Why a stepper motor should not be too big . . . .

I keep picking up bits and pieces of info on why bigger steppers are not necessarily better...

Today Mariss said on his forum:

"We have one of those infamous 1,200 in-oz motors. The problem with the
motor is its poor electrical angle to mechanical angle linearity.

1) Advancing 90 electrical degrees results in 1.8 degrees of motor
motion (one full step). A 10-microstep drive advances 9 electrical
degrees per step and ideally the motor should move 0.18 degrees
mechanical for each step. A graph would show a linear relationship
between electrical and mechanical angles (a linear slope thru the
origin).

Motors with poor linearity have a pronounced S-shaped curve (initial
steep slope rounding to a shallow slope, rounding to a steep slope
again. This means the microsteps bunch at the half-step location and
spread at the full-step locations. This results in motor deceleration
where the microsteps bunch and acceleration where they spread. Since
this acel/decel occurs over the span of one full step, the motor is
pumped into resonance when the acel/decel frequency (full-step rate)
matches the motor's mechanical resonant frequency.

A motor with good linearity will have evenly spaced microsteps and
therefore has no acel/decel component to its motion. It will exhibit
no low-speed resonance.

Our 'figure of merit' for motor linearity is the ratio of the biggest
microstep to the smallest microstep. Good motors have a ratio of 1.1:1
or less. This motor has a rather miserable 1.9:1 ratio and it does
exhibit pronounced low-speed resonance.

2) The "no free lunch" thing comes into play here as well. This
particular motor has a corner speed of only 210 RPM at 80VDC. This
means torque rolls-off rapidly above 210 RPM. Quick calculations show
a power output of 186 Watts mechanical.

You can get more power (about 250W) from a lower torque NEMA-34 motor
and have excellent linearity as well."

Art Fenerty said on his forum:

"Don't be too tempted to get Nema42's with 1200oz inch power, as steppers get larger, the detent torque (that clicking you feel as you turn a stepper) increases, and it can be a devil to tune them or get great acceleration. I recommend (just me..) that you don't go higher than 800oz in for a stepper"

Another exchange of comments last night:

Art:
Yeah, 1200 Oz in motors can be a problem on any system, very high
detent torque. Try .053, it has a ramped backlash correction now, it
may help alot.. hard to say though, Im still tweaking the fine
details.. Let me know how it does.. ( Personally, if one needs more
than 800 oz in or so, I always recommend servos, steppers over 1000
often give bad performance is many situations, on the up side they
have very great power, down side though is that they have very great
power..

Mariss:
Art, I agree. NEMA-34 motors over 800 in-oz holding torque are
designed for high torque at the expense of motor linearity. Motor
linearity is paramount for smooth motion at low speeds.

Secondarily, This high holding torque comes at the expense of speed.
It is part of the "there is no free lunch" principle. Gobs of holding
torque means you have a motor that cannot get out of its own way as
speed increases.

Given the examples I've tested, I think of the over 800 in-oz NEMA-34s
as vanity motors. They give the motor mfgs bragging rights to say "Our
NEMA-34 motors go up to 2,100 in-oz !" Some gullible users buy them
and find what they have is a real dog (or boat anchor) on their hands.

The hardest concepts to pound into people's heads is it's power that
gets things done. Power is proportional to V / SQRT L on a step motor.
Detent torque is always present; high detent torque motors consume
more power to turn themselves than smaller, low detent torque motors
and that power comes at the expense of what's left to apply to the
load. Never mind the non-linearity (low speed vibration) of these
unfortunate monsters.

Use a sub-800 in-oz NEMA-34 motor. A good one will be smooth at all
speeds. Spin it fast enough to develop real power (torque times RPM).
Gear it properly to the load instead of taking the brute-force 1:1
approach. You will out-perform any muscle-bound 2,000 in-oz beast of a
motor.

Art:
Yes, Id rather have a small fast motor on a reduction gearing anyday, than a big honker motor 1:1 connected. The first one IS
impressive, the latter only looks impressive.. Its why I think the faster speed drivers over the next few years may change the way
people design drive systems, small and fast can be pretty impressive when fast is determined to be around 300Khz to the motor..
lots of room for gear reduction there.. as well as much higher resolutions ...

But thats just me. ... truth be known, Ill use anything that works well...

[Gerald D]

Posted by Mariss today:

Motor power output is proportional to V / SQRT L so inductance and
supply voltage determines power output in a perfect world. In the real
world step motors have detent torque. It is present while the motor is
turning and it absorbs power that otherwise would be available for the
load. This absorbed power is proportional to speed. Large frame size
motors have large detent torques and they absorb large amounts of power.

If you compare the power output of a 14mH NEMA 23 step motor to a 14mH
NEMA 42 motor, you will find the NEMA 23 motor will have more torque
at high speeds. The NEMA 23 will continue to spin way past the speed
at which the NEMA 42 stalled.

[Gerald D]

Before Mariss posted, his son Marcus said the following:

Yes, the motor will run, but your high speed performance will be
miserable compared to a modern square NEMA 34 motor. This is because
motor power output is proportional to power supply voltage / SQRT
motor inductance. A modern NEMA 34 rated at 6A will have around 2mH of
inductance, allowing it to deliver better high speed performance.

This was in response to a question "is it possible to drive nema 43 motor with Geckodrive?"

[Gerald D]

From this forum today:
http://groups.yahoo.com/group/mach1m...message/101024:

Just thought I would share this problem I have been troubleshooting
for 2 weeks incase it may help someone else later on down the road.

I discovered as I ran some test on my just completed router bed that
I was missing steps on the Y direction. At first I felt it was
mechanical and tried to eliminate everything I could. I was running
my test cutting foam at 60"/min feed. The X & Z were rock solid but
the Y kept missing steps and wandering all over. The Y can run over
400"/min without stalling so I eliminated the mechanical idea that
it was stalling out at 60"/min in foam. Then I started looking at
noise, printer port voltage, breakout boards, timing tests etc.
everything looked ok.

I swapped computers, drivers, just about everything.

Then I took and replaced the stepper with a motor that was less that
1/2 the size of the 1200 oz one I was using. Bingo that was the
culprit, brand new 1200 oz stepper was randomly missing steps. I
am doing some more checking to make sure it wasn't a problem with
the stepper wiring size or connectors because I use a larger wire
and new connector when I put the smaller motor on.

Hope this might help others in their searches for the elusive
missing step cuplrits.

Arnie

[Gerald D]

Here is someone actually using big motors direct driving a MechMate:
http://www.mechmate.com/forums/showt...1&postcount=31

My first impression from that video is that the speed and control is good. The cut quality is debateable. But, generally speaking, the results are far better than what I was expecting after reading the comments from Mariss, Marcus and Art . . . . . .

[maoxinai4]

Originally Posted by Gerald_D

I keep picking up bits and pieces of info on why bigger steppers are not necessarily better...

Today Mariss said on his forum:

"We have one of those infamous 1,200 in-oz motors. The problem with the
motor is its poor electrical angle to mechanical angle linearity.

1) Advancing 90 electrical degrees results in 1.8 degrees of motor
motion (one full step). A 10-microstep drive advances 9 electrical
degrees per step and ideally the motor should move 0.18 degrees
mechanical for each step. A graph would show a linear relationship
between electrical and mechanical angles (a linear slope thru the
origin).

Motors with poor linearity have a pronounced S-shaped curve (initial
steep slope rounding to a shallow slope, rounding to a steep slope
again. This means the microsteps bunch at the half-step location and
spread at the full-step locations. This results in motor deceleration
where the microsteps bunch and acceleration where they spread. Since
this acel/decel occurs over the span of one full step, the motor is
pumped into resonance when the acel/decel frequency (full-step rate)
matches the motor's mechanical resonant frequency.

A motor with good linearity will have evenly spaced microsteps and
therefore has no acel/decel component to its motion. It will exhibit
no low-speed resonance.

Our 'figure of merit' for motor linearity is the ratio of the biggest
microstep to the smallest microstep. Good motors have a ratio of 1.1:1
or less. This motor has a rather miserable 1.9:1 ratio and it does
exhibit pronounced low-speed resonance.

2) The "no free lunch" thing comes into play here as well. This
particular motor has a corner speed of only 210 RPM at 80VDC. This
means torque rolls-off rapidly above 210 RPM. Quick calculations show
a power output of 186 Watts mechanical.

You can get more power (about 250W) from a lower torque NEMA-34 motor
and have excellent linearity as well."

Art Fenerty said on his forum:

"Don't be too tempted to get Nema42's with 1200oz inch power, as steppers get larger, the detent torque (that clicking you feel as you turn a stepper) increases, and it can be a devil to tune them or get great acceleration. I recommend (just me..) that you don't go higher than 800oz in for a stepper"

I think small stepper motor is wildly used in more application.