Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[Gene Heskett]

On Wednesday 09 September 2020 13:41:53 Jon Elson wrote:

> On 09/09/2020 12:16 PM, Gene Heskett wrote:
> > On Wednesday 09 September 2020 12:38:43 Jon Elson wrote:
> >> On 09/09/2020 11:08 AM, Gene Heskett wrote:
> >>> Is it possible to get an enable signal, per axis, from the mesa
> >>> cards that would become true, hopefully a couple micky-secs ahead
> >>> of the first stepper pulse so no missed steps would normally
> >>> occur, losing its home, and go back false when the last step has
> >>> been sent? It maybe could be timed right if it actually steps on
> >>> the trailing edge of the pulse, and the enabling action was on the
> >>> leading edge.
> >>
> >> This is the TRICKY bit, knowing when an axis is SOON going
> >> to start moving.
> >>
> >> Jon
> >
> > I'd be surprised if these things needed any lead time at all, Jon.
> > The hitachi docs on this chip do not mention any timing constraints
> > in regard to this. So its likely that it could be brought up to full
> > song with the leading edge of the pulse. If that triggered a
> > retriggerable oneshot that was then or'd with the pulse, and its
> > output would be the enable high current signal. Let the oneshot time
> > out in 50 millisecs if no pulse, putting it back in reduced current
> > at the end of the move. If the oneshot was in the mesa, would be
> > ideal, and divorced from the enable granularity of the 1khz servo
> > thread.  That granularity could easily lose enough steps to destroy
> > a home position, something I intend to have with this BS-1 clone.  A
> > oneshot in the hal file would be a disaster in that case because of
> > the millisecond+ delay. A oneshot in the cable to the driver would
> > work, but if its in the mesa card, better yet. :)
>
> Certainly worth a try.  One horrible possibility is that the
> DIP switches are only read once at
> power-up.  The first couple steps should be spaced out as
> the axis accelerates which should give the drive time to
> ramp up the current.
>
> Jon

Nah, I can change them while its running, and hear the change in the 
motors sound as the currents change in real time.  But thats not saying 
that some other board using this power chip might not do it differently.

These came out of the electrics box that came with a 6040 gantry mill I 
paid $1175 for, very carefully designed so as not to be used by anything 
but the very limited Mach3.  The vfd for instance, has a reverseing 
input.  It ignores it.  It also has an analog speed input, also ignored, 
forcing you to use only that mode where the front panel knob is active.  

Then, finding the motor supply was duff, I said to hell with this and 
went to the shop, coming back with the driver kit from my broken hf 
mill.  Runs it well, add another ebay clone of a decent vfd and it all 
Just Works, x at 192 ipm, z at 35 ((needs more motor to lift a 1.5 
horse water cooled spindle) and A at about 15 rpm.  Turn on the mister, 
and carve the panels for the interface box, 12 holes Plus a line cord 
grommet, in about 3m:30s.  And doesn't sound like its working all that 
hard.

Thanks Jon.

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)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[Gene Heskett]

On Wednesday 09 September 2020 13:41:53 Jon Elson wrote:

> On 09/09/2020 12:16 PM, Gene Heskett wrote:
> > On Wednesday 09 September 2020 12:38:43 Jon Elson wrote:
> >> On 09/09/2020 11:08 AM, Gene Heskett wrote:
> >>> Is it possible to get an enable signal, per axis, from the mesa
> >>> cards that would become true, hopefully a couple micky-secs ahead
> >>> of the first stepper pulse so no missed steps would normally
> >>> occur, losing its home, and go back false when the last step has
> >>> been sent? It maybe could be timed right if it actually steps on
> >>> the trailing edge of the pulse, and the enabling action was on the
> >>> leading edge.
> >>
> >> This is the TRICKY bit, knowing when an axis is SOON going
> >> to start moving.
> >>
> >> Jon
> >
> > I'd be surprised if these things needed any lead time at all, Jon.
> > The hitachi docs on this chip do not mention any timing constraints
> > in regard to this. So its likely that it could be brought up to full
> > song with the leading edge of the pulse. If that triggered a
> > retriggerable oneshot that was then or'd with the pulse, and its
> > output would be the enable high current signal. Let the oneshot time
> > out in 50 millisecs if no pulse, putting it back in reduced current
> > at the end of the move. If the oneshot was in the mesa, would be
> > ideal, and divorced from the enable granularity of the 1khz servo
> > thread.  That granularity could easily lose enough steps to destroy
> > a home position, something I intend to have with this BS-1 clone.  A
> > oneshot in the hal file would be a disaster in that case because of
> > the millisecond+ delay. A oneshot in the cable to the driver would
> > work, but if its in the mesa card, better yet. :)
>
> Certainly worth a try.  One horrible possibility is that the
> DIP switches are only read once at
> power-up.  The first couple steps should be spaced out as
> the axis accelerates which should give the drive time to
> ramp up the current.
>
> Jon

I figure on using the regular axis/joint settings for those profiles.
ATM I've no clue what the effective gear ratio is, but I can have that in 
hand an hour after I get it hooked up with an aribtrary home switch. The 
hal code to do that is already sitting, commented out, in that machines 
hal file. With all the gearing down, the scale may run to 5 or 6 digits.  
Heck, my spindle is 5 digits in low gear. 14,xxx per spindle rev in low 
gear.

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)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[Gene Heskett]

On Wednesday 09 September 2020 12:54:45 andy pugh wrote:

> On Wed, 9 Sep 2020 at 17:10, Gene Heskett  wrote:
> > Is it possible to get an enable signal, per axis, from the mesa
> > cards that would become true, hopefully a couple micky-secs ahead of
> > the first stepper pulse so no missed steps would normally occur,
>
> I would look at the "locking indexer" pins.

And in that description it says no other axis can be moved while moving a 
locked axis.  And if thats true, then this whole idea is dead in the 
water and sunk. I want to use this to run the BS-1 while cutting a 
spiral gear so at minimum, it must move WITH X, and maybe Y too 
depending on the cutter. Hardware oneshots in the cable it is then.  If 
that doesn't work, buy a box of 2M542's unless there is a cheaper 
alternative.  Suggestions welcome as I already have a 24 volt supply for 
this.

Or a 48V I can turn down to 42.5.  That is what I'm powering the 
Sheldon's Z with these days, and that 3NM 3 phase stepper/servo fairly 
flies.

And someplace on a truck, is a 2NM version that will be driving the X 
axis about 2 days after it falls out of the truck in front of my place. 
A 3NM on the rear of the aspron won't clear the bed, but a 2 will. ;-) 
But I'll gain another piece of SS pipe for cableing on the right edge of 
the carriage too.

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)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[Jon Elson]

On 09/09/2020 12:16 PM, Gene Heskett wrote:

On Wednesday 09 September 2020 12:38:43 Jon Elson wrote:


On 09/09/2020 11:08 AM, Gene Heskett wrote:

Is it possible to get an enable signal, per axis, from the mesa
cards that would become true, hopefully a couple micky-secs ahead of
the first stepper pulse so no missed steps would normally occur,
losing its home, and go back false when the last step has been sent?
It maybe could be timed right if it actually steps on the trailing
edge of the pulse, and the enabling action was on the leading edge.

This is the TRICKY bit, knowing when an axis is SOON going
to start moving.

Jon

I'd be surprised if these things needed any lead time at all, Jon. The
hitachi docs on this chip do not mention any timing constraints in
regard to this. So its likely that it could be brought up to full song
with the leading edge of the pulse. If that triggered a retriggerable
oneshot that was then or'd with the pulse, and its output would be the
enable high current signal. Let the oneshot time out in 50 millisecs if
no pulse, putting it back in reduced current at the end of the move. If
the oneshot was in the mesa, would be ideal, and divorced from the
enable granularity of the 1khz servo thread.  That granularity could
easily lose enough steps to destroy a home position, something I intend
to have with this BS-1 clone.  A oneshot in the hal file would be a
disaster in that case because of the millisecond+ delay. A oneshot in
the cable to the driver would work, but if its in the mesa card, better
yet. :)


Certainly worth a try.  One horrible possibility is that the 
DIP switches are only read once at
power-up.  The first couple steps should be spaced out as 
the axis accelerates which should give the drive time to 
ramp up the current.


Jon


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[N]

> On 09/09/2020 11:08 AM, Gene Heskett wrote:
> > Is it possible to get an enable signal, per axis, from the mesa cards
> > that would become true, hopefully a couple micky-secs ahead of the first
> > stepper pulse so no missed steps would normally occur, losing its home,
> > and go back false when the last step has been sent? It maybe could be
> > timed right if it actually steps on the trailing edge of the pulse, and
> > the enabling action was on the leading edge.
> >
> >
> This is the TRICKY bit, knowing when an axis is SOON going 
> to start moving.

But it is possible to introduce a delay if needed. Use a translation of time so 
these things that should be done before is done immidiately while the others 
are delayed. Just be aware not to introduce any delay for feedback in control 
loops.


Regards Nicklas Karlsson


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[Gene Heskett]

On Wednesday 09 September 2020 12:38:43 Jon Elson wrote:

> On 09/09/2020 11:08 AM, Gene Heskett wrote:
> > Is it possible to get an enable signal, per axis, from the mesa
> > cards that would become true, hopefully a couple micky-secs ahead of
> > the first stepper pulse so no missed steps would normally occur,
> > losing its home, and go back false when the last step has been sent?
> > It maybe could be timed right if it actually steps on the trailing
> > edge of the pulse, and the enabling action was on the leading edge.
>
> This is the TRICKY bit, knowing when an axis is SOON going
> to start moving.
>
> Jon

I'd be surprised if these things needed any lead time at all, Jon. The 
hitachi docs on this chip do not mention any timing constraints in 
regard to this. So its likely that it could be brought up to full song 
with the leading edge of the pulse. If that triggered a retriggerable 
oneshot that was then or'd with the pulse, and its output would be the 
enable high current signal. Let the oneshot time out in 50 millisecs if 
no pulse, putting it back in reduced current at the end of the move. If 
the oneshot was in the mesa, would be ideal, and divorced from the 
enable granularity of the 1khz servo thread.  That granularity could 
easily lose enough steps to destroy a home position, something I intend 
to have with this BS-1 clone.  A oneshot in the hal file would be a 
disaster in that case because of the millisecond+ delay. A oneshot in 
the cable to the driver would work, but if its in the mesa card, better 
yet. :)

Stay safe and well, Jon.

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)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[andy pugh]

On Wed, 9 Sep 2020 at 17:10, Gene Heskett  wrote:

> Is it possible to get an enable signal, per axis, from the mesa cards
> that would become true, hopefully a couple micky-secs ahead of the first
> stepper pulse so no missed steps would normally occur,

I would look at the "locking indexer" pins.

-- 
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, 1912


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

Re: [Emc-users] Another possibly silly Q for Peter or Andy.

[Jon Elson]

On 09/09/2020 11:08 AM, Gene Heskett wrote:

Is it possible to get an enable signal, per axis, from the mesa cards
that would become true, hopefully a couple micky-secs ahead of the first
stepper pulse so no missed steps would normally occur, losing its home,
and go back false when the last step has been sent? It maybe could be
timed right if it actually steps on the trailing edge of the pulse, and
the enabling action was on the leading edge.


This is the TRICKY bit, knowing when an axis is SOON going 
to start moving.


Jon


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

[Emc-users] Another possibly silly Q for Peter or Andy.

[Gene Heskett]

Greetings;

I'd like to be able to use that quad pack of TB6560 based stepper drivers 
with these nema 17 driven harmonic drives I'm cobbling up.

But, these things are dumber than a rock in that they have a 4 step 
current adjustment, but no provision for an idling current turndown.  
And at full current, their heat will kill these harmonic drives.

The current adjustments are the right hand pair of dip switches, taking 
it from about 750ma when down to about 2.2 amps when both are up.  They 
feed two pins on the tb6560, with what looks like logic levels.  And 
ANAICT, the response to flipping the switches is instant, no powerdown 
reset needed such as the bigger drives like the 2M542's need.

I have I/O to do these with, so that isn't a concern. Developing the 
signal is. I'd hit both pins with the same signal, with the idea of 
banging the motor at the full 2.2 amps only if its moving, and turn it 
back down 50 millisecs after the last step to get it where we sent it to 
has been issued.

Is it possible to get an enable signal, per axis, from the mesa cards 
that would become true, hopefully a couple micky-secs ahead of the first 
stepper pulse so no missed steps would normally occur, losing its home, 
and go back false when the last step has been sent? It maybe could be 
timed right if it actually steps on the trailing edge of the pulse, and 
the enabling action was on the leading edge.

Can this be done?

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)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 


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