Mechanical Vs Electro Magnetic Advantage
Alright, I'm trying to wrap my head around this concept for basic machine
High TPI results in better mechanical force.
Low stepper speed results in better electro magnetic force.
If you wanted to build a high speed machine which trade off will result in
the better net results.
For the sake of this I am assuming accuracy and resolution will be within
desired parameters with either build.
Desired results are good strength for roughing out parts, good speed for
fast pocketing, the ability to overcome any minor drag or bind points in the
machine, and the ability to move a heavy carriage.
Obviously screw and nut types and quality will have a major affect as will
the quality of the machine build itself. I'm looking for a general major
difference here in the one aspect of the build.
While they are a type of screw and nut...blink blink..they impliment the
differences rather dramatically.
Ballscrews are tough, strong, able to take loads that would bust a nut
and screw, far more accurate and last far longer.
With ballscrews..one simply needs to use servos/steppers that are fast
Or are we talking on different pages here?
OK, the basic problem is that stepper torque falls off rapidly with
increasing speed. At some point, the available torque is less than the
drag (friction plus machining forces) and the stepper starts to lose
steps. To REALLY know what is going on, you have to get manufacturer's
data charts on the torque vs. speed for the specific motor, drive and
power supply voltage. This is the reason why so many setups use a
stepper with a 1:1 drive ratio to the leadscrew. If it weren't for this
loss of torque, you could use lots of reduction as is often used with
servo motors, where the torque is pretty much flat across the speed range.
Once you have a chart of torque vs. speed, you can then figure linear
force that can be delivered at a specific IPM feedrate. You can then
compare what happens at 1:1 or at 2:1 reduction. Probably, in most
cases, you will find that the belt reduction doesn't buy you anything
except at standstill. When you near that motor speed where the supply
voltage limits the drive, then the reduction will actually hurt you, you
will get more torque to the screw without reduction.
Trade off what vs. what? It _appears_ that you've settled on stepper
motors, and you're looking at trading off step rate vs. thread pitch (or
In my experience, from a pair of similarly sized motors you can get a
lot more useful power out of a DC motor than you can out of a stepper.
Granted, there are clever circuits out there to increase the power out
of a stepper -- but by the time you've reached the end of that road
you're still not generating as much reliable shaft power as you can with
a DC motor, you've got a circuit that's every bit as complicated as a DC
motor with feedback, and one that's possibly even more obnoxiously
obscure in its workings.
So _I_ think that, given the available motors and drivers out there
today, that your _first_ good tradeoff is to use motors with good
encoders or tachometers on them in closed loop with servo control. If
you just have to follow the stepper model, use controllers that make a
DC motor pretend to be a stepper.
You're right of course, but Bob already has his steppers.
There are engineering calculators out there to tell you how much
torque you need. And your stepper spec sheet will give torque vs. RPM.
A bit of study will get you in the right ballpark for gear ratio to
give the torque you need at low speed. We just don't have the
information to tell you what ratio to go with. I can say err on the
side of extra torque and sacrifice top end speed. To little torque
and you screw up parts with missed steps and it only takes one little
section of your program. Top end speed is just not that important for
My fist CNC was a stepper on a knee mill. After I screwed up a couple
parts, I went from 1:1 to the ballscrew to 4:1. And that was with 2000
oz in steppers.
Basically with a stepper, the stepper motor is what's keeping track of
where it is -- the motor is physically counting steps. If you try to
drive it too fast, the average torque out of the motor falls to zero,
period. That doesn't just mean that the motor won't push any more -- it
also means that motor goes limp, and will allow itself to be backdriven.
So unless you've got feedback external to the motors your machine
position will be unknown until it's homed again. Even if you _do_ have
feedback external to the motors, and a control loop that does something
with it, what that control loop will see is a HUGE disturbance, which it
will want to respond to by asking the steppers to go REALLY FAST --
which will just keep them unlocked.
In my formative years as a control system engineer I had a very bad
experience with steppers. They had let a mechanical guy with no
responsibility for driving the motors select them, with an electrical
guy who didn't understand motors designing the circuits, and a software
guy who barely had the mechanical ability to zip up his pants* writing
the code. I got tossed into a project as a troubleshooter** after we
had product failing in the field, and all I could do was offer
management the choice between being reliable without meeting spec,
meeting spec 70% of the time with spectacular failures the other 30%, or
completely redesigning the mechanism.
So I tend to be down on steppers. But if you have time, space and power
to waste, they're a great way to turn a shaft in a predictable way.
* Not to denigrate the guy on the whole -- he was an absolute wizard
when it came to processing video, and absolutely made up for any lack of
mechanical smarts with his attitude and aptitude in other areas -- he
just couldn't screw a bolt into a threaded hole to save his life, and
should _not_ have been asked to drive a motor!
** The real trouble was that they demoted a vice president of _programs_
for crying out loud into a project management position -- so he's pissed
off, disconnected, and not an engineer. This in a company that doesn't
understand the meaning of systems engineering, which means that the
systems engineering is either a consensus amongst the project members,
or entirely up to the program manager. His response to systems
engineering questions was "you've been hired to build a system -- go
build a system".
Maybe -- too high a TPI (too fine a pitch) means that the shear
strength of the threads is lowered.
Toss the steppers and use servos with encoders instead. They
can put out lots of torque at high speeds.
Toss the (presumably Acme) threads and replace with ball screws
and ball nuts -- very low torque needed to move against cutting forces
compared to Acme screws. And -- the ball nuts and screws are typically
not that fine, either.
Essentially, this is what serious CNC machine tools use.
If you are already considering steppers -- servos and the rest
of what is in a good CNC machine are the better choice.
Your argument is wrong. The premise you have assumed is the use of stepper
motors in a high load solution. Stepper motors, by
their nature fight themselves. Output torque is the differential between
slightly out of phase drive fields and is quite low in
respect to the power applied. Under load they slip. A far better solution is the
use of permanent magnet DC motors driven with PWM
along with velocity and positional feedback. This solution is powerful,
efficient, fast and accurate. It has been the solution of
choice for computer disk drives for 30+ years. Furthermore, the parts required
are available and inexpensive. Just as Gunner
stated, you are locked into using ball screws, because of backlash. Ballscrews
are low TPI. The more gears you use, the more
backlash you build in. So, a high torque, direct drive solution is the only way
I have been looking at servos. inexpensive is in the eye of the beholder.
Controller's, power supply and motors are about twice the price of
comparable in/oz steppers. Given that in many cases the power supply is
interchangeable. Most of that difference is in the servos. I'm not saying
they are not better. I always figured a positive feedback encoder made
perfect sense anyway.
I already have a fair number of steppers accumulate from various sources and
three stepper controllers.
OK, here's where you run into trouble. Since you are already looking at
the difference between steppers and servos, I'll get into that shortly.
I'm not sure what "bind points" are, but mechanical binding will almost
certainly cause steppers to skip steps with no warning to the computer.
This not only can cause ruined parts, it can cause the machine to start
cutting into itself!
Stepper motors are good up to about 200 W of mechanical power, ie. 1/3
Hp. Above that, you really need servos. Larger steppers just don't
deliver much more power, but lots of torque at lower speeds.
All steppers suffer from a loss of torque at higher speeds. For
instance, some particular stepper may generate 200 Oz-In of holding
torque, and that will be the given rating of the motor. But, if you can
get the curves on the motor, you will find that it only gives 100 Oz-In
at even low speeds, like 600 RPM. At 2400 RPM, the delivered torque
will be almost zero, just touching the shaft with your finger will cause
it to stall. (These numbers are just typical for some size 23 motors,
some are better, some worse, and apply ONLY to a specific combination of
motor and drive, at some specific supply voltage.)
Another problem with steppers is resonance. Since the motor moves in
STEPS, it produces a vibration every step. At some speed, this
vibration matches the mechanical resonance of the motor, leadscrew and
other parts. These vibrations grow until the motor loses synch, usually
accompanied by loud rattling sounds. I would not use non-microstepping
drives in ANYTHING today except toys. Comparing the difference between
a full-step drive and a good microstepping drive like a Gecko is an
astonishing night and day difference!
Another problem with steppers is heating. Modern drives reduce motor
current at idle to reduce motor heating, but there is also SELF-heating.
The moving magnets in the motor cause iron losses in the motor's
laminations. If you spin a typical stepper motor at 2000 RPM, with no
electrical connection, it will get quite hot. In fact, it will burn out
in less than 30 minutes! The stepper motor makers will NEVER tell you
about this nightmare! So, if you plan a machine that will run motors at
high speed for a long time, it will self-destruct.
Servo motors generally don't suffer from these problems. Torque is
roughly constant up to the speed where the self-generated voltage of the
motor nearly equals the power supply voltage. Self-heating is usually
quite small, and is proportional mostly to motor LOAD, not speed.
Resonance is non-existent in motors designed for servo use, as the
windings are set up to minimize velocity and torque ripple.
A modern brush motor is just a little more complicated than a stepper,
and brushless motor is no more complicated. Due to the efficient heat
removal of the brushless motor, a smaller, and therefore cheaper motor
can often be used and still provide far better performance.
Keling supplies some awesome size 23 brushless motors for $52 in single
quantity. You still have to add an encoder, but the CUI encoder from
Digi-Key is now down to $28. I make a drive for these motors for $150,
and a 4-axis controller for $250. So, a 3-axis system can be built for
$940, assuming you already have the power supply. Now, this is NOT
picking parts out of your junkbox, but it will perform a LOT better.
You haven't defined "high speed" or "heavy carriage", so I have no idea
what scale of things you are talking about. High speed machining means
75 Hp spindles running at 80,000 RPM and removing 640 Cu In of aluminum,
if you ask Boeing. it might mean 1 Hp at 10,000 RPM and 1 Cu In per
minute to a guy in his garage. Giddings and Lewis might think a 25,000
Lb gantry is pretty heavy, you might think 100 Lbs is heavy.
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.