[kis-mat]

Hi CNC-Zone folks,

I'm new around here and diving into my first CNC machine project. Could really use your expertise.

For my build, I've got three AC Servo motors on the way (1kW power, 3.19Nm, 3000rpm, 5.7A nominal current at 220 Volts, inertia 2.0 * 10^-4 kg*m^2). The project calls for rapid acceleration, lots of stop-starts, and sudden direction switches, hence the aim is to keep the inertia mismatch of the system and the actuators under 1:1.

While waiting for the servos, I'm starting to worry about that all three motors are going to be linked to the same 16A fuse, and here are my concerns:

Inrush Current: I'm planning to use relays to start the motors one at a time, but could the initial power surge of each single motor still trip the fuse?

Acceleration Current: I'm worried the current during acceleration might go over the standard formula (I = torque * RPM / (9.55 * 220V * PF)) due to the constantly changing system dynamics and also trip the fuse. Anyone have a ballpark figure for this kind of current increase?

My plan to tackle these issues is:

- To place an inrush current limiter between the power supply and the servo drivers.

- To add some capacitors (1-2F) alongside the servos to even out acceleration peaks.

- To use a current sensor for real-time monitoring on each driver, and limit the acceleration dynamically if the current gets too high.

Does this make sense to you guys? Oddly enough, I haven't found much online about dealing with these issues, but I can't be the only one worrying about this, right?

Maybe I'm just over-analyzing things, since I have never used servo motors before. After all, I'll be running the motors at no more than 1500 rpm (just 50% of max), and an S-curve velocity should help taper off the acceleration at start and higher speeds.

Any thoughts or advice would be awesome!

Best,
kis-mat

[Al_The_Man]

All depends on the controller system..
Normally servos ore not switched on to full power, there is some accel/decel built in., and the drives take care of current monitoring.
]

[kis-mat]

Thanks for the quick reply!

I'm planning to control the motors myself using pulse/dir, so I reckon the S-curve will be managed within my pulse frequency.

If every drive has a current monitoring feature, that's phenomenal! I haven't found anything about this in the manual. I ordered the motors from StepperOnline in China, but perhaps there might be something in the settings.

So, it looks like my worries and planned solutions (capacitors, etc.) might not be necessary after all?

[pippin88]

Are you talking about a 16amp main circuit breaker? Anything else running on the circuit? What is the power point rated at?

You won't be able to have all three 1kw running at peak at the same time on a 16 amp breaker.

Servos are rated at continuous power limit.

So 1kw is continuous power rating, 3.19nM is continuous torque rating etc.
Usually they will be capable of brief bursts of peak torque/power about 3x this.

But it will come down to what load you are actually putting on the servos, how many at once etc.

You should be able to set a torque limit, which is effectively a current limit.

I don't think you can add capacitors.

I tried to find in rush current figures for servo drive startup (there is not the usual motor in rush current, because the drive controls this, but there is a big bank of capacitors that get charged) but couldn't find much.

[kis-mat]

Thanks for the insight!

Yes, I was referring to a 16amp main circuit breaker, and the only other thing on this circuit would be a light bulb in the room.

Regarding the motors, only two of them (X and Y axis in a CoreXY setup) would be running simultaneously, with the Z-axis motor kicking in later.

Setting a torque limit sounds like a solid start, but the limit might need to vary depending on whether one or two motors are running simultaneously. Having a robust model or measurement would be useful for this.

Actually I am not sure if the drivers are able to set a torque limit. There is no description of such function in the manual.

It's a bummer about the capacitors! But it's reassuring to hear that the drives might counter the inrush to some extent.

Again, thanks for the guidance. It's helping me piece together this puzzle!

[pippin88]

What is this machine? A large printer?

T6 model servo?

T6
Pr0.13 "1st Torque Limit"
Default is 300 (300%)
Page 36 of full manual

Servo continuous torque is 3.19, peak is 9.56, i.e. 3x 3.19.

Rated continuous amps of servo is 5.7amp

Set Pr0.013 to 100, and the servo will limit torque (therefore current) to 3.19nM (about 5.7amps).

[kis-mat]

Thanks for your input!

I'm putting together a vertical Pick and Place robot, with an active workspace of 2x2 meters (2.20 x 2.20 meters in total). The X and Y axes position first, and only when near the target does the Z-axis perform a quick pick or place movement. That's why only two motors are ever active at the same time.

Oh, and they are T6 motors - have you used these before?

I must admit, I've only been handling the simple manual so far - sorry about that oversight. Time to dive into the full manual! The motors arrived today, so it's becoming quite real now.

Thanks for the help!

[joeavaerage]

Hi,
I have 750W servos and 99.9% of the time they draw well under 1A each. The only time they ever draw more than that is at those brief moments of peak acceleration.
You do need D curve or motor rated breakers, as they tolerate these short burst of high current without nuisance tripping, but otherwise I'd say that a 16A breaker will be fine.

Craig

[joeavaerage]

Hi,

- To add some capacitors (1-2F) alongside the servos to even out acceleration peaks.

DO NOT PUT CAPACITORS ON THE SERVO PHASE LINES. The voltage supplied to the servo is pulse width modulated and has a high frequency component. You add a capacitor
and you will VASTLY increase the switching stress on the IGBT's. DONT DO IT! In fact the servo drive is expecting, indeed demanding, that it looks into an inductive impedance to smooth
the current....adding capacitors is exactly the wrong thing do.

If you wish to fit filters to the servo phase lines, usually because of the long distance between the servo and the drive, follow the manufacturers recommendation TO THE LETTER or you risk blowing the drives.

If you wish to put line reactors on the input side to the drive, that's a good practice, conservative but good.

- To use a current sensor for real-time monitoring on each driver, and limit the acceleration dynamically if the current gets too high.

The servo and servo drive will look after themselves. Do not try to second guess the manufacturer, they don't need the precautions you seem to think they require, they will accelerate back and forth
all day without demur, they'll shake your machine to bits before they even look like coming under stress. The manufactures current loop bandwidth will be easily 5kHz, maybe as high as 20kHz,
you'll never even come close to matching that.

If you want to limit the current and/or acceleration that is programmable....use it. There are hundreds of parameter within a modern servo drive....do you have manufacturer supplied setup and
tuning software? You'll need it.

Craig

[kis-mat]

Hi Craig,

Thanks for your input!

In my use case, peak accelerations aren't an exception but the norm. A 1kg mass is accelerated several times per second with 100-200 m/s^2 and stopped again.

My intention for monitoring the power supply to the drivers wasn't because I distrust the drivers. But the maximum current each driver is allowed to draw varies depending on how many motors are active and how strong they are accelerating. There's no reason to limit each driver to 5A unless all three are running simultaneously. That's why I'd hoped for some dynamic regulation. But it seems for the moment that this is too complex for now.

For now I'll limit the power output to the motors with Pr0.13 in the driver to ~4A and hope is that the driver's output current isn't much lower than its input current. In addition I'll buy some D-rated 5A fuses as suggested.

Thanks for the advice!

kis-mat

[joeavaerage]

Hi,

In my use case, peak accelerations aren't an exception but the norm.

High speed production machines do this day in and day out for years at a time. You are worrying about something that you don't need to.

Suggestion: Try a D curve 16A breaker and see what happens. If it proves adequate, for which I'd give you 95% probability, then no more need be done.
If its inadequate THEN start looking at solutions....but don't try to cross a bridge that you will likely never encounter.

Craig

[kis-mat]

....but don't try to cross a bridge that you will likely never encounter.

I guess you are right. Ill wait for the breakers to arrive and than see what will happen.

Thanks a lot!

[joeavaerage]

Hi,
have you used AC servos before?

If you have not then I expect you will be staggered at just how responsive they are.

I use 750W Delta servos, they have a 160,000 count per rev encoder, so are nothing special, just entry level modern servos.

They can accelerate from standstill to 3000rpm (rated speed) in 13ms at rated torque, and 4ms if you want to use overload torque. Its so violent that it makes a lond 'bang' that
you would swear sounds like a breakage or failure....its not, its just responding to your command. Presumably the instantaneous current would peak during the acceleration, but if you are
looking at the ammeter you'll barley see it twitch....hard to credit but thats what happens.

I would be less concerned about supplying power to the servo drives than building you machine is such a way that the violent accelerations don't shake it to bits.

My mill has a 115kg cast iron X axis, and with vice and workpiece I estimate 150kg. With my fourth axis trunnion and fifth axis mounted its more like 200kg. Using just rated torque (2.4Nm)
the axis can accelerate at 2.7m/s², or 0.27g. It will accelerate to 25m/min, travel the length of the axis, 350mm, decelerate to a stop in 1.48s. As you can imagine this
causes my machine to lurch around, its on big castors, so when I use its full potential I have to tie the machine (800kg) down otherwise it takes off across the workshop!

The bottom line is that you need a well designed and built machine just to absorb the potential accelerations that servos are capable of delivering. Supplying power to the servos is the easy bit!

Craig

[kis-mat]

Short update after I put the motors into operation:

You were right on both counts: 1) power won't be the problem and/becuase 2) the driver already handles the problems.

A single motor pulls more at peak acceleration than it should when all three are running at the same time, but the driver reliably limits to the value specified in PR0.13.

Next step is the communication with Arduino via RS485 interface, but that should not be a problem. Thanks a lot guys!


@Craig:
I see what you mean. I think I will have to drill the finished cnc machine into the wall to be safe enough, especially since my machine is upright.

As you might have already guessed: its my first time using a AC motor.

Thanks a lot!