Re: [Emc-users] New stepper-servo --> Phase delay

[dannym]

They put a Cortex M0 on the module that sits on the board, along with the 
stepper driver and encoder.

There's a greater latency inside the Cortex, it will likely have to respond to 
an interrupt thus have ISR latency.  I presume they've got a PLL to prevent 
oscillation within the final step.

It's not a bad case for an FPGA, since with proper coding the latency is 
limited by the response time of the transistors.

Danny


 Nicklas Karlsson  wrote: 
> Best option is probably to generate stepper signal from hardware timer.
> 
> 2016-10-21 8:17 GMT+02:00 Danny Miller :
> 
> > The SPI comm of course has latency within a comm period.
> >
> > However, the pulse output must be super-fast.  It supports up to 14500
> > rpm, so if it's 2000steps/rev the whole step cycle is 2ms, realistically
> > like maybe 1/4 of that for latency?
> >
> > Danny
> >
> > On 10/19/2016 11:34 PM, Nicklas Karlsson wrote:
> > >>> ... The lag also seemed to vary with velocity.
> > >>>
> > >>>
> > >> Most specifically, it was a lag in the encoder's loop to respond to
> > >> acceleration.
> > >> Velocity was fine, but when there is acceleration, the encoder's
> > >> velocity didn't change for a few milliseconds, then it had to produce
> > >> velocity greater than real for position to catch up.
> > >> I thought it was a really severe problem, but it seems it may not be all
> > >> that bad.  But, it can make servo tuning more tricky.
> > >>
> > >> Jon
> > > Delay is called phase in control theory and it is not good at all for
> > fast response.
> > >
> > > 
> > --
> > > Check out the vibrant tech community on one of the world's most
> > > engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> > > ___
> > > Emc-users mailing list
> > > Emc-users@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > >
> >
> >
> > 
> > --
> > Check out the vibrant tech community on one of the world's most
> > engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
> --
> Check out the vibrant tech community on one of the world's most 
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo --> Phase delay

[Nicklas Karlsson]

Best option is probably to generate stepper signal from hardware timer.

2016-10-21 8:17 GMT+02:00 Danny Miller :

> The SPI comm of course has latency within a comm period.
>
> However, the pulse output must be super-fast.  It supports up to 14500
> rpm, so if it's 2000steps/rev the whole step cycle is 2ms, realistically
> like maybe 1/4 of that for latency?
>
> Danny
>
> On 10/19/2016 11:34 PM, Nicklas Karlsson wrote:
> >>> ... The lag also seemed to vary with velocity.
> >>>
> >>>
> >> Most specifically, it was a lag in the encoder's loop to respond to
> >> acceleration.
> >> Velocity was fine, but when there is acceleration, the encoder's
> >> velocity didn't change for a few milliseconds, then it had to produce
> >> velocity greater than real for position to catch up.
> >> I thought it was a really severe problem, but it seems it may not be all
> >> that bad.  But, it can make servo tuning more tricky.
> >>
> >> Jon
> > Delay is called phase in control theory and it is not good at all for
> fast response.
> >
> > 
> --
> > Check out the vibrant tech community on one of the world's most
> > engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
>
>
> 
> --
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo --> Phase delay

[Danny Miller]

The SPI comm of course has latency within a comm period.

However, the pulse output must be super-fast.  It supports up to 14500 
rpm, so if it's 2000steps/rev the whole step cycle is 2ms, realistically 
like maybe 1/4 of that for latency?

Danny

On 10/19/2016 11:34 PM, Nicklas Karlsson wrote:
>>> ... The lag also seemed to vary with velocity.
>>>
>>>
>> Most specifically, it was a lag in the encoder's loop to respond to
>> acceleration.
>> Velocity was fine, but when there is acceleration, the encoder's
>> velocity didn't change for a few milliseconds, then it had to produce
>> velocity greater than real for position to catch up.
>> I thought it was a really severe problem, but it seems it may not be all
>> that bad.  But, it can make servo tuning more tricky.
>>
>> Jon
> Delay is called phase in control theory and it is not good at all for fast 
> response.
>
> --
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo --> Phase delay

[Nicklas Karlsson]

> > ... The lag also seemed to vary with velocity.
> >
> >
> Most specifically, it was a lag in the encoder's loop to respond to 
> acceleration.
> Velocity was fine, but when there is acceleration, the encoder's 
> velocity didn't change for a few milliseconds, then it had to produce 
> velocity greater than real for position to catch up.
> I thought it was a really severe problem, but it seems it may not be all 
> that bad.  But, it can make servo tuning more tricky.
> 
> Jon

Delay is called phase in control theory and it is not good at all for fast 
response.

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[Jon Elson]

On 10/19/2016 01:08 PM, Kirk Wallace wrote:
> On 10/19/2016 10:53 AM, andy pugh wrote:
>> On 19 October 2016 at 18:34, Roland Jollivet  
>> wrote:
>>
>>> While the devices have a high resolution, look at pg.7 for all the
>>> non-linearity specs. So while you may discern 16 383 positions, you may be
>>> well over 50 counts in error.
>> Possibly fixable in software. Especially with a stepper (we don't
>> normally worry about whether a _stepper_ is linear)
> Another issue that might come about is encoder latency. I may be wrong
> but, I seem to recall Jon ran into this problem with capacitive encoders
> which had a lag between the position and the sending of the position
> signal. The lag also seemed to vary with velocity.
>
>
Most specifically, it was a lag in the encoder's loop to respond to 
acceleration.
Velocity was fine, but when there is acceleration, the encoder's 
velocity didn't change for a few milliseconds, then it had to produce 
velocity greater than real for position to catch up.
I thought it was a really severe problem, but it seems it may not be all 
that bad.  But, it can make servo tuning more tricky.

Jon


--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[W. Martinjak]

On 2016-10-19 12:08, Gene Heskett wrote:
> That will give 16,383 discrete positions of 
> the screw per revolution of the screw, 

No, the max pulses are 1024 and the max steps are 4096. (pulses x 4)

Putting the magnet on the pointer of an old dial gauge you have a nice surface 
feeler for
measuring the tools zero radius offset. :)



-- 
"In der Wissenschaft siegt nie eine neue Theorie,
nur ihre Gegner sterben nach und nach"

Max Planck


--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[Kirk Wallace]

On 10/19/2016 10:53 AM, andy pugh wrote:
> On 19 October 2016 at 18:34, Roland Jollivet  
> wrote:
>
>> While the devices have a high resolution, look at pg.7 for all the
>> non-linearity specs. So while you may discern 16 383 positions, you may be
>> well over 50 counts in error.
>
> Possibly fixable in software. Especially with a stepper (we don't
> normally worry about whether a _stepper_ is linear)

Another issue that might come about is encoder latency. I may be wrong 
but, I seem to recall Jon ran into this problem with capacitive encoders 
which had a lag between the position and the sending of the position 
signal. The lag also seemed to vary with velocity.


-- 
Kirk Wallace
http://www.wallacecompany.com/machine_shop/
http://www.wallacecompany.com/E45/

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] New stepper-servo

[Roland Jollivet]

On 19 October 2016 at 19:53, andy pugh  wrote:

> On 19 October 2016 at 18:34, Roland Jollivet 
> wrote:
>
> > While the devices have a high resolution, look at pg.7 for all the
> > non-linearity specs. So while you may discern 16 383 positions, you may
> be
> > well over 50 counts in error.
>
> Possibly fixable in software. Especially with a stepper (we don't
> normally worry about whether a _stepper_ is linear)
>


Sure, but Gene was contemplating attaching them to the ballscrews on his
machine.
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[andy pugh]

On 19 October 2016 at 18:34, Roland Jollivet  wrote:

> While the devices have a high resolution, look at pg.7 for all the
> non-linearity specs. So while you may discern 16 383 positions, you may be
> well over 50 counts in error.

Possibly fixable in software. Especially with a stepper (we don't
normally worry about whether a _stepper_ is linear)


-- 
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

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] New stepper-servo

[Roland Jollivet]

hi Gene

While those AS devices are alluring, I think they're pretty useless as
encoders proper. It bothers me that the datasheet does not say in bold,
"About +- 1 degree accurate only!"

While the devices have a high resolution, look at pg.7 for all the
non-linearity specs. So while you may discern 16 383 positions, you may be
well over 50 counts in error. I think their target market is accelerator
pedals ;-)

Regards
Roland




On 19 October 2016 at 12:08, Gene Heskett  wrote:

> On Wednesday 19 October 2016 01:01:18 Danny Miller wrote:
>
> > A friend showed me this today, not powered up yet:
> >
> > https://www.kickstarter.com/projects/tropicallabs/mechaduino-powerful-
> >open-source-industrial-servo-m
> >
> > At first I was confused, it says it's a stepper, but with an encoder,
> > and lowers or disables drive current when not needed.  Or you can
> > freewheel it and it'll maintain the coordinate system.  It cannot
> > stall without the system knowing, and a stall won't corrupt the
> > coordinate system.
> >
> > But there's no mechanical connection to the motor, it's this new
> > AS5047D high speed, high resolution magnetic rotary position sensor:
> >
> > http://ams.com/eng/Products/Magnetic-Position-Sensors/Angle-Position-O
> >n-Axis/AS5047D
> >
> > You glue a magnet to the rear of the motor shaft, and keep the sensor
> > like 1/4" away.  Note it not only counts delta, it knows absolute
> > rotor position.
> >
> > Gives 2000 counts/rev.  That's enough that it could effectively show
> > rotor phase through a single fullstep, which would allow for more
> > stable electrical control of stepping.  There's an initial calibration
> > phase that drives a phase so it knows where the TDC of Coil A is in
> > absolute position.
> >
> > They set it up so it powers down and freewheels, except it will hold
> > its position by powering coils when needed.  It can also reduce
> > current to product only the torque needed to follow the step commands,
> > instead of always operating at full current.
> >
> > Well, my mind is blown!
> >
> > Danny
> >
> That does indeed sound very handy for stepper users. But the sensor
> magnet I do not envision as being on the motor, but on the ends of the
> screws driving the machine. That will give 16,383 discrete positions of
> the screw per revolution of the screw, irrespective of any stepdowns
> from belt drives etc between the motor and the screw. I did not see a
> price for the completed board and magnet, just for the AMS device.  Ahh,
> did find it on the pre-order button, $47 a copy for version 0.1.
>
> Did anyone else? If not too outrageous, I might be interested in a pair
> of them to put on this Sheldon conversion as a test bed. But I'll do the
> conventional way first & see where it falls over.
>
> What I would really like to see would be a light beam that a tool could
> be advanced into in order to both adjust the tool tip height, and set
> the tools zero radius offset. Something whose shadow could be detected
> by a 4 cell photodetector in the tailstock socket.  Or even a low res
> cmos imager would suffice for the detector.  Almost anything to get us
> into the +- .005" range for a first pass calibration cut.
>
> Comments? Gotchas?
>
> 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 
>
> 
> --
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[Sarah Armstrong]

I,m actually working on a unit to do exactly that Gene. 

Along with tool breakage sense 
Which is tending to be the pain, so I might make 2 units 

Sarah 

⁣

Sent from BlueMail

​

On 19 Oct 2016, 11:13, at 11:13, Gene Heskett  wrote:
>On Wednesday 19 October 2016 01:01:18 Danny Miller wrote:
>
>> A friend showed me this today, not powered up yet:
>>
>>
>https://www.kickstarter.com/projects/tropicallabs/mechaduino-powerful-
>>open-source-industrial-servo-m
>>
>> At first I was confused, it says it's a stepper, but with an encoder,
>> and lowers or disables drive current when not needed.  Or you can
>> freewheel it and it'll maintain the coordinate system.  It cannot
>> stall without the system knowing, and a stall won't corrupt the
>> coordinate system.
>>
>> But there's no mechanical connection to the motor, it's this new
>> AS5047D high speed, high resolution magnetic rotary position sensor:
>>
>>
>http://ams.com/eng/Products/Magnetic-Position-Sensors/Angle-Position-O
>>n-Axis/AS5047D
>>
>> You glue a magnet to the rear of the motor shaft, and keep the sensor
>> like 1/4" away.  Note it not only counts delta, it knows absolute
>> rotor position.
>>
>> Gives 2000 counts/rev.  That's enough that it could effectively show
>> rotor phase through a single fullstep, which would allow for more
>> stable electrical control of stepping.  There's an initial
>calibration
>> phase that drives a phase so it knows where the TDC of Coil A is in
>> absolute position.
>>
>> They set it up so it powers down and freewheels, except it will hold
>> its position by powering coils when needed.  It can also reduce
>> current to product only the torque needed to follow the step
>commands,
>> instead of always operating at full current.
>>
>> Well, my mind is blown!
>>
>> Danny
>>
>That does indeed sound very handy for stepper users. But the sensor 
>magnet I do not envision as being on the motor, but on the ends of the 
>screws driving the machine. That will give 16,383 discrete positions of
>
>the screw per revolution of the screw, irrespective of any stepdowns 
>from belt drives etc between the motor and the screw. I did not see a 
>price for the completed board and magnet, just for the AMS device. 
>Ahh, 
>did find it on the pre-order button, $47 a copy for version 0.1.
>
>Did anyone else? If not too outrageous, I might be interested in a pair
>
>of them to put on this Sheldon conversion as a test bed. But I'll do
>the 
>conventional way first & see where it falls over.
>
>What I would really like to see would be a light beam that a tool could
>
>be advanced into in order to both adjust the tool tip height, and set 
>the tools zero radius offset. Something whose shadow could be detected 
>by a 4 cell photodetector in the tailstock socket.  Or even a low res 
>cmos imager would suffice for the detector.  Almost anything to get us 
>into the +- .005" range for a first pass calibration cut.
>
>Comments? Gotchas?
>
>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 
>
>--
>Check out the vibrant tech community on one of the world's most 
>engaging tech sites, SlashDot.org! http://sdm.link/slashdot
>___
>Emc-users mailing list
>Emc-users@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/emc-users
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[W. Martinjak]

On 2016-10-19 07:01, Danny Miller wrote:
> But there's no mechanical connection to the motor, it's this new AS5047D 
> high speed, high resolution magnetic rotary position sensor:
>
> http://ams.com/eng/Products/Magnetic-Position-Sensors/Angle-Position-On-Axis/AS5047D

It's just austrian art of engineering, no more.
It's no rocket science.
Because austrians are not allowed to build rockets. :D

-- 
"In der Wissenschaft siegt nie eine neue Theorie,
nur ihre Gegner sterben nach und nach"

Max Planck


--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] New stepper-servo

[Gene Heskett]

On Wednesday 19 October 2016 01:01:18 Danny Miller wrote:

> A friend showed me this today, not powered up yet:
>
> https://www.kickstarter.com/projects/tropicallabs/mechaduino-powerful-
>open-source-industrial-servo-m
>
> At first I was confused, it says it's a stepper, but with an encoder,
> and lowers or disables drive current when not needed.  Or you can
> freewheel it and it'll maintain the coordinate system.  It cannot
> stall without the system knowing, and a stall won't corrupt the
> coordinate system.
>
> But there's no mechanical connection to the motor, it's this new
> AS5047D high speed, high resolution magnetic rotary position sensor:
>
> http://ams.com/eng/Products/Magnetic-Position-Sensors/Angle-Position-O
>n-Axis/AS5047D
>
> You glue a magnet to the rear of the motor shaft, and keep the sensor
> like 1/4" away.  Note it not only counts delta, it knows absolute
> rotor position.
>
> Gives 2000 counts/rev.  That's enough that it could effectively show
> rotor phase through a single fullstep, which would allow for more
> stable electrical control of stepping.  There's an initial calibration
> phase that drives a phase so it knows where the TDC of Coil A is in
> absolute position.
>
> They set it up so it powers down and freewheels, except it will hold
> its position by powering coils when needed.  It can also reduce
> current to product only the torque needed to follow the step commands,
> instead of always operating at full current.
>
> Well, my mind is blown!
>
> Danny
>
That does indeed sound very handy for stepper users. But the sensor 
magnet I do not envision as being on the motor, but on the ends of the 
screws driving the machine. That will give 16,383 discrete positions of 
the screw per revolution of the screw, irrespective of any stepdowns 
from belt drives etc between the motor and the screw. I did not see a 
price for the completed board and magnet, just for the AMS device.  Ahh, 
did find it on the pre-order button, $47 a copy for version 0.1.

Did anyone else? If not too outrageous, I might be interested in a pair 
of them to put on this Sheldon conversion as a test bed. But I'll do the 
conventional way first & see where it falls over.

What I would really like to see would be a light beam that a tool could 
be advanced into in order to both adjust the tool tip height, and set 
the tools zero radius offset. Something whose shadow could be detected 
by a 4 cell photodetector in the tailstock socket.  Or even a low res 
cmos imager would suffice for the detector.  Almost anything to get us 
into the +- .005" range for a first pass calibration cut.

Comments? Gotchas?

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 

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] New stepper-servo

[Danny Miller]

A friend showed me this today, not powered up yet:

https://www.kickstarter.com/projects/tropicallabs/mechaduino-powerful-open-source-industrial-servo-m

At first I was confused, it says it's a stepper, but with an encoder, 
and lowers or disables drive current when not needed.  Or you can 
freewheel it and it'll maintain the coordinate system.  It cannot stall 
without the system knowing, and a stall won't corrupt the coordinate system.

But there's no mechanical connection to the motor, it's this new AS5047D 
high speed, high resolution magnetic rotary position sensor:

http://ams.com/eng/Products/Magnetic-Position-Sensors/Angle-Position-On-Axis/AS5047D

You glue a magnet to the rear of the motor shaft, and keep the sensor 
like 1/4" away.  Note it not only counts delta, it knows absolute rotor 
position.

Gives 2000 counts/rev.  That's enough that it could effectively show 
rotor phase through a single fullstep, which would allow for more stable 
electrical control of stepping.  There's an initial calibration phase 
that drives a phase so it knows where the TDC of Coil A is in absolute 
position.

They set it up so it powers down and freewheels, except it will hold its 
position by powering coils when needed.  It can also reduce current to 
product only the torque needed to follow the step commands, instead of 
always operating at full current.

Well, my mind is blown!

Danny



--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users