RE: [CAD_CAM_EDM_DRO] Re: breakout boards plus - hard choice

Hi Ron



Yes it's down to how people perceive things. I am surprised though you see 2
full motor poles of lag, I think you're the guy with the positional control
board feedback board so this is data you would have, somehow I had the name
Robin stuck in my brain.

I thought though that 2 poles lag was the point at which you normally
irrecoverably loose a step, and normally you limit acceleration to stop
loosing pulses with a basic stepper drive which does not have an error
counter and cannot move onto the next step unless the previous one is
committed and almost completed by the motor. Mariss isn't the only one that
allows steppers pulse to lag with an error counter so it can slow and catch
up, whilst steppers aren't my field, my brother has been fitting stepper
drives in some of our CNC grinder and strap cutting machines over ten years
ago whose manufacturers claimed it was impossible to loose pulses and whom
used an error counter. Sadly their product back then wasn't as good as they
claimed, it worked well on small square motors or the modern variety, but we
also had a massive old round motor to drive and that could quite easily
loose pulses, round steppers have though always been a damn nuisance.



As to rounding of corners and the like, it's all down to how hard you push
it, and how long you give it to catch up at the end of a turn. On a servo
system the servo resolution is often great enough that a large error count
is a very small distance, we expect an error of perhaps 100 counts and
commonly get a system to stay within 40 to 60. With steppers people jump in
to use very fine microstepping, but the perception is that errors are small
or you loose steps, so I feel there is potential here for users to mislead
themselves and have a far greater error in terms of real distance than they
expected.

It would be really nice though if we had simple controls for people to
feedback encoder signal so an allowable error could be mixed with the
federate to control accuracy, you'd then have control over the precision
before you start the job. Some servo cards like CNC teknix have an analogue
error output that could easily be summed and fed directly fed to an analogue
federate compensation input. A shame their cards couldn't drive my motors.

This closing the loop issue is something I try to keep nibbling away at so
perhaps the G100, mach or whatever may one day gain such a feature for this
budget field.



Regs Mark





Dr. Mark Vaughan Ph'D. B.Eng. M0VAU

Managing Director

Vaughan Industries Ltd, reg in UK no 2561068

Water Care Technology Ltd, reg in UK no 4129351

Addr Unit3, Sydney House, Blackwater, Truro, Cornwall, TR4 8HH, UK.
Phone/Fax 44 1872 561288

RSGB DRM111(Cornwall)

_____

From: CAD_CAM_EDM_DRO@yahoogroups.com
[mailto:CAD_CAM_EDM_DRO@yahoogroups.com] On Behalf Of R Rogers
Sent: 19 July 2007 19:56
To: CAD_CAM_EDM_DRO@yahoogroups.com
Subject: RE: [CAD_CAM_EDM_DRO] Re: breakout boards plus - hard choice



Mark,

You wrote: This makes steppers and servos very similar now, so they both can
be working
lagging behind the true commanded position.

They already do. I have around ten encoder equipped stepper based machines
out in the field and they all require a following error allowance lest they
they trip the DLL. If what you're claiming is true many problems in
application would not exist like "corner rounding". The event where an axis
cannot keep up with the steps. Those machines arent losing ultimate
position, simply falling behind the commanded.

In my own testing with steppers, I've seen them lag 2 full motor poles (3.2
degrees) and recover. The theory that they presently execute each step
instantaneously is virtually impossible in regard to wide speed ranges and
acceleration.

Your last statement sums it up, its reliant on the design. If someone is
setting up and expecting to spin a stepper in mutli-k RPM range, they'll be
disappointed. If one needs high RPM with torque, they need a servo.

Ron

Mark Vaughan <mark@.... <mailto:mark%40vil.uk.com> com> wrote:
Not quite true Ron.

Servo's receive a command of many steps, which they stack into a counter,
this value in the counter means the motor is not at the correct position, so
based on a PID control systems formulae they then apply effort to the motor
to get it to move in the correct direction. As it moves the counter is
decremented until it becomes zero and the motor stops. Of course in
operation additional pulses are added to the motor is always trying to keep
up. If the counter gets above a certain value the servo trips with a follow
up error. You try to design servo systems with many more steps than you need
accuracy so the counter value can be quite a hi number and it is still
within your tolerance. So they lag but the lag has limits and any error is
recoverable up to the trip point.

Steppers are a little different, normally they try to immediately respond to
the step command. This sounds better than a servo to begin with, but normal
steppers are not unstallable and if too much load is applied they will loose
position steps which is not normally recoverable. Steppers have a different
torque response to servo's. With servo's the torque is related to how much
current you can apply, so slow or fast the have about the same torque.
Steppers have lots of torque at low rpms and loose torque as you speed them
up. Mariss's theory is that if while moving fast the load gets too high, you
can slow them down and they will have more torque. He is trying to run a
stepper like a servo drive, so in order to slow them down he adds data to an
error counter, rather than decrementing that error with an encoder, he
decrements it by electronically detecting that a step pulse has been
successfully implemented in the motor.

This makes steppers and servos very similar now, so they both can be working
lagging behind the true commanded position. If the step resolution is
similar then the response from each will be similar, however I thing the
stepper steps will typically be much bigger so an error will be bigger and
can accrue to a large value, not only that, but because or the torque
response it is possible that that error will be a lot bigger than a servo's
error.

All down to how we weigh it up in our designs really.

Regs Mark

Dr. Mark Vaughan Ph'D. B.Eng. M0VAU

Managing Director

Vaughan Industries Ltd, reg in UK no 2561068

Water Care Technology Ltd, reg in UK no 4129351

Addr Unit3, Sydney House, Blackwater, Truro, Cornwall, TR4 8HH, UK.
Phone/Fax 44 1872 561288

RSGB DRM111(Cornwall)

_____

From: CAD_CAM_EDM_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>
DRO@yahoogroups.com
[mailto:CAD_CAM_EDM_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>
DRO@yahoogroups.com] On Behalf Of R Rogers
Sent: 19 July 2007 18:13
To: CAD_CAM_EDM_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>
DRO@yahoogroups.com
Subject: Re: [CAD_CAM_EDM_DRO] Re: breakout boards plus - hard choice

Peter,

Yeah I do think the unstallable stepper is being developed for the G100.
Sounds like a good system. Personally, I'm not sure it's the direction to go
though.

If I understand it correctly, current stepper and servo drives are all
unstallable in a sense, that they allow a certain following error. The
ability of the motors to lag behind the commanded position and not lose
position ultimately.

A properly setup and operated system is very reliable in regards to
positional accuracy. The only time this is an issue is when some unusual
event takes place thereby restricting an axis for whatever reason. Something
not anticipated. Analytically, is it best to try and overcome the obstacle,
or stop and see what the problem is. The better logic says to stop and see
what the problem is. If we're backing out of the driveway and run over the
kids bike, do we drop down in a lower gear and over come the increased load
:), or do we stop and get the kids bike out, eliminating further damage to
his vehicle and our own?

A system like this has some applications, for basic 3 axis cartesian setups,
I don't see it.

I know I'm biased in this because I sell the encoder board. And I've been
told by many, "thats all it does is stop the machine" Thats by design and
decision, not limitation. When Brian Barker and I were discussing the design
of the DLL plug in, positional correction on the fly was mentioned. Properly
setup digital systems don't need it. Analog systems do. And all hobby and
alot of commercial systems nowadays are digital.

Ron

Peter Homann <groups@homanndesign <mailto:groups%40homanndesigns.com> s.com>
wrote:
Mark,

A couple of things. I believe that the unstallable stepper is designed to
work
with the G100 which will slow down the feed rate on all axes together.

If not used in a syncronised system, it could work independently faulting
once
the stepper has lost say, 256 steps. This is the same as any servo system.

I don't believe that the intention is that you set the feed rate to warp
factor 9. You would still have to set an appropriate rate. The stepper would

then be able to recover from the occasional lost step.

Cheers,

Peter

Mark Vaughan wrote:

> The worrying bit Graham is while one axis slows, what are the others

doing.

>
> To work they all need to be tied together, and by quite a complex set of
> formulae to maintain the shape accuracy.
>
> At present you won't loose steps, but you will still suffer an error in

the

> work, so it becomes debatable whether there is any major benefit.
>
>
>
> Regs Mark
>
>
>
> Dr. Mark Vaughan Ph'D. B.Eng. M0VAU
>
> Managing Director
>
> Vaughan Industries Ltd, reg in UK no 2561068
>
> Water Care Technology Ltd, reg in UK no 4129351
>
> Addr Unit3, Sydney House, Blackwater, Truro, Cornwall, TR4 8HH, UK.
> Phone/Fax 44 1872 561288
>
> RSGB DRM111(Cornwall)
>

--
----------------------------------------------------------
Web: www.homanndesigns.com
email: homann@homanndesign <mailto:homann%40homanndesigns.com> s.com
Phone: +61 421 601 665
www.homanndesigns.com/ModIO.html - Modbus Interface Unit
www.homanndesigns.com/DigiSpeedDeal.html - DC Spindle control
www.homanndesigns.com/TurboTaig.html - Taig Mill Upgrade board

[Non-text portions of this message have been removed]

[Non-text portions of this message have been removed]

[Non-text portions of this message have been removed]





[Non-text portions of this message have been removed]