Servo as Spindle Motor

[Hellbringer]

I have searched and i have seen quite a few posts about this topic but i need a little more info.

i want to change my x2 motor our for a servo so i can control the speed with mach and be able to tap.

for the question

1. i am thinking about getting the KL34-150-90 from Keling (specs posted below) will 600oz-in be enough torque to act as a spindle... i 99% of the time cut aluminum and i would like to be able to use 1/2 - 5/8 endmill and .125-.25in cuts at a time? i am also thinking about running 1:1 so the spindle speed would be (i am guessing) around 5500-5800rpm.

2. i know i will need a driver and a encoder to use the servo but i have no clue what drive and encoder i should use any suggestion/links?

3. i read a post saying get to get the steps per in mach 3 to be correct to the spindle speed that mach is registering you need some other type of sensor. did i misread that? or dose the encoder send that information to mach. if there are other sensors that are needed what are they and do you have links to where they can be bought.


KL34-150-90 Specs

Dual Shaft, Front Shaft
1/2" Rear shaft 1/4"
600 oz-in Peak 90V/38A $129 Special Sale $119
Green is Ground, White is +, Black is Common)
6000 RPM (no load)

thanks
Mike

[kawazuki]

I believe servo encoders only output steps made and direction so controller can determine where tool is compared to where it should be? There is no time reference
to establish a speed of rotation so i'm guessing an additional pulse will be required -
this is guesswork though!

[M250cnc]

Mike milling machines have the spindle power defined in HP

I would go search an online conversion site to see how the numbers stack up

But my guess is that it is not suitable at that power at full speed you cant tap for tapping you need slow revs

Most peeps go the VFD and suitable motor, simpler and cheaper & Mach is not able to rigid tap at this moment

Phil

[irving2008]

To see if thats possible you need to know the feedrate, the number of flutes and the width and depth of the cut. You want to use an 1/2" (say 12mm) endmill and take a .125" (3mm) deep cut in alloy.. Assuming a 4-flute cutter and a cut of 1/2 tool diameter (6mm) then:

The cutting speed for alloy is 120m/min so for a 12mm tool you need 120 * 1000/(pi * 12) = 3200rpm

The chip load for alloy is approx dia/150 for roughing, d/200 for finishing. Assuming roughing for a 3mm deep cut, chip load is 12/150 = .08mm so the feed rate @3200rpm for a 4flute cutter is .08mm * 4 * 3200 = 1024mm/min

The material removal rate is depth * width * feed rate = 3 * 6 * 1024/1000 = 18.4cc/min

Power(W) = specific cutting force * removal rate/1000 = 17 (for alloy) * 18.4/1000 = 313W (~1/2HP) and the torque needed is P * 60/(2pi * revs) = 313 * 60/(6.28 * 3200) = 0.93Nm (130oz-in).

However the motor losses, mechanical losses, tool quality and sharpness will increase this by 100% or more. Your cutting requirement therefore roughly needs a 1HP (750W) motor with an on-load speed of 3200rpm.

So, IN THEORY, this servo motor - assuming you can find a way to control the speed with no loss of torque - will be marginal for the sort of cut you want at a feed rate of about 1000mm/min (40ipm) - it would probably be OK with a 2-flute cutter and a feed rate of 500mm/min (20ipm).

Many factors will affect this.. rigidity of the machine, duty cycle of the motor (power drops as it heats up), quality of the tool, etc. etc. HOWEVER, given the price of small AC 3-phase motors, timing belts (a 4pole motor and a 2:1 gearing up will get to 3400+ rpm) and 1PH to 3PH VFDs I do wonder if the servo route is the expensive option.

[Hans_G]

That motor is similar to the Sherline spindle motor. It probably wouldn't be much of an upgrade power-wise, but it would give you speed control & tapping from gcode.

For Mach3 you probably want a low resolution encoder or an encoder with an index pulse (once per rev) and run the index into Mach3.

[SCzEngrgGroup]

The on-load speed of that motor will be about 4800rpm, assuming the torque of 600oz-in (about 4.2Nm) is the max output torque at those revs the motor power is 4.2 * 4800/60 * 2pi = 2kW or about 2.8HP (for a 90 * 38 = 3.4Kw input, so 60% efficiency is about right)

To see if thats possible you need to know the feedrate, the number of flutes and the width and depth of the cut. You want to use an 1/2" (say 12mm) endmill and take a .125" (3mm) deep cut in alloy.. Assuming a 4-flute cutter and a cut of 1/2 tool diameter (6mm) then:

The cutting speed for alloy is 120m/min so for a 12mm tool you need 120 * 1000/(pi * 12) = 3200rpm

The chip load for alloy is approx dia/150 for roughing, d/200 for finishing. Assuming roughing for a 3mm deep cut, chip load is 12/150 = .08mm so the feed rate @3200rpm for a 4flute cutter is .08mm * 4 * 3200 = 1024mm/min

The material removal rate is depth * width * feed rate = 3 * 6 * 1024/1000 = 18.4cc/min

Power(W) = specific cutting force * removal rate/1000 = 17 (for alloy) * 18.4/1000 = 313W (~1/2HP) and the torque needed is P * 60/(2pi * revs) = 313 * 60/(6.28 * 3200) = 0.93Nm (130oz-in).

However the motor losses, mechanical losses, tool quality and sharpness will increase this by 100% or more. Your cutting requirement therefore roughly needs a 1HP (750W) motor with an on-load speed of 3200rpm.


So, IN THEORY, this servo motor - assuming you can find a way to control the speed with no loss of torque - will allow the sort of cut you want with some margin for error (if anything its overkill) as long as you can get your feed rate to about 1000mm/min (40ipm) - if you can't you'll need a 2-flute cutter and a feed rate of 500mm/min (20ipm) and correspondingly less power from the spindle.

Many factors will affect this.. rigidity of the machine, duty cycle of the motor (power drops as it heats up), quality of the tool, etc. etc.


The key issue here is can you control the motor speed AND maintain the torque? The calculations suggest that even if the torque drops off by a factor of 2 this motor will still do the job.... HOWEVER, given the price of small AC 3-phase motors, timing belts (a 4pole motor and a 2:1 gearing up will get to 3400+ rpm) and 1PH to 3PH VFDs I do wonder if the servo route is the expensive option.

You're overlooking the fact that the motor is rated at 600 oz-in *peak* torque. The continuous torque rating is almost certainly *under* 100 oz-in. If you try to run it at peak torque for any length of time, it *will* overheat and fail quickly. A spindle motor, other than perhaps a BLDC, will need fan cooling....

Regards,
Ray L.

[irving2008]

You're overlooking the fact that the motor is rated at 600 oz-in *peak* torque. The continuous torque rating is almost certainly *under* 100 oz-in. If you try to run it at peak torque for any length of time, it *will* overheat and fail quickly. A spindle motor, other than perhaps a BLDC, will need fan cooling....

Regards,
Ray L.

Ray,

Good point Ray, I misread that , thats what comes of posting while still asleep!

As you say thats the stall (peak) torque... the run torque from the data sheet is actually 119oz-in continuous at 7.9A, or 0.9Nm, Output power is about 450W (0.6HP) at around 4800rpm with 72v/7.6A input (see Mariss' test results here

So, still a useful motor, but probably just marginal for this requirement...

I've edited the original post so as not to give out duff info any more than necessary...

[yantra3d]

This is going to sound bad but... I would urge you to reconsider using the keling servo for this app. Based on my experience with the KL23 on my x3 axis' I can tell you with almost certainty that it will not have the torque you expect from it.

The KL23's barely have the power to drive my axis' when under a reasonable cutting load and I am using a 3:1 gearing with them. Based on the specs of these servo's they should be more than adequate but unfortunately, they are not.

As an example, I just replaced my Z axis KL23 servo with an Aerotech 1000DC that is actually rated less powerful than the keling, but the keling would stall whenever I tightened the Z axis gib even the slightest bit while I ended up stripping the threads on my Z axis lock trying to stall the 1000DC motor and it still never stalled!! The holding torque is tremendous compared to the kl23 and yet it has a weaker rating than the keling!???!

I was able to rotate the ballscrew by hand with the keling servo (even with the 3:1 gearing) and if I try that with the Aerotech servo it will not budge.

I am now in the process of replacing the other two kl23's with high-end servos like the Aerotech. The difference is literally night and day. These kelings just don't have the torque required to hold up to their rating imho. That has been my experience with the ones I have anyway.

You may have better luck with the larger kl34 and it may be fine for your application but I just wanted to warn you that their performance is not going to live up to an industrial-grade servo with the same rating. I honestly would not even consider using one of the kl servo's to drive an R8 spindle.

[irving2008]

This is going to sound bad but... I would urge you to reconsider using the keling servo for this app. Based on my experience with the KL23 on my x3 axis' I can tell you with almost certainty that it will not have the torque you expect from it.

The KL23's barely have the power to drive my axis' when under a reasonable cutting load and I am using a 3:1 gearing with them. Based on the specs of these servo's they should be more than adequate but unfortunately, they are not....

I'm not surprised, the KL23 are only 42oz-in motors, the Areotech is rated at 135oz-in, bigger than the KL34 even...

[yantra3d]

No, my Aerotech 1000DC is only rated at 35oz/in. The kl23 is supposed to be rated at 42oz/in!!!

[Al_The_Man]

The Aerotech 1000DC motors range from stall torque 17 oz-in to 960 oz-in. skewed rotor type.
Aerotech have been manufacturing quality servo's for quite some years now.
Al.

[Hellbringer]

Most peeps go the VFD and suitable motor, simpler and cheaper & Mach is not able to rigid tap at this moment

can someone explain this a little more.... i have done a little research and i think VFD is Variable-frequency drive.

1. if that is the case can you control it with mach,
2. is the mach control rpm accurate enough to be able to use it for tapping?
3. do you need a special style of motor to use this driver? or can it just be any dc motor?

Also did not know that mach cannot do rigid tapping i have been thinking about trying EMC2 and i am pretty sure EMC2 can do it.

thanks
Mike