Re: [Emc-users] Servo and Mesa pulse encoder/velocity feedback

[Andy Pugh]

> On 18 Sep 2016, at 09:46, Frederic RIBLE  wrote:
> 
> This Fanuc drive will be in runaway condition if it does not receive a 
> velocity feedback.
> It has an internal analog closed loop

Internally it compares the feedback and command voltages. 
If both signals are generated by LinuxCNC from the same encoder in the same 
thread then you are using two analogue channels to provide a voltage difference 
when you could use one, at least in theory. 

As this is a lathe and you have 4 channels available the inefficiency probably 
doesn't matter. 



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

Re: [Emc-users] Servo and Mesa pulse encoder/velocity feedback

[Frederic RIBLE]

This Fanuc drive will be in runaway condition if it does not receive a 
velocity feedback.
It has an internal analog closed loop to track the velocity command and 
control the motor rpm.
This is why we absolutely need to derive this feedback from the 
quadrature encoder somewhere.

Frederic.

On 17/09/2016 23:02, andy pugh wrote:
> On 17 September 2016 at 20:24, Frederic RIBLE  wrote:
>
>> This is why we are thinking about using mesa to implement this function.
>
> It might be enough to "fool the drive" but there really is no information
> content in that velocity signal.
>


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

Re: [Emc-users] Servo and Mesa pulse encoder/velocity feedback

[Gene Heskett]

On Saturday 17 September 2016 17:02:51 andy pugh wrote:

> On 17 September 2016 at 20:24, Frederic RIBLE  wrote:
> > This is why we are thinking about using mesa to implement this
> > function.
>
> It might be enough to "fool the drive" but there really is no
> information content in that velocity signal.

I just looked again Andy, and what I see is a faster error feedback to 
the drive than it would get by going around thru the PID in the probably  
much slower because its much older OEM controller.

Am I imagining that would make a difference?  I have seen similar 
treatments in tracking and error correction in high end broadcast 
videotape machines, some more than 30 years old now.  Not wearing the 
same names for the function block but it seemed to me the principle was 
the same.  A faster feedback path is one way around the Nyquist 
stability diagrams. I am recalling a robotic 2" videotape player that 
kept 30 sec to 2 minute commercials on small cassettes that the operator 
could program 4 or 5 station breaks at a time, from a library robot 
feeding two player decks that were self threading.  In addition to the 
robotic load and unload back to a storage rack, it had a several 
kilowatt rated headwheel motor drivers which supplied some of its magic. 
Both headwheels sitting dead, push the start button, the lights blinked 
and the wheel was turning 14,400 rpm in about half a second. Time base 
correctors locked to house synch could output a color phase locked 
picture in under a second, but you could also have the tapes vacuum 
shoes retracted 10 thou so the flying heads wouldn't saw the stopped 
tape in two, which allowed the wheels to be started a minute early, so 
that when the start signal came in, the shoes closed in 5 milliseconds, 
and one frame of video later (16 milliseconds) it was color locked and 
the commercial was playing.

It was a hungry machine, needing an additional 25 kw in AC for every such 
machine in use.  The technology would be considered ancient history 
today, and it cost north of a million in 1975 dollars.  But there was a 
need, and it worked well enough the stations that could afford it, 
bought it in pairs.

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] Servo and Mesa pulse encoder/velocity feedback

[andy pugh]

On 17 September 2016 at 20:24, Frederic RIBLE  wrote:

> This is why we are thinking about using mesa to implement this function.


It might be enough to "fool the drive" but there really is no information
content in that velocity signal.

-- 
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] Servo and Mesa pulse encoder/velocity feedback

[Frederic RIBLE]

The diagram is here: 
https://forum.linuxcnc.org/media/kunena/attachments/16648/Mauriceservo.jpg

Actually, modifying the drive to get access to torque mode is something 
we would like to avoid.
Designing an external pulse to velocity converter as well.
This is why we are thinking about using mesa to implement this function.
Depending on latency for this operation, we may experience additional 
stability issue in the loop.

Frederic.


On 17/09/2016 15:43, andy pugh wrote:
> On 17 September 2016 at 10:20, Frederic RIBLE  wrote:
>
>> We would like to replace the controller with a Mesa 7i48/6i24 combination.
>> We plan to use one pulse encoder input to make the frequency to voltage
>> conversion and generate the velocity feedback for the drive.
>>
> You could use two output channels per drive from any of the servo control
> cards, one for velocity command, and one for velocity feedback.
> But, when you think about it, both are updated from the same source at the
> same time, and there is no extra information in the second channel.
>
> What you might find works perfectly well is to short out one or other of
> the command and feedback channels and to control the other conventionally.
>
> The only way to gain any benefit from the internal velocity loop in the
> drive is if the velocity feedback updates more rapidly than the LinuxCNC
> command channel. And the only ay I see to do that is with an external
> hardware board, as originally used.
>


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

Re: [Emc-users] Servo and Mesa pulse encoder/velocity feedback

[Frederic RIBLE]

Yes, the diagram has been dropped by the mailing list software.

You can find it here: 
https://forum.linuxcnc.org/media/kunena/attachments/16648/Mauriceservo.jpg

Frederic.

On 17/09/2016 15:32, Gene Heskett wrote:
> On Saturday 17 September 2016 05:20:09 Frederic RIBLE wrote:
>
>> Hello,
>>
>> We are working on a retrofit for a Mori Seiki SL-0 CNC lathe.
>> The original X/Z axis servo drives are Fanuc analog.
>> They work in velocity mode with external frequency to voltage
>> conversion embedded in the original CNC controller.
>> We would like to replace the controller with a Mesa 7i48/6i24
>> combination. We plan to use one pulse encoder input to make the
>> frequency to voltage conversion and generate the velocity feedback for
>> the drive.
>> See attached diagram.
> No attached diagram came thru, Frederic, perhaps too big?
>
>> Anyone having success with similar setup? Does it make sense?
>> Do we have to anticipate specific stability problem in the closed
>> loop?
>>
>> Thanks.
>> Frederic (http://blog.f1oat.org/)
>>
>> --
>>  ___
>> Emc-users mailing list
>> Emc-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/emc-users
>
> Cheers, Gene Heskett


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

Re: [Emc-users] Servo and Mesa pulse encoder/velocity feedback

[andy pugh]

On 17 September 2016 at 10:20, Frederic RIBLE  wrote:

>
> We would like to replace the controller with a Mesa 7i48/6i24 combination.
> We plan to use one pulse encoder input to make the frequency to voltage
> conversion and generate the velocity feedback for the drive.
>

You could use two output channels per drive from any of the servo control
cards, one for velocity command, and one for velocity feedback.
But, when you think about it, both are updated from the same source at the
same time, and there is no extra information in the second channel.

What you might find works perfectly well is to short out one or other of
the command and feedback channels and to control the other conventionally.

The only way to gain any benefit from the internal velocity loop in the
drive is if the velocity feedback updates more rapidly than the LinuxCNC
command channel. And the only ay I see to do that is with an external
hardware board, as originally used.

-- 
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] Servo and Mesa pulse encoder/velocity feedback

[Gene Heskett]

On Saturday 17 September 2016 05:20:09 Frederic RIBLE wrote:

> Hello,
>
> We are working on a retrofit for a Mori Seiki SL-0 CNC lathe.
> The original X/Z axis servo drives are Fanuc analog.
> They work in velocity mode with external frequency to voltage
> conversion embedded in the original CNC controller.
> We would like to replace the controller with a Mesa 7i48/6i24
> combination. We plan to use one pulse encoder input to make the
> frequency to voltage conversion and generate the velocity feedback for
> the drive.
> See attached diagram.

No attached diagram came thru, Frederic, perhaps too big?

> Anyone having success with similar setup? Does it make sense?
> Do we have to anticipate specific stability problem in the closed
> loop?
>
> Thanks.
> Frederic (http://blog.f1oat.org/)
>
> --
> ___
> 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