[mrehmus]

According to Micromo, maybe not. The torque falls off rapidly as one increases the number of microsteps.

Stepper Motor Tutorials & Whitepapers | MICROMO

for more details

[awerby]

In practice, any decrease in torque is more than offset by the increased smoothness of motion that microstepping affords. That's why there are very few modern control systems that use full steps; all of them I know of go with at least half-stepping, while 8x or 10x is more common.

[datac]

Microstepping "SAFE" ?

Well personally, I'd microstep thru a cow pasture, especially if it is in the dark,... perhaps I'd opt to microstep on black ice....

For me, full stepping is best left for good, clear trustworthy well traveled walkways.

We all should be careful when stepping so we do not end up on our heads !

(sorry, couldn't resist)

[Eldon_Joh]

I think someone else stumbled across that information and inaccurately interpreted it a year or two ago.

"The real compromise is that as you increase the number of microsteps per full step the INCREMENTAL torque per microstep drops off drastically. Resolution increases but accuracy will actually suffer."

This is true, the incrimental torque per microstep decreases, but accuracy does not suffer, depending on how you look at it.

suppose I want a stepper to hold a position between 19 and 20 out of 256 microsteps, so i command it to hold 20 microsteps. it will be closer to the desired position then if I only had 16 microsteps, and I commanded the motor to arrive at position 1 out of 16 microsteps.

Stepper motors are not magical devices which hold their position and have a defined accuracy, they are electrically driven magnetic SPRINGS!

Additionally: they are basically a hybrid switched reluctance and a synchronous motor (this is why they have to use alnico magnets that get partly wiped when you remove the rotor from the stator). when you apply 375 ounce-in of torque to a 425 ounce-in motor, the stepper motor's actual position will be just less than 2 steps behind the commanded position. suppose you give the shaft a light tap with a hammer (exceeding the 425 oz-in holding torque), the momentary torque will cause the shaft to rotate backwards 4 full steps. (in real life, it may lose holding torque completely and free spin)

In the case of my taig, which has 200 step motors and 20 tpi screws, while power is applied to the steppers and they are not moving, it is convenient that i can manually rotate the steppers in .001" resolution. they are 200 steps but when you exceed holding torque you jump backwards in 4 step increments.

"Simply stated, taking a microstep does not mean the motor will actually move!"

this is true, stiction must be overcome. it doesn't matter how many microsteps you lost, the stiction caused lost motion.

suppose you are interpolating a circle with a cnc machine driven with steppers, and you want the circle's accuracy as high as possible. suppose it only takes 10% of the motor's maximum torque to drive the machine under load.

well, presuming the relationship is linear (which it isn't) we can presume that the motor's true position will be one tenth of 2 steps behind the commanded position. so the hole will be .2 steps smaller in diameter (assuming no backlash and assuming the cutting forces are always resisting the motor's torque)

suppose the machine has some stiction. let us presume that stiction is 1% of the motor's peak torque.
so at the zero crossing when the motor changes direction, for some finite amount of time the motor will be .02 steps behind the commanded postion. so now we have a witness mark even if there is no backlash.

now suppose instead of 10 microsteps per step we drive the stepper at 256 microsteps.

how much less lost motion will there be?
in this example, practically no difference, but the surface finish will be much better. (supposing you can see the 10 microstep "steps" in the finished part.

[Torchhead]

All microstepping works on opposing currents in the coils. Whereas at full step it uses all of the current in one coil and pulls it to the nearest pole you have to use both coils pulling gin opposite direction ths make the rotor "hover" between a hard pole. Its useful at lower RPM on a stepper where you want smoothmess of motion and there tend to be resonance points from the hard landing on each pole. It does reduce the torque. How much more for higher step counts is debatable and yes the motion will be smoother (again at low RPM/ After you get past a sertain RPM the effect of microstepping starts to dimish because the inertia of the rotor and the load tends to overcome the "hover" forces. It is at higher RPM where you need higher torque (because ti already drops off due the inductive effect) . Most modern Stepper drivers have "step morphing " that moves the micro stepping to a low or even full step mode. It no longer needs it for smoothness but it does need all its torque applied in the direction of rotation. So the statement that microstepping improves resolution (or accuracy) is only partially true. It CAN improve resolution at lower RPM but not at the higher RPM / Anyone that uses the micro-stepping number to show resolution and/or accuracy is playing with numbers. The more steps between poles the more sine wave the current. the current and the smoother it runs. Unfortunately sine current only as .707 of the current of a square wave at the same voltage and duration. If you want true resolution increase you put in a transmission (belt reduction) . You trade speed for resolution (and torque) But there is a dilemma The more you increase the reduction ratio the faster you have to spin the motor to hit the same linear speeds and the LESS torque you have because of the increased RPM (Rats!) . You still have all of the resolution gain but the motor stalls sooner . So the trade off point is where you have a ratio between the belt reduction and the final drive element that essentially cancels each other out. Another way is to state that the ratio is right when one full revolution of the motor = 1 inch (25.4mm) of linear movement. If most of your cutting is done at lower feedrates then the ratio can be higher but for a plasma or mixed use table the "golden ratio" above works the best/

I would not run a stepper for a continuous motion machine without micro stepping. I would not run a stepper driver that did not have step morphing to be able to extend the torque curve out to the motor specs. If you look at any set of curves for stepper motors you will note that the torque drops off linear to about 1/2 the RPM then goes non-linear afterwards. The curves are always done at full step or they would drop off a lot sooner.

TOMcaudle
Home

[Eldon_Joh]

where did you read that microstepping works by opposing currents?

This is not the case at all, in fact full stepping gets the most torque when you have both coils fully "on" at the same time.. you are in fact, at this point, holding the stepper half way between two steps.. basically half way between the off state detent postion..

when you feed a stepper two sine waves 90 degrees out of phase, at one point you have one coil fully "on" and the other fully "off" this is why you only get 70% of "full stepping torque"

when both coils are at 71% current, or 45 degrees, you still have the same holding torque because the vector sum of the magnetic field remains 1.


the vector sum of both coils fully "on" is 1.41. but you also get twice the heat generation.

you may be able to run a micro stepping drive at 41% extra amps to compensate, i'll have to research that because i've not heard of anyone trying it. but you may demagnetize the magnets.

[109jb]

Torque is only decreased at the intermediate steps or microsteps. Say you have 1/16 microstepping. so for every 16 microsteps you are at a full motor step and the motor coils are energized as if full stepping at every 16th microstep. At those positions you have full torque. lets say the motor is at full step position at 0, 16, 32, 48 ... So then at 1-15, 17-31, 33-47, etc the torque will be reduced, but as said at every 16th step in this example you have full torque.

[Torchhead]

Since torque and power are not instant values they are integral over time. Power is expressed by Voltage X Current X duration. The amount of power in a square wave is the area under the curve. The amount of power in a sine wave is also the area under the curve. So its the RMS current (average current across one full wave form). Simple AC electronics says that the RMS voltage and RMS current of a sine wave is about 70% that of a square wave with the same peak voltage. An AC waveform with a peak of 160VAC will show 120VAC as an average (over time) voltage and the current will track the voltage (in a pure restrictive load) . In a microstepping motor currents in parts of the waveform vary and the higher the step count the lower each current step is , The end result is smoother motion but less torque. At some point there is not enough current to actually move the motor rotor let alone the an external load. Torque comes from power. It takes 4 times the power to double the torque. The answer is to either up the voltage or up the current limit so you get more power (torque) but either option comes with consequences. Upping the current increase heat by I^2 X R . All this would be pretty easy to understand if the motor was not huge old inductor with a moving rotor that has inertia.

Power is power (Watts is Watts) Power in always exceeds mechanical power out. Usable power produces work. Work is force over distance/time The opposing currents are vectors moving in some angle less than 90

[Eldon_Joh]

Torque is only decreased at the intermediate steps or microsteps. Say you have 1/16 microstepping. so for every 16 microsteps you are at a full motor step and the motor coils are energized as if full stepping at every 16th microstep. At those positions you have full torque. lets say the motor is at full step position at 0, 16, 32, 48 ... So then at 1-15, 17-31, 33-47, etc the torque will be reduced, but as said at every 16th step in this example you have full torque.

no, this is not the case. the only way to increase the torque is to supply current to both coils simultaniously, both at 100%

microstepping feeds 0-100% into both coils but the vector sum of the magnetic field is alway 100%. when both coils are on, each one carries only 70% of the current.
the vector sum of full stepping is 141%.

[Eldon_Joh]

In a microstepping motor currents in parts of the waveform vary and the higher the step count the lower each current step is , The end result is smoother motion but less torque. At some point there is not enough current to actually move the motor rotor let alone the an external load. Torque comes from power. It takes 4 times the power to double the torque. The answer is to either up the voltage or up the current limit so you get more power (torque) but either option comes with consequences. Upping the current increase heat by I^2 X R . All this would be pretty easy to understand if the motor was not huge old inductor with a moving rotor that has inertia.

Power is power (Watts is Watts) Power in always exceeds mechanical power out. Usable power produces work. Work is force over distance/time The opposing currents are vectors moving in some angle less than 90

the vector sum of the magnetic field generated inside a microstepping stepper is 100% of a single coil. this is why the microstepping torque is 70% of "full stepping" which i previously explained is when two coils are energized at all times.

as i already explained full stepping in which two coils are always fully on generates twice as much heat as does microstepping where each coil is feed a sine wave out of phase by 90 degrees. when one is fully on, the other is off. at 45 degrees both are 70.7%

steppers are magnetic springs.. at near rated output torque they are 2 full steps behind the commanded position, regardless of how many microsteps your driver is. so if you have a 10x microstepping drive, at full torque they are 20 microsteps behind commanded position.

since all cnc applications are dynamic, we don't really care about the incrimental torque between individual microsteps.

it might be true that a 256 microstep stepper motor won't move the machine until you command it to move +/-8 microsteps.
however, if you're interpolating a circle, you will more accurately trace that circle with a 256 micro step stepper motor than you will with a 10 microstep stepper motor, this is regardless of the amount of stiction.

[devo9er]

Quite a lot of info in this thread. Most of the discussion so far certainly a bit over my head and probably others too but interesting to read.

That article is misleading in a lot of ways and poorly biases the pros and cons in my opinion. It reminds me of click bait journalism. Something like, "How the vegetables you eat may actually be killing you." Or something along those lines..lol. It does mention the benfits of microstepping later in the read but I think it does a poor job giving examples of any real world situations. It doesnt discuss 1/4 or 1/8 stepping, it jumps right to picking on microstepping 1/256, which ive never personally seen equipment that operates past maybe 1/8 or 1/16. Im sure there is some specialized equipment out there that uses these insanley smaller microstep divisions but the reality is for 98% of the everyday shop mills etc, the accuracy would never be realized anyway. Machine rigidity, backlash, runout, tool flex and even just material temp fluctuation would become the limiting factor, right?

I guess, Ive just never had an application that didnt benefit from at least 1/2 or 1/4 stepping to smooth things out and get the resolution desired. Its almost like the article suggests we should all be full or half stepping and using some drive reduction to increase resolution. Thats way more complicated and potentially disasterous than just microstepping at 1/4 or 1/8 steps... Size your motors right and dont rely on the rated full step torque in your equations while you build something. Just my opinion.

[UA_Iron]

This exact topic has been discussed in the sticky - starting back in January to October of 2015

http://www.cnczone.com/forums/steppe...-torque-6.html

[Jon.N.CNC]

This debate is great, I've tried to understand it to the best of my abilities from this debate and past debates but no closer to understanding why my stall speed decreases if I go over 10x microstepping. Will remain a mystery I feel.