows like the mdi line
> without loosing sight of the Scope.

Or of the showconfig panel. being able to click focus between the mdi 
line and the command line of showconfig would also facilitate not losing 
track of ones train of thought.  So having that "always on top" would 
also help.

> The developers have to work inside the desktop environment that you
> use. Things like fonts and colors and whos on top are governed by
> other rules. If the devs take over that stuff they may borq something
> else. Its really your job to choose how your desktop is laid out. You
> do have the power to do so.
>
> HTH TomP tjtr33

Yes it does TomP, thanks.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

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

Re: [Emc-users] This ones for PCW

[andy pugh]

On 11 September 2015 at 05:33, Gene Heskett  wrote:
> One hopefully last question then Andy, can hal or axis do the math such
> that there is one master place to specify this "scale", and set the
> pid's maxoutput to 0.98 * scale?


No, but yes.

Mathematical expressions are not allowed in a HAL file.
Unless the HAL file is a TCL HAL file...
http://linuxcnc.org/docs/html/hal/haltcl.html

-- 
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] This ones for PCW

[TJoseph Powderly]

On 09/11/2015 03:59 AM, andy pugh wrote:
> On 11 September 2015 at 05:33, Gene Heskett  wrote:
>> One hopefully last question then Andy, can hal or axis do the math such
>> that there is one master place to specify this "scale", and set the
>> pid's maxoutput to 0.98 * scale?
>
>
> No, but yes.
>
> Mathematical expressions are not allowed in a HAL file.
> Unless the HAL file is a TCL HAL file...
> http://linuxcnc.org/docs/html/hal/haltcl.html
>
the 1st line in the referenced url says the halcmd language
"offers no computational capabilities"

which is a bit confusing
( slight of hand, it says halcmd not haltcl )

but certainly the later example
shows the haltcl language can do math
"""
setp scale.0.gain "6.28 / 360.0 * $::AXIS_0(radius) * 60.0"
"""

I'd have to try the section on auto-converting
from .hal to .tcl by simple find
Hard to believe its sufficient, but I bet its useful beginning.
I'm not from Missouri but E.StLouis was just across the river ;)
( who's afraid of the big bad tickle? :)
TomP tjtr33

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

Re: [Emc-users] This ones for PCW

[TJoseph Powderly]

On 09/11/2015 10:50 PM, Gene Heskett wrote:

>
> I was feathered, sorta resembling a chicken, so I called in a scale
> module, which once configured worked really well.
>
> But in the 2.8.0-pre's I have a bitch, a loud one.  The halmeter is shown
> forground for as long as it exists, and that is nice, very nice.
>
> But neither halscope nor the hal show config do.  I had halscope watching
> some stuff an external scope cannot access.  And everytime I clicked on
> the mdi line to try a spindle speed change today, all that stuff went
> behind the axis screen, rendering what you wanted to watch invisible
> because you simply cannot get it re-selected in time to see the response
> the system made from the command you typed.
>
Gene, can you right click on top bar of Halscope and choose 'Always on 
top"? and if you change desktops, try "Always on visible workspace"
you should be able to use the other windows like the mdi line
without loosing sight of the Scope.
The developers have to work inside the desktop environment that you use.
Things like fonts and colors and whos on top are governed by other 
rules. If the devs take over that stuff they may borq something else.
Its really your job to choose how your desktop is laid out. You do have 
the power to do so.

HTH TomP tjtr33

> IMO thats a bug big enough to get out my Ackley-06 & do some target
> shooting.  I do not recall that it was a problem 2 years ago when I was
> doing the same spindle tuning exercises.  Please fix that I beg you.


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

Re: [Emc-users] This ones for PCW

[Gene Heskett]

On Friday 11 September 2015 04:59:55 andy pugh wrote:

> On 11 September 2015 at 05:33, Gene Heskett  wrote:
> > One hopefully last question then Andy, can hal or axis do the math
> > such that there is one master place to specify this "scale", and set
> > the pid's maxoutput to 0.98 * scale?
>
> No, but yes.
>
> Mathematical expressions are not allowed in a HAL file.
> Unless the HAL file is a TCL HAL file...
> http://linuxcnc.org/docs/html/hal/haltcl.html

Something I haven't played with, yet.

However, after hacking around in the hal file to get rid of a pin does 
not exist error, I am back to an error that is a showstopper for running 
it, on that machine, but sending the x display stuff back to this 
machine.

Logged into "lathe" with an ssh -Y session, an attempt to 
run 'linuxcnc -l' stumbles over, and exists with this error:

_tkinter.TclError: unknown color name "BACKGROUND"
Shutting down and cleaning up LinuxCNC...

I think we hacked up something in the environment on this machine to fix 
those two errors (there will also be an unknown color name "FOREGROUND" 
error to deal with when we define BACKGROUND) but that was pre-wheezy on 
this machine.  Do you recall how to fix that?

FWIW, other stuff that uses X Just Works(TM), like geany or 
synaptic-pkexec, have no such problem.

Thanks Andy

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

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

Re: [Emc-users] This ones for PCW

[Gene Heskett]

On Friday 11 September 2015 10:01:34 TJoseph Powderly wrote:

> On 09/11/2015 03:59 AM, andy pugh wrote:
> > On 11 September 2015 at 05:33, Gene Heskett  
wrote:
> >> One hopefully last question then Andy, can hal or axis do the math
> >> such that there is one master place to specify this "scale", and
> >> set the pid's maxoutput to 0.98 * scale?
> >
> > No, but yes.
> >
> > Mathematical expressions are not allowed in a HAL file.
> > Unless the HAL file is a TCL HAL file...
> > http://linuxcnc.org/docs/html/hal/haltcl.html
>
> the 1st line in the referenced url says the halcmd language
> "offers no computational capabilities"
>
> which is a bit confusing
> ( slight of hand, it says halcmd not haltcl )
>
> but certainly the later example
> shows the haltcl language can do math
> """
> setp scale.0.gain "6.28 / 360.0 * $::AXIS_0(radius) * 60.0"
> """
>
> I'd have to try the section on auto-converting
> from .hal to .tcl by simple find
> Hard to believe its sufficient, but I bet its useful beginning.
> I'm not from Missouri but E.StLouis was just across the river ;)
> ( who's afraid of the big bad tickle? :)
> TomP tjtr33

I was feathered, sorta resembling a chicken, so I called in a scale 
module, which once configured worked really well.

But in the 2.8.0-pre's I have a bitch, a loud one.  The halmeter is shown 
forground for as long as it exists, and that is nice, very nice.

But neither halscope nor the hal show config do.  I had halscope watching 
some stuff an external scope cannot access.  And everytime I clicked on 
the mdi line to try a spindle speed change today, all that stuff went 
behind the axis screen, rendering what you wanted to watch invisible 
because you simply cannot get it re-selected in time to see the response 
the system made from the command you typed.

IMO thats a bug big enough to get out my Ackley-06 & do some target 
shooting.  I do not recall that it was a problem 2 years ago when I was 
doing the same spindle tuning exercises.  Please fix that I beg you.

But while I wasn't able to use a lot of Pgain, 2.5 or so seemed fairly 
stable, but Igain at 15.00 it tolerated well enough I was able to get 
some of the work I wanted to do done.  With those changes, the spindle 
speed is being disturbed by a slightly eccentric encoder disk, but its 
stiff enough under load that the load does not show on the speed slider, 
even at 100 revs in high backgear.

But Murphy is still here. I get it together, and some of the hal changes 
so even with that heavy a chuck, I can from 400 revs, type m4, m3, m4, 
m3 and turn the rotation around in a couple seconds, which is much 
faster than before, 4x faster maybe.  But I was hearing some sound 
effects, it seems one of the jackshaft bearings is not entirely happy, 
sounding like an inner race is squeeking on the shaft.  Anything above 
200 revs and its doing the intermittent squeek, occasionaly even at 100 
revs.

And the darned things are supposedly glued pretty solid with green, 
penetrating permagoo thread-locker.  Hopefully I can change that drive 
belt for a longer one that will reduce the bearing load.  The 6 rib j 
sized belt can be played like Jack Benny's violin.

Which makes me say that vbeltsupply in Sioux Falls SD is NOT a good place 
to get product.  At the price ($14) they charge for shipping a $4 belt, 
I figured 2 days should get it here.  So I called them.  Nope, not till 
at least the 15th.  Those switching buck regulator modules from 
someplace I never heard of in China, ordered last thursday, came 
yesterday, free shipping via China Post.  And the one I put in already 
works flawlessly.

China I hate to say it, is eating our Merican suppliers lunch.  And doing 
it better, faster, and cheaper.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

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

Re: [Emc-users] This ones for PCW

[andy pugh]

On 11 September 2015 at 00:56, Gene Heskett  wrote:
> So if I am using a scale factor there, to tame the axis big jumps when
> clicking on the + and - buttons, it seems to me that the value I apply
> to a limit2 in front of your 5i25 pwmgen should also be scaled up by
> this same scale factor.

I think you should use RPM all the way to the pwmgen and limit the pid
to 98% of the pwmgen.scale

The alternative is to divide motion.spindle-speed-out by max spindle
speed if you insist that you want to work on a 0 to 0.98 scale.

-- 
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] This ones for PCW

[andy pugh]

On 11 September 2015 at 01:44, andy pugh  wrote:
> I think you should use RPM all the way to the pwmgen and limit the pid
> to 98% of the pwmgen.scale

Actually, given that encoder velocity is returned in RPS then use RPS
all the way, from motion.spindle-speed-out-rps all the way through the
pid to the pwmgen.
Limit the pid max output to 98% of the pwmgen.scale parameter.


-- 
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] This ones for PCW

[Gene Heskett]

On Thursday 10 September 2015 20:46:50 andy pugh wrote:

> On 11 September 2015 at 01:44, andy pugh  wrote:
> > I think you should use RPM all the way to the pwmgen and limit the
> > pid to 98% of the pwmgen.scale
>
> Actually, given that encoder velocity is returned in RPS then use RPS
> all the way, from motion.spindle-speed-out-rps all the way through the
> pid to the pwmgen.
> Limit the pid max output to 98% of the pwmgen.scale parameter.

And that will get rid of the limit2 to do that I believe.

Thanks Andy.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

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

Re: [Emc-users] This ones for PCW

[Gene Heskett]

On Thursday 10 September 2015 22:10:50 Gene Heskett wrote:

> On Thursday 10 September 2015 20:46:50 andy pugh wrote:
> > On 11 September 2015 at 01:44, andy pugh  wrote:
> > > I think you should use RPM all the way to the pwmgen and limit the
> > > pid to 98% of the pwmgen.scale
> >
> > Actually, given that encoder velocity is returned in RPS then use
> > RPS all the way, from motion.spindle-speed-out-rps all the way
> > through the pid to the pwmgen.
> > Limit the pid max output to 98% of the pwmgen.scale parameter.
>
> And that will get rid of the limit2 to do that I believe.
>
> Thanks Andy.
>
> Cheers, Gene Heskett

One hopefully last question then Andy, can hal or axis do the math such 
that there is one master place to specify this "scale", and set the 
pid's maxoutput to 0.98 * scale?  Or do I have to loadrt a scale module, 
addf it at a suitable time, setp its gain to 0.98, set its input to the 
same inif  files S_SCALE, and feed its output to pid.s.maxoutput?

I can do either of course, but the math is a one time thing at init time 
whereas the addf refreshes it, wasting a few cpu cycles every 
servo-thread.

But I just did the other using a scale module.  I'll see how it works 
tomorrow.

Thanks Andy

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

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