[sizala]

hi friends,
i have received my sieg sx3 10 days ago, i am very happy with the result that i got from the machine.
and like everyone else i want to convert it to cnc. i have made a cnc router in past with stepper motors. and thats why i want to go with the servos for the sx3.
i have found a wonderfull link about "Digital Calliper". here it is=
Voltages: Logic and Power

but using a digital calliper and a arduino we can read the position of the calliper on a computer. SO MY IDEA IS to us 3 calliper and connect it with PC and make a software (VB6) which will read the value from calliper and then gives step dir commands to the stepper driver. it works the same way like if we hocked up a encoder at back of an servo motor.
so as far as i think, by using calliper as DRO and with the help of an self made software i can use any stepper as servo.

so guyz, what do you think? am i missing something? will this concept work or not? and is there anyone on cnczone who have tried anything like this?

i am waiting for your suggestions guyz.

[SignTorch]

it's still a stepper

with some sort of feedback (which is really only useful to detect missed steps or maybe do backlash compensation)

it won't never function like a servo (because it's a stepper)

that is interesting, maybe for a rudimentary DRO, but I wouldn't want to rely on 3 cheap calipers for cnc

really the only way to get the benefits of servos is with servos

[Mariss Freimanis]

A motor is a motor is a motor. Because a motor is a stepper doesn't mean it can't be made a servomotor. For example, a brush DC servomotor is just another utility DC motor unless it is run closed-loop. A step motor can be run closed-loop. A step motor is actually a high pole-count permanent magnet synchronous motor (PMSM) and in theory is perfectly capable of being passivated by field oriented control (FOC).

The reason you don't see the world flooded with FOC step motor servodrives is because of the step motor's high pole-count.

There is little fundamental difference between a 6-pole, 3-phase PMSM motor (also known as a BLDC motor) and a step motor (50-pole, 2-phase motor). Both are perfectly amenable to to being controlled by the Clarke-Park transforms and their inverses better known as FOC.

The difference is the step motor has more than an 8 times higher pole-count (50 / 6 = 8.333) than a BLDC.

DSP (digital signal processors) type MCUs sample at 20kHz to run the Clarke-Park transforms / inverses and do a barely adequate job when the motor turns at 3,000 RPM. To do the same using a step motor, the DSP would have to sample at 167kHz. This is way beyond what a commercial DSP processor can do.

The Clarke-Park transform requires 4 sine and cosine multiplications plus some minor time doing 2 cross-additions. The inverse transforms require the same computational burden for a total of 8 multiplications and 4 2's compliment additions. Throw in 4 proportional-integral calculations and you have a very busy DSP processor. It is completely overwhelmed running a step motor at 3,000 RPM.

But what if there was a way to short-cut to bypass the burdensome 8 multiplications and 4 additions? What if there was a way to reduce them to no computational burden at all? What if it all this burden was placed in the analog domain? Then a step motor servodrive becomes very easy, in fact extraordinarily easy.:-)

The last significant hurdle is field-weakening (the Qd term in the Park transform). For a BLDC motor this can often be neglected because its effects become significant only beyond 3,000 RPM. Not so with a step motor; the effect arrives at 400 RPM and above.

BTW: A step motor is already a two-phase system. The Clarke transform and its inverse are unnecessary. For what it's worth, the Clarke transform is trivial. It's just a 3-phase to 2-phase conversion algorithm any could come up with. All the meat is in the Park transform.

Mariss

[SignTorch]

A motor is a motor is a motor. Because a motor is a stepper doesn't mean it can't be made a servomotor. For example, a brush DC servomotor is just another utility DC motor unless it is run closed-loop. A step motor can be run closed-loop. A step motor is actually a high pole-count permanent magnet synchronous motor (PMSM) and in theory is perfectly capable of being passivated by field oriented control (FOC).

The reason you don't see the world flooded with FOC step motor servodrives is because of the step motor's high pole-count.

There is little fundamental difference between a 6-pole, 3-phase PMSM motor (also known as a BLDC motor) and a step motor (50-pole, 2-phase motor). Both are perfectly amenable to to being controlled by the Clarke-Park transforms and their inverses better known as FOC.

The difference is the step motor has more than an 8 times higher pole-count (50 / 6 = 8.333) than a BLDC.

DSP (digital signal processors) type MCUs sample at 20kHz to run the Clarke-Park transforms / inverses and do a barely adequate job when the motor turns at 3,000 RPM. To do the same using a step motor, the DSP would have to sample at 167kHz. This is way beyond what a commercial DSP processor can do.

The Clarke-Park transform requires 4 sine and cosine multiplications plus some minor time doing 2 cross-additions. The inverse transforms require the same computational burden for a total of 8 multiplications and 4 2's compliment additions. Throw in 4 proportional-integral calculations and you have a very busy DSP processor. It is completely overwhelmed running a step motor at 3,000 RPM.

But what if there was a way to short-cut to bypass the burdensome 8 multiplications and 4 additions? What if there was a way to reduce them to no computational burden at all? What if it all this burden was placed in the analog domain? Then a step motor servodrive becomes very easy, in fact extraordinarily easy.:-)

The last significant hurdle is field-weakening (the Qd term in the Park transform). For a BLDC motor this can often be neglected because its effects become significant only beyond 3,000 RPM. Not so with a step motor; the effect arrives at 400 RPM and above.

BTW: A step motor is already a two-phase system. The Clarke transform and its inverse are unnecessary. For what it's worth, the Clarke transform is trivial. It's just a 3-phase to 2-phase conversion algorithm any could come up with. All the meat is in the Park transform.

Mariss

a motor may be a motor, but a stepper is not a servo, and anything in between is neither, it's not all that complicated....

it doesn't look to me like step-motor-servo-drive is even a real word, but for all I know FOC means full of crap (I'm joking there, I know you're not FOC, I'm not even doubting you, you know more than I do), but to me it doesn't seem practical to use cheap calipers for encoders or to use steppers for servos, your argument is over my head and doesn't change my mind, but maybe that's just what the OP needed to know, in which case, I digress....

[Mariss Freimanis]

I didn't mean to imply using digital caliper signals as a servo feedback is a viable option. Digital caliper signals update 3 to 5 times a second. This update rate is about 10,000 times to slow to be usable as a real-time servo feedback rate.

Mariss