[bunalmis]

I see some technical specs of step motor drivers.

This specs says;

Chopping frequency is 20 Khz.
Max puls rate is 100Khz

How is possible?

[NC Cams]

The output side is not necessarily directly linked to the input side.

This is probably what is happening:

A digital signal is first fed to a F-V converter. The output V from this is being fed to a comparator circuit. The comparator circuit tells the "chopper" mosfet driver the speed and direction to drive the motor. The chopper drive is independandly operating at 20 Khz.

A basic description of how this is done can be found via the application notes for a couple of IC's - LM2917 F-V converter and UC3637 motor controller. Although the 3637 is a 10khz chip, the operation is essentially the same for a 20khz chip.

Anyway, feed that 100Khz max input freq into the 2917 (or one suitable to operate at that freq) and out comes an analog control voltage that is proportional to the input freq.

Feed that voltage to the input comparator of the 3637 and you ultimately control speed and direction of the motor.

Pretty simple, huh???

[bunalmis]

Chopping frequency brings a limit for the step rate at the real time operation.

Controller must accumulate the step pulses if step rate bigger than chopping frequency. But this is mislead.

Are there another idea?

[fkaCarel]

Let's bring out the calculator. Let's chop and step. At 20Khz and 200 step/rev we get: 20000 / 200 = 100 revolutions per sec. * 60 = 6000 rotations per minute. Spindle 5mm pitch: 30000mm / min = 30 meters a minute! With halfstep this comes down to 15 meter a minute. So the 20 Khz chopper frequency covers the stepping range neatly. The 100khz will be the electrical limits. It looks all nice, ones and zero's, but the signals are sloped depending on circuit, load etc. etc., so there is a maximum speed where the zeros no longer become zeros and the ones no longer becomes ones. It's in between and as such no longer switching.

Carel

[Mariss Freimanis]

People, this is apples and oranges; one has nothing to do with the other.

Chopping frequency is the rate at which the winding current ramps up then down to regulate current. The frequency has no practical meaning except that 20kHz means it is inaudible.

Step frequency is the maximum step rate the drive will accept. Every received step pulse starts a new switching cycle; that it may not finish before the next pulse is received doesn't matter.

Step pulse frequency trumps switching frequency. Ignore the chop frequency.

Mariss

[fkaCarel]

Let's disagree for disagreement. Here we use apples and peers. The question posted was about conflicting frequencies without specifying a circuit. The question did'nt assume a reset for statechange. That's not general, it's a feature. If you scale down the frequencies, let's say a 100 fold, the problem is real. Dos wiedanja,

Carel

[Mariss Freimanis]

I think you meant 'pears' not 'peers'.:-)

Any good drive should reset the chopper monostable on step input; if not, aliasing (beat) frequencies would desynchronize the motor at all but the lowest of speeds.

Which frequencies would you scale down 100-fold? Certainly not the 20kHz. The 100kHz? Probably not. There are real (though rare) applications where you do need to have a stepper turning 30,000 RPM (100kHz, full-step, 1.8-degrees/step). Portescap disk motors spinning a mirror for laser scanning is one that comes to mind. The motor winding frequency is 25kHz then (100kHz / 4), clearly above the 20kHz switching frequency.

Lab dien,

Mariss

[Mariss Freimanis]

An unrelated point of interest:

We have run into some interesting applications over the years. One of these (laser scanner) resulted in my personal record for the highest RPM I have ever run a step motor to, 128,000 RPM.

This involved a Portescap 751 series ceramic disk step motor. Though miserable for CNC applications, the motors featured extremely high acceleration and speed rates.

Out of curiosity, a spare motor was run up to see just how fast it could go. The motor exploded as the speed passed thru 128,000 RPM. The explosion was contained within the motor housing. Internal examination showed the 2" diameter ceramic disk that formed the rotor was reduced to a fine sand.

Calculations later showed a curious coincidence; the tangential velocity at the rotor edge happened to equal the speed of sound at sea-level. Either mach effects or more mundane centripital forces tore the rotor apart.

I still like to think I took a step motor past Mach 1.:-)

Mariss

[fkaCarel]

Yes pears. Using a second language creates misspellings.

Any good drive should, do they? I studied a few and I consider it a feature.

Hydraulic servo's use switching frequencys of about 300Hz and pose therefore these scalable problems, and dictate with this frequency the real time behaviour. And have some further irritating issues also.

Laser mirrors is a field I know. They are depending heavily on inertia, vs dynamic beheavior. The effect of the switching should be totally lost in the required constant RPM. They are mechanical devices posing as electronics.

Carel

[Mariss Freimanis]

"They are mechanical devices posing as electronics." I like that; well put.

Step motors can be modeled as resonant low-pass filters. Any frequency that falls below the cutoff frequency (about 100Hz or so for a NEMA-23) will result in a mechanical response. If you are misfortunate enough for this frequency to be at the motor's resonant freq or sub-harmonic, the motor will be "pumped" into mechanical resonance.

This happens easily and frequently if the switching cycle isn't phase-locked with the step rate. Example: 20kHz switching freq and a 1.99 kHz full-step rate generates beat frequency and harmonics of 10Hz, 20Hz, ... 100Hz, etc. If 100Hz is the motor's resonant freq then the 10th harmonic still carries enough energy to pump the motor to desynchronization.

Only the switching frequency "sloppiness" of choppers (spread-spectrum?) saves the day somewhat but then you have to put up with the audible hissing, squealing, grunting, etc. these drives make.

Mariss

[pminmo]

"so there is a maximum speed where the zeros no longer become zeros and the ones no longer becomes ones."

Hmmm, I guess this means that a chopping frequency of zero means you can't step. :-) Good thing linear current regulation methods circuits don't read.

[Mariss Freimanis]

An L/R drive has a chopping frequency of 0Hz.

I don't understand the point of the mystical ones and zeros quote nor the celebration of drive illiteracy.

Mariss

[bunalmis]

Not any problem for L/R drivers.

There are problem by chopper.

For example, chopping frequency is 20 Khz. This means Ton time (energy time for coil) change between 0 to 50 usec. But next current period can't start before the 50 usec.

If your clk puls come at the every 10 usec (100 kHz) driver can't change the new coil currents before the 50usec.

Problem this.

[jeffs555]

I can understand your confusion. There is a lot of incorrect information in this thread, but listen to Mariss. He designs these drives and knows what he is talking about. I hope he will correct me if I get this wrong.

Like he said, for most modern drives, the chopper is a monostable, not a free running oscillator. It only controls the off time for the outputs. The on time is controlled by the current comparator. When the step pulse comes in, the monostable is reset and the output is turned on.

Also, at high step rates it is likely that the current will never have time to reach the set current, so the drive won't be chopping anyway.

[bunalmis]

@jeffs555

If chopper use monostable we cant talk from fix chopper frequency.

Yes monostable can control the off time but chopping frequency depended to motor inductance. Frequency continuously change by the mechanical load or motor to motor.

If your said true, I can use high speed logic circuits in the driver and I write

My driver use 20Khz for chopping frequency and max step rate is 1Mhz

But this is mislead. 1Mhz is hoodwink.