the only significant difference is that steppers can loose
steps and position, servos don't and servos have a higher torque at
higher speeds

Besides losing step and position (which has a minor irritation drift) ,
I fail to see how the other axes on a stepper based system will know if
one axis stalled completely for example. And here I'm not referring to
hitting some extreme where there's a limit switch. Suppose the Z axis
on a CNC drill or mill got jammed in the workpiece won't the other axes
continue until they too are forcefully disabled if an operator isn't
around? Do stepper drives have a "fault bus" like servo drives?

wrote in news:1127495435.696903.21470
@g14g2000cwa.googlegroups.com:

the only significant difference is that steppers can loose
steps and position, servos don't and servos have a higher torque at
higher speeds

Besides losing step and position (which has a minor irritation drift) ,
I fail to see how the other axes on a stepper based system will know if
one axis stalled completely for example. And here I'm not referring to
hitting some extreme where there's a limit switch. Suppose the Z axis
on a CNC drill or mill got jammed in the workpiece won't the other axes
continue until they too are forcefully disabled if an operator isn't
around? Do stepper drives have a "fault bus" like servo drives?

Totally correct, one axis loosing position will not stop the other axis
in the movement. No fault bus at all, but the step/direction servo drives
will fault and some software will halt the motion of all axis if an axis
faults, some will not. Again this is the cheaper end of the scale and not
as comprehensive as the big stuff.

I don't know of *any* metalworking machines that will scoot with such tiny
amounts power

Both scoot and tiny are relatives. Whatever, there's woodworking,
robotics and several other fields where lower torque levels are
commonplace.

150 watts for $80, vs Gecko's 1600 watts for $120, I'd choose the latter.

Its more like $195.00 vs $360.00 if you're catering to 3 axes.

It *is* all about the oomph, you know

No, I don't know. How old are you? You seem like a impressionable
teenager who drives a car with one of those 6" tail pipes that's good
for nothing but noise.

No fault bus at all, but the step/direction servo drives
will fault and some software will halt the motion of all axis if an axis
faults, some will not

This is where the MCU based servo drives shine and are destined to be
the state of the art IMO. Software is not required to halt operation
when any axes faults. In fact the software doesn't even know a fault
bus exists. Each drive is connected to the fault bus via open drain
inputs. When any drive detects a fault associated with the motor it is
driving, it will switch its input to output and pull the bus low
causing all other drives to fault. This all takes place in 3 or 4 clock
cyles so it's practically instantaneous.

wrote in news:1127499300.458419.75650
@o13g2000cwo.googlegroups.com:

No fault bus at all, but the step/direction servo drives
will fault and some software will halt the motion of all axis if an axis
faults, some will not

This is where the MCU based servo drives shine and are destined to be
the state of the art IMO. Software is not required to halt operation
when any axes faults. In fact the software doesn't even know a fault
bus exists. Each drive is connected to the fault bus via open drain
inputs. When any drive detects a fault associated with the motor it is
driving, it will switch its input to output and pull the bus low
causing all other drives to fault. This all takes place in 3 or 4 clock
cyles so it's practically instantaneous.


Yes, agreed. I've tried to do something similar and link each drivers state
into the E stop. Crude at best, but while it still might break an endmill
it would save the part if you where lucky. Gecko is/has? developed a box
that sits between the controller and the drive with an MCU that does most
of what you suggest along with other stuff.

"Richard J Kinch" wrote in message
. ..

http://www.embedded.com/2000/0010/0010feat3.htm

Thanks for that. It explains the control problem quite well.

So, taking all that was said before, I take it the 320 solves to the
commanded position by controlling the servo's speed. An evenly spaced step
pulse train results in smooth movement at the target speed, which settles in
quickly if the gains are adjusted correctly. The smooth movement doesn't
suffer from a stepper's herky jerky motion, and is thus immune from the
related stepper resonance issues. The discrete nature of the step pulses and
the encoder stream would seem to be inherently noisy, but is easily averaged
out in steady state motion. OTOH, acceleration and deceleration would seem
strongly dependent on the controller being able to maintain sync, rather
than solely on the motor's available torque as is the case with a stepper. I
don't know how big the difference is, or if it's even measurable or
objectionable in CNC use.

So, how much can I expect to pay for a 400 oz-in servo?

According to :
I'm not sure what it was you were trying to say.

Based on your last post, I think it suffices to say that there is no
simple way under the sun you can use a Gecko 320 to drive a stepper
motor even if it was equipped with an encoder. Repeating....It is a DC
servo drive.

Yes -- it allows a DC servo motor to look like a stepper to the
controller.

Enjoy,
DoN.

--
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

Mike Young writes:

The discrete nature of the step pulses and
the encoder stream would seem to be inherently noisy, but is easily
averaged out in steady state motion.

Dunno what you mean by noise. There is no noise of any consequence.

OTOH, acceleration and
deceleration would seem strongly dependent on the controller being
able to maintain sync, rather than solely on the motor's available
torque as is the case with a stepper.

Dunno what you mean by this, either. Whatever the motor is capable of
doing, the controller and software will actuate.

So, how much can I expect to pay for a 400 oz-in servo?

You can get a new 1/3 or 1/2 HP PMDC motor surplus, and retrofit an
encoder, for less than $100. Ready-made DC servos are more expensive.

http://www.truetex.com/servomod.htm
http://www.truetex.com/mcgpd34002.htm

"Richard J Kinch" wrote in message
.. .
Mike Young writes:

The discrete nature of the step pulses and
the encoder stream would seem to be inherently noisy, but is easily
averaged out in steady state motion.

Dunno what you mean by noise. There is no noise of any consequence.

Measurement noise, as opposed to EMI or audible noise. Not of consequence.


OTOH, acceleration and
deceleration would seem strongly dependent on the controller being
able to maintain sync, rather than solely on the motor's available
torque as is the case with a stepper.

Dunno what you mean by this, either. Whatever the motor is capable of
doing, the controller and software will actuate.

Again, probably of little consequence. I was thinking about the the PID
controller characteristics and settling time during accel and decel.

So, how much can I expect to pay for a 400 oz-in servo?

You can get a new 1/3 or 1/2 HP PMDC motor surplus, and retrofit an
encoder, for less than $100. Ready-made DC servos are more expensive.

http://www.truetex.com/servomod.htm
http://www.truetex.com/mcgpd34002.htm

(Thanks for the links. Interesting pages...)

And now we've come full circle. Why servos for hobbyist CNC? What benefits
do you gain for the added cost and complexity?

Mike Young writes:

Why servos for hobbyist CNC? What benefits
do you gain for the added cost and complexity?

Superior characteristics and performance, less cost.

Steppers made sense only when encoders and digital servo controllers were
expensive. Encoders and controllers are cheap now, so steppers are
obsolete, except for the smallest and lowest-torque applications.

"Richard J Kinch" wrote in message
. ..
Mike Young writes:

Why servos for hobbyist CNC? What benefits
do you gain for the added cost and complexity?

Superior characteristics and performance, less cost.

Hence the question. "Hobby" range PMDC motors cost about twice as much as
"hobby" steppers of roughly equal speed and torque. Simpler drives can be
used with steppers, and the encoder is optional. So what are the superior
characteristics? The only that comes to mind is static current draw for
steppers; they draw power just sitting still. That seems to be of very
little consequence.


Steppers made sense only when encoders and digital servo controllers were
expensive. Encoders and controllers are cheap now, so steppers are
obsolete, except for the smallest and lowest-torque applications.

So what are the superior
characteristics?

Here's just one......One of the advantages of a CNC setup is that one
can walk away and leave it to finish what was started. You've probably
seem a drill bit get jammed in the workpiece while manually drilling.
The same thing can happen while CNC drilling or milling. If this
happens with a stepper setup chances are the X and Y axes will do their
best to not only destroy the part but destroy themselves and Z axis
components as well, behind your back, since there is no switch or
feedback to tell the software or the other axes that something went
wrong.

wrote in message
oups.com...
So what are the superior
characteristics?

Here's just one......One of the advantages of a CNC setup is that one
can walk away and leave it to finish what was started. You've probably
seem a drill bit get jammed in the workpiece while manually drilling.
The same thing can happen while CNC drilling or milling. If this
happens with a stepper setup chances are the X and Y axes will do their
best to not only destroy the part but destroy themselves and Z axis
components as well, behind your back, since there is no switch or
feedback to tell the software or the other axes that something went
wrong.

Closed loop is undoutedly a Good Thing (TM). It's not exclusive to servos,
however. You only need an encoder, and software or controller that reads it.

Why else might I consider discarding the steppers for servos? Motor and
drive cost is about $300 per axis; I already have the encoders.

Mike Young writes:

Hence the question. "Hobby" range PMDC motors cost about twice as
much as "hobby" steppers of roughly equal speed and torque. Simpler
drives can be used with steppers, and the encoder is optional. So what
are the superior characteristics? The only that comes to mind is
static current draw for steppers; they draw power just sitting still.
That seems to be of very little consequence.

This kind of cost comparison isn't appropriate. Steppers perform
dismally. They are only "cheap" when they aren't moving, but the
purpose of the system is to move. Steppers are rated with holding
torque, but nothing like that torque is available in motion. The
"simpler" drive only works up to a few hundred rpms. To get higher
speeds you need a sophisticated drive that costs as much or more than a
servo drive. And the torque still dribbles off.

Torque is misleading you. Motion components should be compared in terms
of power, that is, torque times speed. Steppers start losing their
torque at rather low speeds, and that makes them much more expensive for
the equivalent performance of servos. The top speed is a fraction of a
servo.

Steppers are also absurdly inefficient in converting electrical power to
mechanical work. This means you need a power supply many times larger
than what a servo would require, and this adds further up-front expense.

Servos have guts. You can run them at multiples of their rated torque
for short periods. This is exactly what you need to move a mechanism
with stiction, which is to say any type of machine tool. Or to avoid
stalling on a cut with a difficult momentary load. With steppers,
moving and cutting is always stuck underneath the reciprocal of the peak
force requirements, instead of the average.

Steppers are appropriate for slow, tiny, low-torque applications, like
printer mechanisms and disk drive heads. They have no place in machine
tools any more, other than as a hobbyist toy to get the cheapest
possible motion no matter how dismal the speed.

It's not exclusive to servos,
however.

Actually, closed loop in our context is exclusive to servos. It is not
exclusive to DC servos. So far we have been using servo to mean DC
servo and stepper to mean open loop stepper. A closed loop stepper
setup (using encoder etc.) is also a servo. We seldom have to make the
distinction since stepper based servos are very rare.

You only need an encoder, and software or controller that reads it.

You just need an encoder and a closed loop stepper drive (encoder
feedback). The most popular CNC software doesn't have a clue as to what
is receiving the steps it ouputs since there is no feedback from the
drive to the software. All it does is output a finite number of steps
for a particular move. It will do that even if nothing is connected to
the computer.

Why else might I consider discarding the steppers for servos?

Don't discard anything if you're satisfied with it.

"Richard J Kinch" wrote in message
. ..
Servos have guts. You can run them at multiples of their rated torque
for short periods. This is exactly what you need to move a mechanism
with stiction, which is to say any type of machine tool. Or to avoid
stalling on a cut with a difficult momentary load. With steppers,
moving and cutting is always stuck underneath the reciprocal of the peak
force requirements, instead of the average.

So I studied up a little on servos.

A 200W AC servo will do what I want. If you're going to drop the dollars,
get what you want for the money, right? (List for a 200W Sanyo-Denki P5 and
PY2 drive is $675 + $1125.) 4500 RPM, 100 oz.in. continuous, 280 oz.in at
current limit.

What does 100 oz.in. continuous duty torque buy you? Operating at 3000 RPM
max with a 90% efficient ballscrew, 5mm lead (590 ipm):

* Continuous duty cutting against 150 lb cutter force. (I've been using a
high limit of 100 lbs for years, but don't recall where I pulled that from.
Probably out of my arse, and at any rate, the servo won't be the limiting
factor.)

* From standstill, accelerate 200 lbs of table, fixtures, and work to 590
ipm in 80 ms. Roughly 1.5 G's, accomplished in the first 2" of travel.

That's pretty jaw dropping performance, and low maintenance on the brushless
motors. The only thing to not like is the price, about $2k per axis. I
didn't think it would be possible, but that makes the [PMDC + encoder +
Gecko 320] package look cheap. Insignificant even.

In comparison, a lowly $40 double stack NEMA 23 stepper can hold the same
100 oz in up to 450 rpm, a little less than 90 ipm. That's absolutely
lackluster next to the servo, but a pretty good match for small machine
milling. (In good hardened aluminum, 90 ipm is a 1/2" HSS 4 flute end mill
feeding .015 per rev at 1500 rpm, for 180 sfm. Not a bad bad match.) Control
is simple and direct: position, directly, rather than PIV this or PID that.
So simple even a layperson (me) can understand it. Not that the simple part
matters... PCs do more and more of my thinking each day.

(List for a 200W Sanyo-Denki P5 and
PY2 drive is $675 + $1125.)

You can get brand new 1kW Sanyo-Denki P5 AC servo motors for $39.95 in
abundance on the surplus market. Yes, that's forty dollars. Now go
price the drives and the 3 phase power supplies.

* From standstill, accelerate 200 lbs of table, fixtures, and work to 590
ipm in 80 ms. Roughly 1.5 G's, accomplished in the first 2" of travel.

In theory perhaps. Motor torque is one thing, the drives ability to
accelerate the motor to the required speed in 80ms is another. Why do
you think CNC software has start velocity and acceleration options?

wrote in message
oups.com...
(List for a 200W Sanyo-Denki P5 and
PY2 drive is $675 + $1125.)

You can get brand new 1kW Sanyo-Denki P5 AC servo motors for $39.95 in
abundance on the surplus market. Yes, that's forty dollars. Now go
price the drives and the 3 phase power supplies.

Wow. It's amazing what pops up by simply adding "surplus" to the Google
search. In general, are these units OK? $40 is almost beneath notice, but
I'm not fond of just tossing it away. Also, are there reasons to buy a
properly sized, smaller motor rather than the multi kW motors, now that $$
isn't really an issue?

So, what about the drives and PS? I don't know jack about either; was just
pretty much just glassy eyed when I saw the mfr list price. What do I need
to start?

(Good Lord. 3 hp DC spindle motors for less than $100? Thanks for the nudge;
a lightbulb just came on.)

In general, are these units OK?

Brand new, never been used AFAIK.

So, what about the drives and PS? I don't know jack about either

Me neither and never did any further research. I fell back on the adage
"If it's too good to be true then it probably is". In other words, if
the motors are costing less than an encoder and still work as designed
then it's something to do with the drives and PS why nobody is
interested. Hopefully, someone with any AC servo experience can chime
in here.

In article 1127924078.465626.307350
@g44g2000cwa.googlegroups.com, says...
In general, are these units OK?

Brand new, never been used AFAIK.

So, what about the drives and PS? I don't know jack about either

Me neither and never did any further research. I fell back on the adage
"If it's too good to be true then it probably is". In other words, if
the motors are costing less than an encoder and still work as designed
then it's something to do with the drives and PS why nobody is
interested. Hopefully, someone with any AC servo experience can chime
in here.


Though Sanyo Denki has very little data online, they look
just like the motors that come with some Aromat/Panasonic
servo systems, a closer comparison of the specs would
likely tell the story...

http://www.motiononline.com/home.htm?overview_py.htm&2

http://www.ohlheiser.com/products/aromat/panasonic/low%
20inertia/Minas%20A%20MSMA.pdf

I've used several of the Panasonic drive/motor combinations
as stepper replacements. The motors appear to be standard
brushless servos with a standard encoder. I expect a
brushless amp from AMC or Copley would drive the motors
with no problems (beyond the hair pulling that often goes
along with commutating a new brushless motor and drive
combination.)

http://www.copleycontrols.com/

http://www.a-m-c.com/

I do have the documentation and a couple of the Panasonic
motors on the shelf - if anyone is really interested I can
take a closer look. That is an excellent price from Surplus
Center.

Ned Simmons