Low speed spindle motor / gearbox selection

[ZeroBacklash]

I need to run the spindle at low speed around 5 to 10 rpm. So I was thinking for going in for a 3phase, 8 pole motor and then add a gearbox to further reduce the rpm.

I need at least a 2.2kW motor here. I need to drive this not as a spindle but as an axis because I need position control. The motor will be running continuosly at 5 to 10 rpm.

My questions are:
1. Is a AC motor + gearbox + VFD + encoder a good fit for this application. I mean what would be the min, max speeds I can safely run the motor without overheating issues.

2. What about dc or ac servo motor or even a bldc motor.Are these well suited for low rpms / high torque. Or do I even need a gearbox for these too.

Thanks in advance.

Similar Threads:

• cnc spindle motor selection
• The Science of Spindle and Stepper Motor Selection
• NEED HELP spindle motor selection
• Hi-speed motor/spindle???
• Spindle Motor Selection

[ger21]

Get an AC servo with the appropriate power. No gearbox required.

https://www.automationdirect.com/adc...m_(Med_Inertia)

[Jim Dawson]

I need at least a 2.2kW motor here. I need to drive this not as a spindle but as an axis because I need position control. The motor will be running continuosly at 5 to 10 rpm.

You describe two different conditions here. A 3 phase motor / VFD combination would work fine for continuous operation, but for position control you would need a servo or stepper.

If you geared down any standard 3 phase motor with a 150:1 gearbox it would get you into the RPM range you need with enough torque to twist about anything into a pretzel. You say you need a 2.2KW motor, more importantly what torque do you need? A 2.2KW motor geared down 150:1 would output more torque than my Diesel truck engine is rated at.

My questions are:
1. Is a AC motor + gearbox + VFD + encoder a good fit for this application. I mean what would be the min, max speeds I can safely run the motor without overheating issues.

2. What about dc or ac servo motor or even a bldc motor.Are these well suited for low rpms / high torque. Or do I even need a gearbox for these too.

Thanks in advance.

Again if you need positioning then a servo would be best option. Most servos will output near 100% torque at 0 RPM with a pretty flat torque curve through the RPM range.

[ZeroBacklash]

ger21

Get an AC servo with the appropriate power. No gearbox required.

Good choice but pricey. (package coming to $2000)
How about DMM AC servos their drive + ac servo is around $600.

[ZeroBacklash]

Again if you need positioning then a servo would be best option. Most servos will output near 100% torque at 0 RPM with a pretty flat torque curve through the RPM range.

How about a Automation Direct VFD GS3 series with encoder connected to a 3phase motor. Will this give sufficient accuracy to do position control?
My torque requirements are minimum 100Nm.
Unfortunately DMM servo's are not meeting that They have max torque of around 12Nm.

[Jim Dawson]

How about a Automation Direct VFD GS3 series with encoder connected to a 3phase motor. Will this give sufficient accuracy to do position control?
My torque requirements are minimum 100Nm.
Unfortunately DMM servo's are not meeting that They have max torque of around 12Nm.

Unfortunately no, I have a GS3 and encoder on the 7KW spindle motor on my lathe. I thought I might be able to get it to position, but the GS3 is just too ''loose'' to make that work, I played with it for days trying to make it work. When I can find one at my price, I'm going to replace the spindle motor with a 10KW servo.

Since you only need 10 rpm, a 10:1 gearbox would give you 120Nm, to be really safe go with a 15:1 gearbox and you would have 180Nm available.

[ZeroBacklash]

Unfortunately no, I have a GS3 and encoder on the 7KW spindle motor on my lathe. I thought I might be able to get it to position, but the GS3 is just too ''loose'' to make that work, I played with it for days trying to make it work. When I can find one at my price, I'm going to replace the spindle motor with a 10KW servo.

Since you only need 10 rpm, a 10:1 gearbox would give you 120Nm, to be really safe go with a 15:1 gearbox and you would have 180Nm available.

Now thats a dampener. I mean regarding the VFD. I was planning to use an 8 pole 3ph motor (already low on rpm) with a VFD and encoder and with the GS3 drive I thought I would get sufficient accuracy to do position control.

Is threre no way to use a 3 ph AC motor with VFD and get sufficient accuracy for spindle operations with position control even with a high resolution encoder?. How much accuracy were you getting from your system.?

[Jim Dawson]

Now thats a dampener. I mean regarding the VFD. I was planning to use an 8 pole 3ph motor (already low on rpm) with a VFD and encoder and with the GS3 drive I thought I would get sufficient accuracy to do position control.

Is threre no way to use a 3 ph AC motor with VFD and get sufficient accuracy for spindle operations with position control even with a high resolution encoder?. How much accuracy were you getting from your system.?

I am not able to position at all. I am using a Galil motion controller so I have a lot of control over the way the system reacts. Thus far, no matter what settings I use in the drive or in the motion controller I can not control any positioning move. I can make it run at a constant speed or even programmatically vary the speed from about 5 to 5500 RPM and it's rock solid at any set speed. But any time I adjust any PID parameter above zero I lose control, I tried every trick I know and I've been doing this for about 25 years. The drive just does not react to speed change commands fast enough to act as a servo positioning system.

One thing I have thought about doing is adding a DMM servo to the spindle drive to use for low speed positioning for mill/turn functions and add a electro-mechanical system to clutch it in when needed, that is one way I've see it done. I have 1.8KW DMM servos on the axes, live tooling drive, and turret drive. Been very happy with them. But what I will probably do is just install a 10KW servo motor and be done with the problem.

In your case, a servo with low backlash gearbox would be the ideal setup. Low RPM and positioning resolution in fractions of an arc second.

[ZeroBacklash]

I am not able to position at all. I am using a Galil motion controller so I have a lot of control over the way the system reacts. Thus far, no matter what settings I use in the drive or in the motion controller I can not control any positioning move. I can make it run at a constant speed or even programmatically vary the speed from about 5 to 5500 RPM and it's rock solid at any set speed. But any time I adjust any PID parameter above zero I lose control, I tried every trick I know and I've been doing this for about 25 years. The drive just does not react to speed change commands fast enough to act as a servo positioning system.

One thing I have thought about doing is adding a DMM servo to the spindle drive to use for low speed positioning for mill/turn functions and add a electro-mechanical system to clutch it in when needed, that is one way I've see it done. I have 1.8KW DMM servos on the axes, live tooling drive, and turret drive. Been very happy with them. But what I will probably do is just install a 10KW servo motor and be done with the problem.

In your case, a servo with low backlash gearbox would be the ideal setup. Low RPM and positioning resolution in fractions of an arc second.

I'm trying to do something like this

And the vfd would be connected to the spindle drive. If you skip to 3 minutes into the video it will show you winding on a cylinder with dome shaped ends and you can notice how the spindle slows down slightly to allow for winding on the ends. So will this be achievable on a VFD and 3ph ac motor?

[Jim Dawson]

I'm trying to do something like this And the vfd would be connected to the spindle drive. If you skip to 3 minutes into the video it will show you winding on a cylinder with dome shaped ends and you can notice how the spindle slows down slightly to allow for winding on the ends. So will this be achievable on a VFD and 3ph ac motor?

Yes, that is doable. You would need to electronically gear the axes to the spindle with an encoder on the spindle. The ability to use electronic gearing would depend on what motion controller you are using, I don't know if all of them have it available. Looks like that winder would need some kind of custom software written to coordinate the motions, but it might be possible with standard CNC G code, maybe something like a rigid tapping routine.

[ZeroBacklash]

Yes, that is doable. You would need to electronically gear the axes to the spindle with an encoder on the spindle. The ability to use electronic gearing would depend on what motion controller you are using, I don't know if all of them have it available. Looks like that winder would need some kind of custom software written to coordinate the motions, but it might be possible with standard CNC G code, maybe something like a rigid tapping routine.

Ok thanks. I don't understand the electronic gearing part but will look it up. So do you still think that a servo would be best for this application for the spindle or a VFD is equally well suited to do the job? I mean if I use a VFD I would I get the position control for co-ordinated motion of the axis?

[Jim Dawson]

I mean if I use a VFD I would I get the position control for co-ordinated motion of the axis?

Yes, you make the spindle the master axis, and the slave the motion of the other axes to the spindle. That way, the other axes follow the spindle and all of the timing is done without regard to the spindle speed. This is how I rigid tap on my lathe without having position control of the spindle, I simply turn the spindle at some speed, and tell the Z axis to follow the spindle encoder at the correct ratio to tap the hole. Once the tap has reached the set threading depth, the spindle reverses and the Z axis backs out following the spindle rotation until the tap has reached its starting point. Winding a filament wound tube would work exactly the same way, for every spindle rotation, you would move the X axis some distance down the tube, then on reaching the end of travel, you would reverse the X axis direction and apply the filament in the opposite direction. As shown in the video, the X axis dwelled at the end of travel for some amount of spindle rotation before it headed back in the other direction. It's all done in the software.

Electronic gearing is simply slaving one or more axis to a master axis and setting a ratio of encoder pulses for the slave axes to move per the master axis encoder. For instance, you could tell the X axis to move 5 encoder pulses for every 2 encoder pulses of the spindle for a 5:2 gear electronic gear ratio. I hope that makes sense.

[ZeroBacklash]

Yes, you make the spindle the master axis, and the slave the motion of the other axes to the spindle. That way, the other axes follow the spindle and all of the timing is done without regard to the spindle speed. This is how I rigid tap on my lathe without having position control of the spindle, I simply turn the spindle at some speed, and tell the Z axis to follow the spindle encoder at the correct ratio to tap the hole. Once the tap has reached the set threading depth, the spindle reverses and the Z axis backs out following the spindle rotation until the tap has reached its starting point. Winding a filament wound tube would work exactly the same way, for every spindle rotation, you would move the X axis some distance down the tube, then on reaching the end of travel, you would reverse the X axis direction and apply the filament in the opposite direction. As shown in the video, the X axis dwelled at the end of travel for some amount of spindle rotation before it headed back in the other direction. It's all done in the software.

Electronic gearing is simply slaving one or more axis to a master axis and setting a ratio of encoder pulses for the slave axes to move per the master axis encoder. For instance, you could tell the X axis to move 5 encoder pulses for every 2 encoder pulses of the spindle for a 5:2 gear electronic gear ratio. I hope that makes sense.

That's some pretty amazing information. Many Many Thanks! So I can either use the electronic gearing way or the convetional way where I tell both the spindle axis and the x axis to rotate and move in a specified amount of time right?
Do you think the VFD + encoder will be accurate enough for this operation.
I mean if we take the threading analogy would you have done threading with a VFD encoder combo on the spindle?

[Jim Dawson]

That's some pretty amazing information. Many Many Thanks! So I can either use the electronic gearing way or the convetional way where I tell both the spindle axis and the x axis to rotate and move in a specified amount of time right?
Do you think the VFD + encoder will be accurate enough for this operation.
I mean if we take the threading analogy would you have done threading with a VFD encoder combo on the spindle?

For that application electronic gearing is a good option. It would be possible to do an interpolated rotary/linear move and accomplish the same thing, but this would require a servo on the rotary axis. The only place time enters into the equation is peripheral, not part of the function. In other words, you set a speed in encoder pulses / second and it takes some amount of time to accomplish the motion segment. The exception to this is when operating in contouring mode where you are telling the axes to arrive at a point and at a time, this is not that common in normal 4 axis moves. More something you might see in robot arm movements.

I'm not sure which method I would use for threading if I had a servo on the spindle, both accomplish the same movement. Not sure that one would be more accurate than the other. But electronic gearing was my only option without the servo spindle.

[ZeroBacklash]

For that application electronic gearing is a good option. It would be possible to do an interpolated rotary/linear move and accomplish the same thing, but this would require a servo on the rotary axis. The only place time enters into the equation is peripheral, not part of the function. In other words, you set a speed in encoder pulses / second and it takes some amount of time to accomplish the motion segment. The exception to this is when operating in contouring mode where you are telling the axes to arrive at a point and at a time, this is not that common in normal 4 axis moves. More something you might see in robot arm movements.

I'm not sure which method I would use for threading if I had a servo on the spindle, both accomplish the same movement. Not sure that one would be more accurate than the other. But electronic gearing was my only option without the servo spindle.

Currently I have 2 options
1.) VFD + Encoder + 3ph AC Motor + GearBox + Electronic Gearing
2.) AC Servo (Built in Encoder) + AC Servo Drive + Gear Box

The AC Servo may be more expensive system but it would give more acurracy. The VFD option much easy available I mean the ac motor is widely used and replacements easily available but on the whole its slightly low on accuracy. I talking about the spindle axis here.

[Jim Dawson]

The AC servo, while slightly more expensive, would give you more control options and perhaps greater accuracy. If I were building this machine and had the budget, I would go with the servo option. The most expensive part of the servo system would probably be the gearbox, required in either case, the servo and drive is relatively inexpensive. As I recall the 1.8 KW DMM servo and drive are in the US$600 range. The low backlash servo gearbox is in the US$1200 range from Automation Direct, there may be other options.