Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 21:36:09 Jon Elson wrote:

> On 04/10/2019 10:05 AM, Gene Heskett wrote:
> > And if I copy the ini to one with no A references in it, do I also
> > have separate .hal files, or does hal ignore the A references?  Or
> > do you have 3 hal files? A basic, and a load vertical and a load
> > horizontal?
>
> Yes, you would have an .ini and at least one .hal file for
> each setup.  The .ini file would call at least one different
> .hal file to correspond with the different setup.  You could
> create two toolbar launchers that start LinuxCNC with
> different .ini files for the two setups.  Or, just choose
> the desired one from the config selector menu.
>
> Jon

That sounds like the least after config trouble, so once I get the A back 
running, I put a motor with about 3x the holding power on it this 
afternoon, and am how looking for a gs16-5 male inline to hook its home 
switch up in removeable config, using the extra contacts for a hey I'm 
here (or not) tally.  They seem to be similar to hens teeth in 
availability, and I may be forced to make a back shell for the panel 
mount version. I have a dozen or so of that.

But I'm busy in the morning, our local daytimer finally bought a new 
transmitter to replace his 60 yo gates 1kw, got a license to goto 2500 
watts. He  bought a BE, but he asked me about a hot ferrite on the 
output transmission line today n the grocery store parking lot.  I've 
known he had an swr problem, loading out of tuning range that was 
perfect when it was installed. When he mentioned it was way too hot to 
touch, all the red flags in my mind went to the top of the pole.  
Ferrites have a very low curie temperature, some below 100C.  If excited 
when at or above the curie point, they are permanently ruined, often 
turning into so much air forever.  So I need to take my IR thermometer 
up and measure it, then call BE and ask them whats the curie point of 
the junk they used. I'd almost bet its running hotter than that, and 
that its an indicator they need to retune the antenna matcher.  In 
addition to replaceing that one with a good one.  Its target practice 
for my 6.5 Creedmoor, sure as hell.

And I don't have the rf bridge to do that. I may wind up buying one yet, 
but bring around $1500 for a modern, even draws smith charts version of 
that puppy. If I do, I'll have to drum up some business to pay for it. 

Redpitaya has such a kit, 900 euro's. 799 for the ultimate kit + 100 more 
for the network vector analyzer. Dunno what the exchange rate or the VAT 
brings it up to. Or how long to get it across the pond.

The life of a trying to retire broadcast engineer. My 1st phone is older 
than the guys designing this stuff today. It dates from 1962!

Cheers Jon, 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] Leaving an axis out of the home sequence AND (continued:)

[Jon Elson]

On 04/10/2019 10:05 AM, Gene Heskett wrote:


And if I copy the ini to one with no A references in it, do I also have
separate .hal files, or does hal ignore the A references?  Or do you
have 3 hal files? A basic, and a load vertical and a load horizontal?
Yes, you would have an .ini and at least one .hal file for 
each setup.  The .ini file would call at least one different 
.hal file to correspond with the different setup.  You could 
create two toolbar launchers that start LinuxCNC with 
different .ini files for the two setups.  Or, just choose 
the desired one from the config selector menu.


Jon


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 13:29:54 Gene Heskett wrote:

> On Wednesday 10 April 2019 12:47:23 John Dammeyer wrote:
> > Hi Andy,
> > See below.
> > [...] 
> > > > The question then is whether or not the system, when starting
> > > > can/will
> > > recognize an input which is a unit present signal.
> > >
> > > This is a solved problem :-) (I have exactly the same 4th axis as
> > > you)
> > >
> > > I have a loop of wire on two spare pins on the connector where the
> > > 4th axis plugs in to the machine:
> > > This drives an input on the STMBL. The STMBL HV enable is only set
> > > (HAL Logic) if this input indicates motor-present.
> >
> > My HP_UHU drives have that feature built in.  The encoder connector
> > has to have a jumper to show the encoder is connected.  Probably to
> > avoid a runaway condition.
> >
> > > I just accept having an A axis display even when the axis isn't
> > > connected. And the machine will f-error if you try to move or jog
> > > it.

I could tolerate that.

> > I'm testing with both MACH3 and LinuxCNC so I can write up a neat
> > little conversion document that takes a parallel port BoB based
> > system from MACH3 into LinuxCNC with minimal pain.  In the MACH3
> > system, on my CNC router, when I tell the open loop stepper
> > controlled axis to home all that MACH3 does is zero out the DRO.
> >
> > But Gene has a home switch on his rotary indexer.  Now thinking
> > about it I think the solution might be simple. When he pulls the
> > indexer off the mill he has to unplug the encoder cable.  If he's
> > using safe type limits then the switch is normally closed and a home
> > (or limit) happens when it opens.
>
> Exactly what I'm doing.  Except no endoder, stepper driven and likely
> needs a bigger motor, and more speed reduction to improve the holding
> power.  Thats a downright puny motor on it now, and I have a 270 oz/in
> currently holding down a stepladder step. 8 wire so slower top speed
> but easy on the psu. In fact I think I'll check and see if its shaft
> is compatible yet today.

It was, and its been installed. Not tested yet, I need a male GS-16-4 or 
5 to hook up the home switch, AND send a signal back to hal that tells 
it its plugged in. I haven't figured out yet what to do with the signal 
though. But from what I can find on fleabay, it looks like I'll have to 
make a backshell and strain relief, the males I've found are all panel 
mount.  Nothing but a 16mm panel nut for a back end.  Sigh. I was born 
either 40 years early, or 40 years too late. I'd flip a quarter but it 
stands a measurable chance of not coming back down. :)

> > So run the home switch circuit through the encoder connector and
> > wire the mating encoder connector to have the connection closed. 
> > From the CNC controller perspective the system is already homed.
>
> Thats not how lcnc homes. In fact if the switch is closed, it jogs
> away till it opens, THEN goes into the normal home search routine. I
> use that to advantage on the Sheldon to assure it actually searches
> for the switch closure.  By sensing the front gib, which is NOW a
> solid bar the width of the carriage, once it has approached the
> spindle and closes, it stays closed, At homing, it will back away to
> the right till it opens before the normal search begins.
>
> I found this undocumented behavior when I put home switches on TLM.
> Putting in on the back of the bed because the screw is in the way on
> the front, and set the switch roller to touch the tapered gib holders
> bottom as it goes by. In fact I had to do a minor rebuild because I
> added the gibs.


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] Leaving an axis out of the home sequence AND (continued:)

[John Dammeyer]

> Exactly what I'm doing.  Except no endoder, stepper driven and likely
> needs a bigger motor, and more speed reduction to improve the holding
> power.  Thats a downright puny motor on it now, and I have a 270 oz/in
> currently holding down a stepladder step. 8 wire so slower top speed but
> easy on the psu. In fact I think I'll check and see if its shaft is
> compatible yet today.

Gene,
I just tried an experiment with my CNC router which is MACH3 driven.  I set up 
a spare input on parallel port 2 to be the Home switch input for the A axis.  
The rotary axis is a small 4" rotary indexer driven by a small stepper.

With the home set not activated I do a ref all home which sends each axis home 
in the following order.  First the router is moved up to the top limit/home 
switch.  Then the Y axis is moved towards me to the limit/home switch.  Then 
the router carriage is moved to the left until it reaches the limit/home 
switch.  Finally the rotary indexer starts turning and stops when I activate 
the switch.  Then it moves back in the other direction until I deactivate the 
switch.

That's behaviour as expected.

Now I move the rotary axis away from my 0 position to 60 degrees.  Move all the 
other axis away from their home.  Then do a ref all home again. Like before, Z 
first, then Y and X, finally it wants to move A but it can't.  Instead I get a 
message that the Home switch is already active and that I should move it off 
the switch first.  That completes the homing cycle.

The A axis is still usable.  I can manually set A to 0. by clicking a 
button on the screen.   I can enter in a G0 and then an A5 and see the rotary 
axis move to 5 degrees.

So even if I had a switch for an A axis home, all I get is a warning that the 
switch is activated and fix it.  It doesn't seem to cause any other problems.  
So I can run my CNC router with or without the A axis connected.  As long as 
the switch is open (NC home switch).

I would guess there must be a similar operation for LinuxCNC.  I can't test 
that until I put the big system back together and swap the hard drive for the 
LinuxCNC version.  (Should really make this dual boot but that's a different 
issue).

So perhaps play with your switch polarity (NC verses NO for Home Sensing) and 
see if you can make that work.

John






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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 12:47:23 John Dammeyer wrote:

> Hi Andy,
> See below.
>
> > -Original Message-
> > From: andy pugh [mailto:bodge...@gmail.com]
> > Sent: April-10-19 2:46 AM
> > To: Enhanced Machine Controller (EMC)
> > Subject: Re: [Emc-users] Leaving an axis out of the home sequence
> > AND (continued:)
> >
> > On Wed, 10 Apr 2019 at 04:49, John Dammeyer 
> >
> > wrote:
> > > Since my Harmonic Drive/STMBL are an AC Servo
> >
> > ...
> >
> > > The question then is whether or not the system, when starting
> > > can/will
> >
> > recognize an input which is a unit present signal.
> >
> > This is a solved problem :-) (I have exactly the same 4th axis as
> > you)
> >
> > I have a loop of wire on two spare pins on the connector where the
> > 4th axis plugs in to the machine:
> > This drives an input on the STMBL. The STMBL HV enable is only set
> > (HAL Logic) if this input indicates motor-present.
>
> My HP_UHU drives have that feature built in.  The encoder connector
> has to have a jumper to show the encoder is connected.  Probably to
> avoid a runaway condition.
>
> > I just accept having an A axis display even when the axis isn't
> > connected. And the machine will f-error if you try to move or jog
> > it.
>
> I'm testing with both MACH3 and LinuxCNC so I can write up a neat
> little conversion document that takes a parallel port BoB based system
> from MACH3 into LinuxCNC with minimal pain.  In the MACH3 system, on
> my CNC router, when I tell the open loop stepper controlled axis to
> home all that MACH3 does is zero out the DRO.
>
> But Gene has a home switch on his rotary indexer.  Now thinking about
> it I think the solution might be simple. When he pulls the indexer off
> the mill he has to unplug the encoder cable.  If he's using safe type
> limits then the switch is normally closed and a home (or limit)
> happens when it opens.

Exactly what I'm doing.  Except no endoder, stepper driven and likely 
needs a bigger motor, and more speed reduction to improve the holding 
power.  Thats a downright puny motor on it now, and I have a 270 oz/in 
currently holding down a stepladder step. 8 wire so slower top speed but 
easy on the psu. In fact I think I'll check and see if its shaft is 
compatible yet today.

> So run the home switch circuit through the encoder connector and wire
> the mating encoder connector to have the connection closed.  From the
> CNC controller perspective the system is already homed.

Thats not how lcnc homes. In fact if the switch is closed, it jogs away 
till it opens, THEN goes into the normal home search routine. I use that 
to advantage on the Sheldon to assure it actually searches for the 
switch closure.  By sensing the front gib, which is NOW a solid bar the 
width of the carriage, once it has approached the spindle and closes, it 
stays closed, At homing, it will back away to the right till it opens 
before the normal search begins.

I found this undocumented behavior when I put home switches on TLM. 
Putting in on the back of the bed because the screw is in the way on the 
front, and set the switch roller to touch the tapered gib holders bottom 
as it goes by. In fact I had to do a minor rebuild because I was sensing 
the bottom of the OEM gib, and the tapered holder is thicker.
> The only 
> problem will be if LinuxCNC uses a find home switch sequence, and then
> backs off so the switch just opens again before it declares it's home.
>
> In that case it would try to move off a non-existent home switch and
> try turning the indexer forever.  And perhaps that's again where that
> extra input comes in.  If the input is asserted then a home switch
> exists.  If it's not asserted then a home switch doesn't exist and he
> runs it like you run yours or I run mine.
>
> Again in the MACH3 situation, all I have to do is select the ports and
> pins menu selection under the config menu.  Select inputs and uncheck
> Home switch for A axis.  That also hits the software ESTOP and shuts
> power down to the drives when I click OK to save.  I can now unplug
> the A axis and remove it.  Then press the now blinking RESET button,
> power is back and a HOME command just clears the A axis DRO.  Simple. 
> No file editing.  No multiple config copies.
>
> That's what is really needed on LinuxCNC.
>
> John
>
> > It might be possible to load different configs depending on that
> > signal. A bash script can send halcmd commands to load the hostmot2
> > driver, start the threads, read a pin, and then run one of two or
> > more linuxcnc startups (after unloading Hostmot2 and realtime,
> >

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[John Dammeyer]

Hi Andy, 
See below.

> -Original Message-
> From: andy pugh [mailto:bodge...@gmail.com]
> Sent: April-10-19 2:46 AM
> To: Enhanced Machine Controller (EMC)
> Subject: Re: [Emc-users] Leaving an axis out of the home sequence AND
> (continued:)
> 
> On Wed, 10 Apr 2019 at 04:49, John Dammeyer 
> wrote:
> 
> > Since my Harmonic Drive/STMBL are an AC Servo
> ...
> > The question then is whether or not the system, when starting can/will
> recognize an input which is a unit present signal.
> 
> This is a solved problem :-) (I have exactly the same 4th axis as you)
> 
> I have a loop of wire on two spare pins on the connector where the 4th
> axis plugs in to the machine:
> This drives an input on the STMBL. The STMBL HV enable is only set
> (HAL Logic) if this input indicates motor-present.
My HP_UHU drives have that feature built in.  The encoder connector has to have 
a jumper to show the encoder is connected.  Probably to avoid a runaway 
condition.
> 
> I just accept having an A axis display even when the axis isn't
> connected. And the machine will f-error if you try to move or jog it.

I'm testing with both MACH3 and LinuxCNC so I can write up a neat little 
conversion document that takes a parallel port BoB based system from MACH3 into 
LinuxCNC with minimal pain.  In the MACH3 system, on my CNC router, when I tell 
the open loop stepper controlled axis to home all that MACH3 does is zero out 
the DRO.

But Gene has a home switch on his rotary indexer.  Now thinking about it I 
think the solution might be simple. When he pulls the indexer off the mill he 
has to unplug the encoder cable.  If he's using safe type limits then the 
switch is normally closed and a home (or limit) happens when it opens.  

So run the home switch circuit through the encoder connector and wire the 
mating encoder connector to have the connection closed.  From the CNC 
controller perspective the system is already homed.  The only problem will be 
if LinuxCNC uses a find home switch sequence, and then backs off so the switch 
just opens again before it declares it's home.

In that case it would try to move off a non-existent home switch and try 
turning the indexer forever.  And perhaps that's again where that extra input 
comes in.  If the input is asserted then a home switch exists.  If it's not 
asserted then a home switch doesn't exist and he runs it like you run yours or 
I run mine.

Again in the MACH3 situation, all I have to do is select the ports and pins 
menu selection under the config menu.  Select inputs and uncheck Home switch 
for A axis.  That also hits the software ESTOP and shuts power down to the 
drives when I click OK to save.  I can now unplug the A axis and remove it.  
Then press the now blinking RESET button, power is back and a HOME command just 
clears the A axis DRO.  Simple.  No file editing.  No multiple config copies.

That's what is really needed on LinuxCNC.  

John

> 
> It might be possible to load different configs depending on that
> signal. A bash script can send halcmd commands to load the hostmot2
> driver, start the threads, read a pin, and then run one of two or more
> linuxcnc startups (after unloading Hostmot2 and realtime, probably.)
> 
> Maybe even an edited version of the LinuxCNC startup script:
> 
> Personally I haven't cared enough about the vestigial axis to bother :-)
> 
> --
> atp
> "A motorcycle is a bicycle with a pandemonium attachment and is
> designed for the especial use of mechanical geniuses, daredevils and
> lunatics."
> � George Fitch, Atlanta Constitution Newspaper, 1916
> 
> 
> ___
> 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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 11:33:05 Jon Elson wrote:

> On 04/09/2019 10:47 PM, John Dammeyer wrote:
> > The question then is whether or not the system, when
> > starting can/will recognize an input which is a unit
> > present signal. Can not the HAL files be set up to
> > #include other files? So based on an input select one of
> > two HAL sub files to load. One with an A axis and one
> > without. Don't know enough about that yet but it's
> > definitely part of what I will need to learn.
>
> Well, I don't think there is a simple way to have a signal
> in hal control whether a hal file is loaded or not.  In
> other words, there is no conditional form of the HALFILE =
> line in the [HAL] section of the .ini file.
>
We are on the same page, Jon

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] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 11:28:41 Jon Elson wrote:

> On 04/09/2019 10:16 PM, Gene Heskett wrote:
> > And the interlock needs to know if the loaded code
> > contains an A|B|C move.
>
> OK, now this is going beyond the capabilities of LinuxCNC as
> it is. I think the best way to handle this is to have two
> configs.  If you ever load a G-code file with an A move when
> the config has no A axis, that will generate a fault as soon
> as you LOAD the program. If you really need this protection,
> then two configs is probably the best way.
>
> > Pretty complex... And I'm beat up from yard work, tried to
> > trim a hedge that has some 40mm diameter branches. I'll
> > hurt tomorrow, sure as heck. Humm, maybe I could use a
> > timer trick like I once did in the simulation to generate
> > a dummy home switch. Or a dummy search switch closure
> > thats actually a timedelay, I've done that so code useing
> > a G38.2 could be simulated.
>
> Another way is to use a spare general purpose input that is
> controlled by a jumper in the A axis connector.  Then, when
> no plug (or a shorting plug to keep noise out of the encoder
> inputs) is there, it can be sensed whether the A axis is
> plugged in.  Now, exactly how you use this is a good
> question.  I think you could set up some hal code that will
> cause an error if you run the config with the A axis when it
> is not plugged in.  Or, of course, it could just refuse to
> home A when not present, and that would tell you you have
> loaded the wrong config, or that you need to plug A in and
> rehome.
>
> Jon
>
That idea might be further developed by using the nc contacts, then if 
they aren't closed jog for 1 sec to see if it does close. If not, then 
its not there. But I don't believe LCNC can do that today, and I've no 
clue how much coding it would take to add that. If you knew it was 
clear, one could jog for long enough to tickle the switch, on this one a 
2.5 second jog at search_vel should register a hit. And considering that 
opens a big can of worms, I'm NOT going to campaign for that. So at some 
point when I've nothing better to do, I'll see if I can cobble up a 
halfile list and files that work w/o any great duplication. In the 
meantime I'll muddle along until I paint myself into a corner and must 
have a 3 axis version. The new vfd should be here yet this week, and 
I've more yard work than I want till then.

Cheers Jon, 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] Leaving an axis out of the home sequence AND (continued:)

[Ralph Stirling]

Carbon is soluble in iron, which is why you can't
use diamond to machine steel.  When the steel
heats up, it just starts absorbing carbon from the
diamond particles.  Your high carbon steel becomes
higher carbon steel and your diamond goes away.

-- Ralph

From: Gene Heskett [ghesk...@shentel.net]
Sent: Wednesday, April 10, 2019 9:02 AM
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] Leaving an axis out of the home sequence AND 
(continued:)

On Wednesday 10 April 2019 11:21:11 Jon Elson wrote:

> On 04/09/2019 08:31 PM, Gene Heskett wrote:
> > Greetings all;
> >
> > out of the logic that determines whether the machine will run gcode
> > when that A axis has not been homed because its unplugged, off the
> > machine and stored and either is unhomed, or was never homed since
> > this power up.
> >
> > How is this best done? Preferably done so that a powerdown to switch
> > its homing on or off, won't disturb LCNC's ability to run gcode just
> > because the A isn't homed??
>
> Well, there usually is not a huge reason to have home
> switches on a rotary axis, unless you have setups that will
> be reused on following days.

Or need a reference for setting angles while shaping HSS tools in it.  If
I can find a CBN cup small enough. I've already tore up $200 worth of
diamond stuff trying to use diamond on HSS. No matter how slow you turn
the diamond, it will fail, usually before the desired shape has been
attained.

> So, you can set the A to home
> wherever it is (search and latch velocity of zero).  Ground
> encoder inputs with a shorting plug, so noise doesn't cause
> LinuxCNC to think the axis is moving without command.  Then,
> you can even unplug the axis while LinuxCNC is running, and
> plug in the shorting plug.

No, not with stepper drivers. You do any connection foolery with power
off.  With the relay I put in yesterday, disabling machine motion does
that.
>
> Another scheme is to have two LinuxCNC config file sets, one
> with the A axis and one without.

I'm leaning somewhat in that direction. I see by the Docs that the
halfile spec says files.  Plural.  So that and the #INCLUDE may show a
way fwd. Time will tell. Might be 2 .ini's and 4 or more .halfiles
before thats all done. And becoming hell to maintain. :( The addf's
should probably be two separate but complete lists just to maintain a
working addf ordering. Twopass should help with that if I put the A addf
stuff as an #INCLUDE right in the correct location in the 2nd halfile's
addf list.

I can see my coffee consumption, already too high, going up before thats
finished. :(

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


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 <http://geneslinuxbox.net:6309/gene>



___
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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[andy pugh]

On Wed, 10 Apr 2019 at 16:07, Gene Heskett  wrote:

> So I don't see a tasty way to do it as I'd still have two pairs of files
> to keep in synch with such as MAX_VELS etc. With the sheer number of
> variables to keep synched, the N! could rapidly exceed 69, the maximum
> my now elderly ti-51 calculator can do.

You set these things _once_ I really don't see your problem.

You _can_ do conditional stuff in HAL files, if you write them in Tcl.
I have never considered it worth the trouble of learning Tcl.

Example: 
https://github.com/LinuxCNC/linuxcnc/blob/master/lib/hallib/wheeljogpins.tcl

(And I think that file actually has a typo...)

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 11:21:11 Jon Elson wrote:

> On 04/09/2019 08:31 PM, Gene Heskett wrote:
> > Greetings all;
> >
> > out of the logic that determines whether the machine will run gcode
> > when that A axis has not been homed because its unplugged, off the
> > machine and stored and either is unhomed, or was never homed since
> > this power up.
> >
> > How is this best done? Preferably done so that a powerdown to switch
> > its homing on or off, won't disturb LCNC's ability to run gcode just
> > because the A isn't homed??
>
> Well, there usually is not a huge reason to have home
> switches on a rotary axis, unless you have setups that will
> be reused on following days.

Or need a reference for setting angles while shaping HSS tools in it.  If 
I can find a CBN cup small enough. I've already tore up $200 worth of 
diamond stuff trying to use diamond on HSS. No matter how slow you turn 
the diamond, it will fail, usually before the desired shape has been 
attained.

> So, you can set the A to home 
> wherever it is (search and latch velocity of zero).  Ground
> encoder inputs with a shorting plug, so noise doesn't cause
> LinuxCNC to think the axis is moving without command.  Then,
> you can even unplug the axis while LinuxCNC is running, and
> plug in the shorting plug.

No, not with stepper drivers. You do any connection foolery with power 
off.  With the relay I put in yesterday, disabling machine motion does 
that.
>
> Another scheme is to have two LinuxCNC config file sets, one
> with the A axis and one without.

I'm leaning somewhat in that direction. I see by the Docs that the 
halfile spec says files.  Plural.  So that and the #INCLUDE may show a 
way fwd. Time will tell. Might be 2 .ini's and 4 or more .halfiles 
before thats all done. And becoming hell to maintain. :( The addf's 
should probably be two separate but complete lists just to maintain a 
working addf ordering. Twopass should help with that if I put the A addf 
stuff as an #INCLUDE right in the correct location in the 2nd halfile's 
addf list.

I can see my coffee consumption, already too high, going up before thats 
finished. :(

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


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] Leaving an axis out of the home sequence AND (continued:)

[Jon Elson]

On 04/09/2019 10:47 PM, John Dammeyer wrote:
The question then is whether or not the system, when 
starting can/will recognize an input which is a unit 
present signal. Can not the HAL files be set up to 
#include other files? So based on an input select one of 
two HAL sub files to load. One with an A axis and one 
without. Don't know enough about that yet but it's 
definitely part of what I will need to learn.
Well, I don't think there is a simple way to have a signal 
in hal control whether a hal file is loaded or not.  In 
other words, there is no conditional form of the HALFILE =

line in the [HAL] section of the .ini file.

Jon


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Jon Elson]

On 04/09/2019 10:16 PM, Gene Heskett wrote:
And the interlock needs to know if the loaded code 
contains an A|B|C move.
OK, now this is going beyond the capabilities of LinuxCNC as 
it is. I think the best way to handle this is to have two 
configs.  If you ever load a G-code file with an A move when 
the config has no A axis, that will generate a fault as soon 
as you LOAD the program. If you really need this protection, 
then two configs is probably the best way.
Pretty complex... And I'm beat up from yard work, tried to 
trim a hedge that has some 40mm diameter branches. I'll 
hurt tomorrow, sure as heck. Humm, maybe I could use a 
timer trick like I once did in the simulation to generate 
a dummy home switch. Or a dummy search switch closure 
thats actually a timedelay, I've done that so code useing 
a G38.2 could be simulated.
Another way is to use a spare general purpose input that is 
controlled by a jumper in the A axis connector.  Then, when 
no plug (or a shorting plug to keep noise out of the encoder 
inputs) is there, it can be sensed whether the A axis is 
plugged in.  Now, exactly how you use this is a good 
question.  I think you could set up some hal code that will 
cause an error if you run the config with the A axis when it 
is not plugged in.  Or, of course, it could just refuse to 
home A when not present, and that would tell you you have 
loaded the wrong config, or that you need to plug A in and 
rehome.


Jon


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Jon Elson]

On 04/09/2019 08:31 PM, Gene Heskett wrote:

Greetings all;

out of the logic that determines whether the machine will run gcode when
that A axis has not been homed because its unplugged, off the machine
and stored and either is unhomed, or was never homed since this power
up.
  
How is this best done? Preferably done so that a powerdown to switch its

homing on or off, won't disturb LCNC's ability to run gcode just because
the A isn't homed??


Well, there usually is not a huge reason to have home 
switches on a rotary axis, unless you have setups that will 
be reused on following days.  So, you can set the A to home 
wherever it is (search and latch velocity of zero).  Ground 
encoder inputs with a shorting plug, so noise doesn't cause 
LinuxCNC to think the axis is moving without command.  Then, 
you can even unplug the axis while LinuxCNC is running, and 
plug in the shorting plug.


Another scheme is to have two LinuxCNC config file sets, one 
with the A axis and one without.


Jon


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Wednesday 10 April 2019 06:00:50 andy pugh wrote:

> On Wed, 10 Apr 2019 at 10:50, Gene Heskett  
wrote:
> > Two separate configs means it will be hard to keep them in synch,
>
> Two configs (ie INI files) can share the same folder, so the HAL
> files, .var file and tool table are all shared.
>
And if I copy the ini to one with no A references in it, do I also have 
separate .hal files, or does hal ignore the A references?  Or do you 
have 3 hal files? A basic, and a load vertical and a load horizontal?

I've just perused the docs again, looking for a way to #IF an #INCLUDE 
with that in mind, but that would appear to need a preliminary loadusr 
of a panel.ui just to select things.

So I don't see a tasty way to do it as I'd still have two pairs of files 
to keep in synch with such as MAX_VELS etc. With the sheer number of 
variables to keep synched, the N! could rapidly exceed 69, the maximum 
my now elderly ti-51 calculator can do.  So the best way might yet be to 
do the 2 configs by 2 .ini's in one directory, along with two .hals, so 
the -3 version is just the -4 with the A stuff stripped.  Problems like 
this beg for an IF [condition] in both the .ini and the hal.

I do see that panel.ui has gfx that pyvcp doesn't, like the little dot 
boxes. Or a teeny square with a check if condition=true. pyvcp is stuck 
with color controllable leds, but you can't click the led so you still 
need a button to click.

I'd also complain if hung with a used rope ;-)

> I do this with my mill to switch between horizontal and vertical
> spindle modes. Most of the HAL files for spindle etc are the same, but
> the different INI files call different sub-sets of HAL files and set
> the kinematics differently.
> (JA and the new trivkins made that part really easy)

Maybe a chapter on this in the Documentation.pdf?  Sounds like it would 
be quite usefull for this.

Can the .ini reliably spec 2 or more .hal files? Separate out the A stuff 
for a 2nd .halfile, change the number of axises from 3 to 4, and call 
the A stuff in an extra halfile in one of 2 separate .ini's?
>From section 8.2.2.7 it appears to be workable, with the exception of 
properly ordering the resultant addf's. That will involve some 
#INCLUDE's, and that boils down to two separate halfiles to maintain 
because we don't have an IF/ELSE/ENDIF.  Makes me get all growly. So I 
guess it gets clamped to the back of the shelf for storage when not it 
use.  Actually, I have a mental image of a extra platform on the rear to 
hold it where its out of the way, but a 1" dowel stuck in the end of a 
blank would suffice to hold and rotate the blank for gunstock work. But 
due to Z height limits, most of the buttstock end will need to be done 
with it laying flat. I can see a problem because of the buttstock toe 
hanging down and needing Z height I don't have while doing the barrel 
channel too, but the G0704 can do that.  That stock setup would leave 
room beside it for the hack-a-day style of tool changer.

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] Leaving an axis out of the home sequence AND (continued:)

[andy pugh]

On Wed, 10 Apr 2019 at 10:50, Gene Heskett  wrote:

> Two separate configs means it will be hard to keep them in synch,

Two configs (ie INI files) can share the same folder, so the HAL
files, .var file and tool table are all shared.

I do this with my mill to switch between horizontal and vertical
spindle modes. Most of the HAL files for spindle etc are the same, but
the different INI files call different sub-sets of HAL files and set
the kinematics differently.
(JA and the new trivkins made that part really easy)

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Tuesday 09 April 2019 23:47:24 John Dammeyer wrote:

> I believe I'll have a similar problem.
>
> > > Assuming you unplug the rotary axis when you remove it, you could
> > > simply use a normally closed homing switch that unplugs with the
> > > axis. Problem with this is it wouldn’t let you use the fast
> > > approach, slow home sequence with a single switch.
> > >
> > > Thaddeus Waldner
> >
> > unplugging it isn't practical without adding a plug in the switch
> > cable, but some timers could simulate that, and a mux2 to steer the
> > timers or the switch could work. I'll have see if I can set
> > something up along those lines.
> > Cheers, Gene Heskett
>
> Since my Harmonic Drive/STMBL are an AC Servo it's unlikely that it's
> a good idea to unplug it while the system is powered.  So one has to
> either ESTOP which removes DC or power down the entire system.  Then
> unbolt the rotary table/indexer assembly and remove/unplug it.Then
> release ESTOP or power up the system again.  Mostly because my Mill
> Table isn't really large enough to keep it there permanently.
>
> The question then is whether or not the system, when starting can/will
> recognize an input which is a unit present signal.   Can not the HAL
> files be set up to  #include other files?  So based on an input select
> one of two HAL sub files to load.  One with an A axis and one without.
>  Don't know enough about that yet but it's definitely part of what I
> will need to learn.
>
> John

Thats a thought I should follow up on. I'm going to make a "use A" 
checkbox in the post gui but that might be a timing problem. I'll 
concentrate on the checkbox first, then figure out how to use it. Or 
maybe a postgui for the postgui?

Two separate configs means it will be hard to keep them in synch, a 
headache I'd druther not have. AFAIK, the ini can't use an IF, and 
neither can hal. That would be a welcome addition, so one could build 
his config on the fly as it was loading based on ENV settings that could 
be set with a wrapper script.  Hint hint... Or a bash wrapper to 
assemble the config according to ENV vars? I suspect there are at least 
50 ways to skin this cat. ;-)
>
>
>
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


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] Leaving an axis out of the home sequence AND (continued:)

[andy pugh]

On Wed, 10 Apr 2019 at 04:49, John Dammeyer  wrote:

> Since my Harmonic Drive/STMBL are an AC Servo
...
> The question then is whether or not the system, when starting can/will 
> recognize an input which is a unit present signal.

This is a solved problem :-) (I have exactly the same 4th axis as you)

I have a loop of wire on two spare pins on the connector where the 4th
axis plugs in to the machine:
This drives an input on the STMBL. The STMBL HV enable is only set
(HAL Logic) if this input indicates motor-present.

I just accept having an A axis display even when the axis isn't
connected. And the machine will f-error if you try to move or jog it.

It might be possible to load different configs depending on that
signal. A bash script can send halcmd commands to load the hostmot2
driver, start the threads, read a pin, and then run one of two or more
linuxcnc startups (after unloading Hostmot2 and realtime, probably.)

Maybe even an edited version of the LinuxCNC startup script:

Personally I haven't cared enough about the vestigial axis to bother :-)

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916


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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Tuesday 09 April 2019 23:40:54 Thaddeus Waldner wrote:

> On Apr 9, 2019, at 10:26 PM, Gene Heskett  wrote:
> > On Tuesday 09 April 2019 23:02:46 Thaddeus Waldner wrote:
> >> Assuming you unplug the rotary axis when you remove it, you could
> >> simply use a normally closed homing switch that unplugs with the
> >> axis. Problem with this is it wouldn’t let you use the fast
> >> approach, slow home sequence with a single switch.
> >>
> >> Thaddeus Waldner
> >> Newdale School
> >> Elkton, SD 57026
> >>
> >> 
> >> From: andy pugh 
> >> Sent: Tuesday, April 9, 2019 9:34 PM
> >> To: Enhanced Machine Controller (EMC)
> >> Subject: Re: [Emc-users] Leaving an axis out of the home sequence
> >> AND (continued:)
> >>
> >> On Wed, 10 Apr 2019 at 02:33, Gene Heskett 
> >
> > wrote:
> >>> out of the logic that determines whether the machine will run
> >>> gcode when that A axis has not been homed because its unplugged,
> >>> off the machine and stored and either is unhomed, or was never
> >>> homed since this power up.
> >>
> >> The simplest is probably to set it for immediate homing (home
> >> search velocity zero) but have it in the homing sequence.
> >>
> >> If you want it to physically home when present, but not home when
> >> absent, automatically, then it's harder. I imagine that some HAL
> >> logic to detect presence or absence and a mux2 to switch between
> >> real and simulated home (wcomp looking at apparent position of the
> >> stepgen?)
> >
> > unplugging it isn't practical without adding a plug in the switch
> > cable, but some timers could simulate that, and a mux2 to steer the
> > timers or the switch could work. I'll have see if I can set
> > something up along those lines.
> >
> > More ideas.  Thanks Thaddeus.
>
> Another idea:
> Instead of simulating a home switch with a timer, you could use an
> actual home switch from a different axis (eg. z-axis).
>
Not sure about that one. I'll investigate the "check to use" idea first. 
My homes, except for A are very close to the switches, so a rehome 
doesn't take long. 5 seconds maybe.

One of the things I have planned for A is sharpening/shaping HSS lathe 
tooling, but have not found a mounted CBN wheel of a size suitable for 
spinning in this spindle. Seems to me the CBN people are missing a 
market niche by not making a fairly rigid miniature cup mounted version. 

The diamond wheel dremel now sells for $26, is so thin and flimsy you 
can't accurately maintain a face angle, plus the diamond isn't 
compatible with HSS, so its not a good choice for HSS. I have made an 
arbor to spin a 4.5" CBN cup in the G0704, and have used it a time or 7, 
but even though I've fine tuned the face of the arbor 2 times, it still 
runs out of true by several thou at the cutting face.  Thats hard on the 
CBN.

And I need to make a base wedge to hold the A nose up (or down) by 7 
degrees for the clearance angles. One almost needs a 2 axis combo table 
for that, at which point LCNC needs a 5th axis.  And that means a 
different interface, like a 7i90HD, or different firmware in the 5i25 to 
put the 5th axis on the p2 connector. The 7i76D only has 4.  Humm, I 
supposed a software stepper could be loaded, and a base-thread setup, 
hal can certainly do that. 2 problems with that though, first, and show 
stopper being the limited Z motion available in a gantry. I bought one 
of those $70 touch probes, and had to make a shorter stylus to get it to 
clear the bed. Piece of fragile plastic junk IMO.  As a probe, its 
usable for xy locations, but will need a round ball tip for z profiles. 
No where near concentric although that can be fiddled with given enough 
time.

I'll probably get a Wild Horse probe before I'm happy. But thats a few 
pages down the log I'm not keeping.

Simple hack-a-day carousel Tool changer first. 6 pocket probably.

With this vfd, I left it running at 675 revs for a few hours yesterday, 
to check out the slow speed burn-up horror stories. At 675 revs, and I 
don't currently have any low speed boost programmed in, in 2 hours the 
motor was up 2C and the coolant tank was up 1C. I can stop it with two 
fingers, so obviously I can add some low speed boost when the new vfd 
gets here. It was back to running the spindle fwd yesterday. Go figure.

Cheers Thaddeus, 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 <http://geneslinuxbox.net:6309/gene>



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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[jrmitchellj]

Maybe I am thinking too simplistically, but my approach would be:
Start with my 3-axis config
Copy it
add the fourth axis information to the copy
change the name of he config to 4 axis.

The steppers I used cannot have the motor connected while powered.  So this
method suits me.

--J. Ray Mitchell Jr.
jrmitche...@gmail.com



"Good enough is the enemy of excellence"author unknown


On Tue, Apr 9, 2019 at 8:50 PM John Dammeyer  wrote:

> I believe I'll have a similar problem.
>
> > > Assuming you unplug the rotary axis when you remove it, you could
> > > simply use a normally closed homing switch that unplugs with the axis.
> > > Problem with this is it wouldn’t let you use the fast approach, slow
> > > home sequence with a single switch.
> > >
> > > Thaddeus Waldner
>
> > unplugging it isn't practical without adding a plug in the switch cable,
> > but some timers could simulate that, and a mux2 to steer the timers or
> > the switch could work. I'll have see if I can set something up along
> > those lines.
> > Cheers, Gene Heskett
>
> Since my Harmonic Drive/STMBL are an AC Servo it's unlikely that it's a
> good idea to unplug it while the system is powered.  So one has to either
> ESTOP which removes DC or power down the entire system.  Then unbolt the
> rotary table/indexer assembly and remove/unplug it.Then release ESTOP
> or power up the system again.  Mostly because my Mill Table isn't really
> large enough to keep it there permanently.
>
> The question then is whether or not the system, when starting can/will
> recognize an input which is a unit present signal.   Can not the HAL files
> be set up to  #include other files?  So based on an input select one of two
> HAL sub files to load.  One with an A axis and one without.  Don't know
> enough about that yet but it's definitely part of what I will need to learn.
>
> John
>
>
>
>
>
> ___
> 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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[John Dammeyer]

I believe I'll have a similar problem.

> > Assuming you unplug the rotary axis when you remove it, you could
> > simply use a normally closed homing switch that unplugs with the axis.
> > Problem with this is it wouldn’t let you use the fast approach, slow
> > home sequence with a single switch.
> >
> > Thaddeus Waldner

> unplugging it isn't practical without adding a plug in the switch cable,
> but some timers could simulate that, and a mux2 to steer the timers or
> the switch could work. I'll have see if I can set something up along
> those lines.
> Cheers, Gene Heskett

Since my Harmonic Drive/STMBL are an AC Servo it's unlikely that it's a good 
idea to unplug it while the system is powered.  So one has to either ESTOP 
which removes DC or power down the entire system.  Then unbolt the rotary 
table/indexer assembly and remove/unplug it.Then release ESTOP or power up 
the system again.  Mostly because my Mill Table isn't really large enough to 
keep it there permanently.

The question then is whether or not the system, when starting can/will 
recognize an input which is a unit present signal.   Can not the HAL files be 
set up to  #include other files?  So based on an input select one of two HAL 
sub files to load.  One with an A axis and one without.  Don't know enough 
about that yet but it's definitely part of what I will need to learn.

John





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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Thaddeus Waldner]

On Apr 9, 2019, at 10:26 PM, Gene Heskett  wrote:
> 
> On Tuesday 09 April 2019 23:02:46 Thaddeus Waldner wrote:
> 
>> Assuming you unplug the rotary axis when you remove it, you could
>> simply use a normally closed homing switch that unplugs with the axis.
>> Problem with this is it wouldn’t let you use the fast approach, slow
>> home sequence with a single switch.
>> 
>> Thaddeus Waldner
>> Newdale School
>> Elkton, SD 57026
>> 
>> 
>> From: andy pugh 
>> Sent: Tuesday, April 9, 2019 9:34 PM
>> To: Enhanced Machine Controller (EMC)
>> Subject: Re: [Emc-users] Leaving an axis out of the home sequence AND
>> (continued:)
>> 
>> On Wed, 10 Apr 2019 at 02:33, Gene Heskett  
> wrote:
>>> out of the logic that determines whether the machine will run gcode
>>> when that A axis has not been homed because its unplugged, off the
>>> machine and stored and either is unhomed, or was never homed since
>>> this power up.
>> 
>> The simplest is probably to set it for immediate homing (home search
>> velocity zero) but have it in the homing sequence.
>> 
>> If you want it to physically home when present, but not home when
>> absent, automatically, then it's harder. I imagine that some HAL logic
>> to detect presence or absence and a mux2 to switch between real and
>> simulated home (wcomp looking at apparent position of the stepgen?)
>> 
> unplugging it isn't practical without adding a plug in the switch cable, 
> but some timers could simulate that, and a mux2 to steer the timers or 
> the switch could work. I'll have see if I can set something up along 
> those lines.
> 
> More ideas.  Thanks Thaddeus.
> 

Another idea:
Instead of simulating a home switch with a timer, you could use an actual home 
switch from a different axis (eg. z-axis).

>> --
>> atp
>> "A motorcycle is a bicycle with a pandemonium attachment and is
>> designed for the especial use of mechanical geniuses, daredevils and
>> lunatics."
>> — George Fitch, Atlanta Constitution Newspaper, 1916
>> 
>> 
>> ___
>> Emc-users mailing list
>> Emc-users@lists.sourceforge.net <mailto:Emc-users@lists.sourceforge.net>
>> https://lists.sourceforge.net/lists/listinfo/emc-users 
>> <https://lists.sourceforge.net/lists/listinfo/emc-users>
>> 
>> ___
>> Emc-users mailing list
>> Emc-users@lists.sourceforge.net <mailto:Emc-users@lists.sourceforge.net>
>> https://lists.sourceforge.net/lists/listinfo/emc-users 
>> <https://lists.sourceforge.net/lists/listinfo/emc-users>
> 
> 
> 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 <http://geneslinuxbox.net:6309/gene 
> <http://geneslinuxbox.net:6309/gene>>
> 
> 
> 
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net <mailto:Emc-users@lists.sourceforge.net>
> https://lists.sourceforge.net/lists/listinfo/emc-users 
> <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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Tuesday 09 April 2019 23:02:46 Thaddeus Waldner wrote:

> Assuming you unplug the rotary axis when you remove it, you could
> simply use a normally closed homing switch that unplugs with the axis.
> Problem with this is it wouldn’t let you use the fast approach, slow
> home sequence with a single switch.
>
> Thaddeus Waldner
> Newdale School
> Elkton, SD 57026
>
> 
> From: andy pugh 
> Sent: Tuesday, April 9, 2019 9:34 PM
> To: Enhanced Machine Controller (EMC)
> Subject: Re: [Emc-users] Leaving an axis out of the home sequence AND
> (continued:)
>
> On Wed, 10 Apr 2019 at 02:33, Gene Heskett  
wrote:
> > out of the logic that determines whether the machine will run gcode
> > when that A axis has not been homed because its unplugged, off the
> > machine and stored and either is unhomed, or was never homed since
> > this power up.
>
> The simplest is probably to set it for immediate homing (home search
> velocity zero) but have it in the homing sequence.
>
> If you want it to physically home when present, but not home when
> absent, automatically, then it's harder. I imagine that some HAL logic
> to detect presence or absence and a mux2 to switch between real and
> simulated home (wcomp looking at apparent position of the stepgen?)
>
unplugging it isn't practical without adding a plug in the switch cable, 
but some timers could simulate that, and a mux2 to steer the timers or 
the switch could work. I'll have see if I can set something up along 
those lines.

More ideas.  Thanks Thaddeus.

> --
> atp
> "A motorcycle is a bicycle with a pandemonium attachment and is
> designed for the especial use of mechanical geniuses, daredevils and
> lunatics."
> — George Fitch, Atlanta Constitution Newspaper, 1916
>
>
> ___
> 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


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 <http://geneslinuxbox.net:6309/gene>



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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

On Tuesday 09 April 2019 22:31:26 andy pugh wrote:

> On Wed, 10 Apr 2019 at 02:33, Gene Heskett  
wrote:
> > out of the logic that determines whether the machine will run gcode
> > when that A axis has not been homed because its unplugged, off the
> > machine and stored and either is unhomed, or was never homed since
> > this power up.
>
> The simplest is probably to set it for immediate homing (home search
> velocity zero) but have it in the homing sequence.
>
> If you want it to physically home when present, but not home when
> absent, automatically, then it's harder. I imagine that some HAL logic
> to detect presence or absence

That appears to be the show stopper, unless a pyvcp button to click on 
when its going to be used might be used to tally that, run the mux etc. 
But that would be after the below was already initialized I think. But a 
rehome after the tally was made true or false might show a way.  It will 
take some imagination. Assuming the mux2 was the zero or search vel 
switch, how do I get the mux2 output back into the search vel link?

And the interlock needs to know if the loaded code contains an A|B|C 
move. Pretty complex...  And I'm beat up from yard work, tried to trim a 
hedge that has some 40mm diameter branches. I'll hurt tomorrow, sure as 
heck.

Humm, maybe I could use a timer trick like I once did in the simulation 
to generate a dummy home switch.  Or a dummy search switch closure thats 
actually a timedelay, I've done that so code useing a G38.2 could be 
simulated.   
 
> and a mux2 to switch between real and 
> simulated home (wcomp looking at apparent position of the stepgen?)

Thats another way.  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] Leaving an axis out of the home sequence AND (continued:)

[Thaddeus Waldner]

Assuming you unplug the rotary axis when you remove it, you could simply use a 
normally closed homing switch that unplugs with the axis. Problem with this is 
it wouldn’t let you use the fast approach, slow home sequence with a single 
switch.

Thaddeus Waldner
Newdale School
Elkton, SD 57026


From: andy pugh 
Sent: Tuesday, April 9, 2019 9:34 PM
To: Enhanced Machine Controller (EMC)
Subject: Re: [Emc-users] Leaving an axis out of the home sequence AND 
(continued:)

On Wed, 10 Apr 2019 at 02:33, Gene Heskett  wrote:

> out of the logic that determines whether the machine will run gcode when
> that A axis has not been homed because its unplugged, off the machine
> and stored and either is unhomed, or was never homed since this power
> up.

The simplest is probably to set it for immediate homing (home search
velocity zero) but have it in the homing sequence.

If you want it to physically home when present, but not home when
absent, automatically, then it's harder. I imagine that some HAL logic
to detect presence or absence and a mux2 to switch between real and
simulated home (wcomp looking at apparent position of the stepgen?)

--
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916


___
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

Re: [Emc-users] Leaving an axis out of the home sequence AND (continued:)

[andy pugh]

On Wed, 10 Apr 2019 at 02:33, Gene Heskett  wrote:

> out of the logic that determines whether the machine will run gcode when
> that A axis has not been homed because its unplugged, off the machine
> and stored and either is unhomed, or was never homed since this power
> up.

The simplest is probably to set it for immediate homing (home search
velocity zero) but have it in the homing sequence.

If you want it to physically home when present, but not home when
absent, automatically, then it's harder. I imagine that some HAL logic
to detect presence or absence and a mux2 to switch between real and
simulated home (wcomp looking at apparent position of the stepgen?)

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916


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

[Emc-users] Leaving an axis out of the home sequence AND (continued:)

[Gene Heskett]

Greetings all;

out of the logic that determines whether the machine will run gcode when 
that A axis has not been homed because its unplugged, off the machine 
and stored and either is unhomed, or was never homed since this power 
up.
 
How is this best done? Preferably done so that a powerdown to switch its 
homing on or off, won't disturb LCNC's ability to run gcode just because 
the A isn't homed?? 

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