Devlin:
As an electronics engineer, I often laugh at companies who claim that the strength of an industry-standard part is
somehow proprietary. It might be secret to THEM, and THEY might refuse to provide you that information, but it
is not at all illegal for you to ascertain, using standard and normal things, what the characteristics of the system
are. And since NWA definitely does NOT have the patent on stepper motors, your investigation and characterization
of that motor does not legally constitute "reverse-engineering". In addition, since your intention is not to compete
by selling systems that include these proprietarily-specified motors, there's zero harm, and therefore zero case.
What I'm saying is, you're free to know they're NEMA-23s somewhere in the 280-310 oz-in range, but the specific
size doesn't matter. You have to get driver modules of sufficient strength, and you have to adjust those so they
work properly with the motor IN THE SYSTEM. That means any current numbers are just a starting point, and some
small amount of tuning will be needed. (Have no fear, I have a robot-torture test for you). A single-axis 4.5A stepper
driver board from eBay, needing a 24-32/4.8A VDC supply (Amps is per-axis) will set you back about $5.
And, it's really easy to tell the motors. The two digits are the numeral and fraction of the side dimension of the face place.
So a NEMA 17 has a 1.7" square mounting face, a NEMA 23 has 2.3", and the NEMA 34s have 3.4" face plates.
The torque is variable across a range in each motor size, and motor length tends to increase as torque increases, though
there are overlaps between vendors. So, a stronger motor from mfg A might be the same size as a weaker one from
mfg B, even if mfg B makes that torque motor in a different size. The trick is when you buy them, know the mfg and
specific model number.
The thing about steppers is that they have a max torque and holding torque that at some point is independent of pushing
more current through. I'm not an expert on the underlying phenomena, but it seems like the metal rotor and stator become
magnetically saturated. That means that if your stepper driver board has a variable current-limit control (a trim-pot, nearly
all of them have them these days that I've seen), you don't REALLY care what torque you have. You have the motors you have,
and those motors have the torque they do. What YOU need to find out is how much current to drive them with. Enough to not
lose steps, but not so much as to start a fire from heat.
And for me, that testing is of an almost impossible-to-imagine, truly PERSONALLY satisfying nature. It really is. Let's assume
you're starting with a known motor + mechanicals and a variable-output, new stepper-driver board. For our size motors, a 4.5A
board (~$5 on ebay, per axis) is more than enough. Turn the current all the way UP on the driver board. Create a GCODE program
that cuts in AIR, in the same circle, over and over and over. Not a square, too much acceleration. Then let it run. For hours.
Whatever you do, do NOT put a fan on the driver board. But, DON'T touch it...a non-contact thermometer is best.
After a nice little piece, the steppers should start to heat up Watch how FAST they are heating up. If they're moving up at a
strong clip, maybe the current is too high. But, we don't really know yet. You just kinda need a feel for how quickly it's heating
up. Let that run around in circles until one of these things happens:
1) everything overheats and fails: turn down the current a little and try again. No, you don't need to let it cool down, unless
it's somehow still not working after the adjustment;
2) The motor reaches 125C (257F)...see, I told you not touch it!...that's good, but only a little bit too hot. turn the current
down just a smidge and leave it there. The plastic, HDPE is good until 200C+, so the motor isn't going to harm it. Inside the
motor is magnet wire in 260C (500F) epoxy-enamel insulation, held in place on a stationary aluminum structure, protected in
nylon (250C melting) coil-holders. 125C (257F) is not going to hurt anything.
3) The stepper driver board somehow reaches over 85C: stop immediately and lower the current a significant amount. This
would indicate that the driver is not strong enough to drive the motor. If it's a real 4.5A driver, it'll be a module about 2x3.2"
with a good, beefy heatsink. If it's a lot smaller than something that size, then be cautious.
4) After about 15 minutes, the stepper motor and drivers settle in, reasonably above 60C, and below 100C. In that case, you've
got it dialed in and it should be quite reliable.
5) Things lose steps, the cutting starts to drift, and the movement doesn't seem "crisp": you know what to do
Give it some
more juice.
Add a fan to those driver modules, and it should be fully reliable on all cuts, even hours-long ones. The motors should stay below
100C, and the drivers should stay not much more than the 70C. With a decent electronics-fan, the heatsink should be in the 50C
and under range. Of course, be careful that the fan doesn't pack your electronics with dust...or in my case, fine aluminum chips.
=================================
I've been doing some engineering with steppers these days, and I can tell you these things:
Our motors are NEMA 23, in the 285-300 Oz-in class. How can I tell? Because the next step up in torque, in the 430 oz-in
range, requires drivers that put out 4.0-4.5A. Given the size of chip and heatsink provided in current drivers that CAN
handle that current, and given the size and amount of heatsinking that our modules actually DO have, there is no way
they can push 4.5A, and therefore would be horribly under-driving the larger size. Thus, they must be somewhere in the
280 oz-in range for a NEMA 23 stepper. Those motors tend to take 2.8-3.5A, which is just the right amount for the chips
and heatsink size we have in our controllers. Don't just go on coil resistance, however, since they use different size wire
AND different number of wire loops ("turns" in the vernacular) in the different motors.
I'm too am currently investigating replacing my controller, but in my case with a LinuxCNC unit. Total controller cost without
the old PC is on the order of a couple hundred bucks, with the main parts coming in ~$60-75. Mine's a little strange because
I intend to actually embed the Shark controller board, minus the driver modules, into the new controller. I have a lot of already-
verified toolpaths that I don't really want to re-engineer.
The shark controller won't work with Mach....if you want to use Mach, expect to put in a new controller. If you want to
run Mach on Windows 7, you can use Mach 3 and a PCI-X (or PCI) add-in parallel port. However...and this is important:
in you want to run Windows 8 or Window 10 with Mach, you can only run Mach 4, and you can only use the Ethernet
smooth-stepper interface boards. Why? Because Windows 8 and Windows 10 changed how application software (like Mach)
accesses hardware. And direct parallel-port access is now forbidden.
LinuxCNC doesn't have those Windows restriction, however it does run on Linux, which can be confusing for some.
That said, both the OS and CNC controller software are FREE, and the LinuxCNC community is updating and fixing the
software FAR more frequently than do the Mach people.
Hope that helps. You can contact me via private-message, or email thom at thomr dot com if you're interested in more
details on the controller(s) I'm putting together.
Regards,
Thom Randolph