1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  2. PLEASE: Read the FORUM RULES BEFORE registering!

    Dismiss Notice

Servo Drive for Lathe Spindle

Discussion in 'CNC IN THE HOME SHOP' started by barnbwt, Apr 28, 2017.

  1. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    Anyone having luck using a servo-driven C-axis spindle to turn *and* mill stuff?

    I am trying to determine my best options for an upcoming CNC lathe project. As is always the case, the goal is a machine that 'does eveything' but compromises for the resources available. Threading at the very least is a necessity, since this project needs to deliver more functionality than my manual lathe to be worthwhile.

    The requirements thus far:
    - 110VAC/15A-1ph wall power source
    - 1.0-1.5hp spindle power (so I still have enough reserve juice for the axis steppers & controller)
    - 4000 spindle speed, belted to motor (fast enough for proper carbide sfm on 1"-2" aluminum parts)
    - Articulating spindle (single point threading, live-tool engraving, live-tool light milling/slotting/broaching, possibly rigid tapping if motor is strong enough)
    -widget-sized parts in small quantities (so no need for very long duty cycles or aggressive cutting)

    Points 3 and 4 seem to suggest a servo, so I am following that rabbit hole at this time.

    The questions thus far:
    -If using the servo as a controllable C-axis for engraving or light milling (1/8" shank or smaller), will the servo's resistance & mass be sufficient to control tool chatter & cutting forces? These are 4N-m range.
    -Is it more efficient to run a high-speed stepped-up-high-voltage servo geared down to the desired RPM, or gearing up a low-speed high-torque servo that operates at near line voltage?
    -Most drivers run on Asian/industrial 220-240VAC sources; are there any downsides to using a driver that operates closer to my line voltage?
    -Is there any advantage to high vs. low voltage AC brushless servos, or is it really just a function of which controller you want to use?
    -Any estimate of losses for transformer or switching voltage booster systems would be helpful to determine my final current draw.
    -At what region of the operation envelope do servos+controllers draw the most source current? Is it whenever they are applying max torque, or rather when they are running at high speeds under maximum voltage?
    -Could I feasibly use a larger motor and limit its current through the driver software, and would I be giving up torque or speed (or both)?

    Anyone having luck using a servo-driven C-axis spindle to turn *and* mill stuff?
     
  2. JimDawson

    JimDawson Global Moderator Staff Member Director

    Likes Received:
    3,968
    Trophy Points:
    113
    City:
    Sandy
    State:
    Oregon

    -Return to Top-

    The limiting factor is your input power. If there is a way you can get 240 volts at 15 amps your options increase dramatically.
    Take a look at these motors http://www.dmm-tech.com/ac_servomotor_main_a1.html

    Normally you get better performance from the higher voltage servos. The torque curves are better on the high voltage servos and normally are capable of higher RPM.

    Adding in the mass of the chuck & spindle should be adequate to control chatter. Running the motor at a higher RPM and gearing it down would be some advantage.

    Max current draw is normally at max torque.

    Transformer losses are negligible.

    There is no reason that a servo C axis won't work. It just requires enough power to run it. All modern horizontal machining centers have servos on the spindle.

    Looking forward to seeing your build.:encourage:
     
  3. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    Yeah, a proper 30A connection (or full 240V connection) is not in the cards at this time; residential standard outlet is the only option in my garage --why do you think I asked the question here vs. CNCzone or Practical Machinist, lol? ;) Now, I'm not opposed to 'growth' potential, as far as leaving some additional capacity that a better power connection could realize at a later date (i.e. limiting driver current draw by holding voltage/speed below maximum levels)

    I fully expect to have some trade-offs relating to input power limitations, same as with my 2hp 12x28 manual lathe; I'm just hoping to get my surface speeds up fast enough to get more optimum cut quality with shallower passes made convenient by the CAM automation. This machine will be far more rigid than my manual lathe with its gimpy gib contact surfaces, so I think it's doable (well, assuming it doesn't ring like a bell in practice, which is entirely possible since I'm not bedding this in epoxy granite or cladding it in cast iron)

    I've been looking into the Baldor offerings since there seem to be a lot of 110VAC driver options in the kW class I'm interested in, and also used motors this side of the oceans. The one thing that seems off is their motors are rated to a much higher voltage than the controllers can reach (600V vs. 250V), but I understand that just means top speed is limited as opposed to torque; I assume their industrial drivers running on higher line voltage are what can reach the rest of the way up there. It would really be helpful if that company would publish a cross-reference of compatible motors/drivers like most of the others do. Parker ('compumotor series') also has some products that look usable, but seem to run at slower speeds than Baldor's.

    Teaser picture for ya'll...
    Assembly Concept.png
    Work envelope of 4"OD (up to 9" in a pinch) x 8" length; tool travel 6" in X & 15" in Z; overall size 32" long x 26" deep x 30" high. 5C spindle bore and AXA toolpost in the current incarnation. NEMA23 steppers on X & Z axis, probably 400 oz-in range, run off a Gecko controller or something similarly well-known. Future planned additions are a tailstock that hangs down from the lower step surface, a rotating tool changer mounted on the large inclined surface, and a chip-evacuation auger along the very bottom/front corner --not really pursuing these goals at this time, but leaving some room for them.
     
    JimDawson likes this.
  4. Metal

    Metal United States Active Member Active Member

    Likes Received:
    95
    Trophy Points:
    28
    City:
    Buffalo
    State:
    New York

    -Return to Top-

    Likely silly idea
    A servo motor isn't too different from a plain old DC motor (in fact the ones on my mill are just expensive DC motors with encoders on them running @72v)

    Ampflows are pretty powerful motors (up to around 11 horsepower at 48vdc) you'd just need to figure out an encoder
     
  5. cmantunes

    cmantunes United States Active Member Active Member

    Likes Received:
    15
    Trophy Points:
    8
    City:
    Herndon
    State:
    Virginia

    -Return to Top-

    There is no advantage to using brushed DC motors anymore. High performance DC brushed motors aren't that much cheaper than equivalent brushless DC motors but they do have some disadvantages: 1) they are noisy; 2) they require periodic replacement of brushes; 3) commutation is fixed so they don't offer precise torque control; 4) they aren't capable of field weakening, allowing them to run above their rated RPM for corresponding rated voltage (constant power rather than constant torque region).
     
  6. Metal

    Metal United States Active Member Active Member

    Likes Received:
    95
    Trophy Points:
    28
    City:
    Buffalo
    State:
    New York

    -Return to Top-

    Sensorless brushless motors have zero startup torque as they need to go through a startup sequence in order to begin spinning under power
    Sensored ones and controllers to go with get very pricey very quickly.

    a couple hundred dollar magmotor and hundred dollar servo controller gives enough power for me, an equal sensored brushless and controller would cost ten times as much.
     
    Eddyde likes this.
  7. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    *quick question at last paragraph

    There's also the aspect that a permanent-magnet async motor has far lower turing inertia than anything using windings at a given power level (even brushless). That's why servos are able to start/stop/accelerate so quickly, and why they bothered to develop control systems more responsive than a VFD. Also why a servo the size of a Heineken can puts out the same power as an induction motor the size of a bowling ball.

    But yeah, much like how data transfer tech advances are rapidly blurring the line between (and dis/advantages of) servos and steppers, power supply advances are blurring the line between speed-controlled motors (of various types) and true servo systems. It's really confusing for a neophyte like myself when ignorant/misleading sellers are offering "closed loop stepper motor servos" as if that isn't a soup-sandwich of jargon, lol. Spent a half hour looking over an Allen Bradley catalog at 110V-fed servo drives, before I realized it was really a list of VFD's ('inverter drives') that simply had many of the same features/terms, but really weren't set up for precise closed-loop position control.

    If I didn't need to do anything under 30rpm or so, I'd be going simple speed controller all the way. But I really want to be able to mill helixes & do engraving, and a VFD simply won't control things precisely enough (well, the really nice modern ones that are indistinguishable from servo drivers might)

    Here's what I'm picking out so far;
    Baldor BSM80C-375AF: ~1kW, max speed 4000rpm, 2500ppr encoder, 3.6N-m
    Baldor FMH2A09TR-EN43: 9A max cont. (18A peak will pop my breaker), 110VAC & 24VDC input, step/dir input (also analog +/-10V for speed/torque)

    Given the speed & encoder, I should able to drive the thing on pulses alone just like a stepper motor across its full range so long as I can generate them at 167kHz from the controller. That's certainly the most convenient option, that won't require swapping into analog speed control mode halfway through a job & back. It looks like several of the common hobby-level motion controllers are now capable of this output speed (e.g. smoothstepper, etc) without losing low-speed resolution.

    Rough block diagram:
    Mach3/4 laptop -> ethernet -> TBD Motion Controller/Breakout Board -> Baldor Microflex Driver & Servo, TBD qty(2) Stepper Drives/Motors

    "Ampflows are pretty powerful motors (up to around 11 horsepower at 48vdc) you'd just need to figure out an encoder"
    Also how to run an 11hp motor from a residential wall outlet...figuring out the encoder feedback such that it is reliable & accurate is no small task from what I understand. I'm having a hard enough time just figuring out 'turn key' factory paired setups, let alone designing my own closed loop system; I studied Mechanical precisely so I wouldn't have to do that stuff, lol.

    *Quick question; tons and tons of Baldor/etc motors & controllers on ebay & elsewhere in all sorts of configurations for pennies on the dollar; where the heck are their cables? Do surplussers just throw them away? That seems crazy, since they rival the cost of the used components wherever a few are listed, and the components are useless without them, so it's not like the sellers simply aren't bothering with the small-fry items. A new set of power/encoder cables for those Baldor units (still in production, but not top-flight items) is well over a grand! I'll be making my own from bailing wire & chewing gum at those prices.

    TCB
     
    Last edited: May 3, 2017
  8. cmantunes

    cmantunes United States Active Member Active Member

    Likes Received:
    15
    Trophy Points:
    8
    City:
    Herndon
    State:
    Virginia

    -Return to Top-

    That Baldor is a beauty and, as usual, expensive as hell. One thing about this motor is that it is actually a permanent magnet synchronous motor, a cousin to the brushless DC motor but with sinusoidal rather than trapezoidal BEMF. You can't do much better than that motor, it is pretty much state of the art. Looking forward to seeing it in action, if you decide to go that route.
     
  9. cmantunes

    cmantunes United States Active Member Active Member

    Likes Received:
    15
    Trophy Points:
    8
    City:
    Herndon
    State:
    Virginia

    -Return to Top-

    You are right about sensorless brushless motors and low startup torque. The "equivalent" to a brushed DC is a brushless DC with hall sensors. There are several hobby lathes and mills being sold with this kind of spindle motor, usually providing around 1-1.5kW.
     
  10. Metal

    Metal United States Active Member Active Member

    Likes Received:
    95
    Trophy Points:
    28
    City:
    Buffalo
    State:
    New York

    -Return to Top-

    Absolutely
    However these are manufactured overseas (typically) and the watts shown are the watts the motor consumes, not the actual output watts, which makes it very difficult to spec out. (ex: I have a local with a 2.something KW spindle, my 1hp bridgeport cuts much faster, and better) Don't buy anything you can't get a power curve from the manufacturer for.
     
    cmantunes likes this.
  11. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    Well, I've 'gone that route' since those two units are on the way here, for 390$ used (not bad, assuming I find a decent cabling solution)

    So, as my last question for those with experience, if I'm running my spindle servo via step/dir like my axis stepper motors, and I'm using the servo's encoder feedback to ensure the spindle is following steps, do I have the control/coordination needed to do threads/tapping from an off-the-shelf unit (let's say Smooth Stepper for argument), or is additional feedback to the actual motion controller necessary? From what I understand it absolutely is in the case of the speed control "spindle" output these cards have (the +/- 10VDC analog signal*) to bring additional encoder feedback to the motion controller so it can make double-sure the feed axis is keeping up with the rotary, but what about the same maneuver between two proper stepper drives, assuming no missed steps? If so, then I need to make sure to use a controller that can parse the simulated encoder output of the servo driver.

    TCB

    *That output may ultimately control the spindle of a live tool, I think
     
  12. cmantunes

    cmantunes United States Active Member Active Member

    Likes Received:
    15
    Trophy Points:
    8
    City:
    Herndon
    State:
    Virginia

    -Return to Top-

    Awesome! Did you buy this on eBay?

    For threads/tapping, usually you run your servo in speed, rather than position, mode. The CNC controller gets the pulses either from the encoder directly or from a "synthesized" encoder provided by the servo drive. The CNC controller then synchronizes the Z-axis movement to the angular position of the spindle. It is only when you want to do milling on the lathe that you switch the servo to position mode and the spindle becomes the C-axis. Then you control the X, Z and C axes to do a (limited) amount of milling on the lathe with live tooling. This is the way a "real" CNC turning controller works. I don't know enough about the SmoothStepper to know if it supports this kind of operation. Here's an example of what I'm talking about:
    [/QUOTE]
     
    JimDawson likes this.
  13. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    I've not even gotten to the programming side; is that control method switch just a G-command, and can be changed mid-tape? I'd figured it was dependent upon how you wired things, and it was either/or. I wonder if the big machines have an entirely parallel driver that a controller selects depending on local parameters like commanded spindle speed.

    TCB
     
  14. cmantunes

    cmantunes United States Active Member Active Member

    Likes Received:
    15
    Trophy Points:
    8
    City:
    Herndon
    State:
    Virginia

    -Return to Top-

    A Fanuc controller, for example, has a pair of M-codes to switch-on/off C-axis mode and it can certainly be done mid-tape. This usually controls a digital I/O to tell the servo drive whether to do speed or position control.
     
  15. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    M code was actually my second guess, lol (see, I am learning ;))

    TCB
     
  16. JimDawson

    JimDawson Global Moderator Staff Member Director

    Likes Received:
    3,968
    Trophy Points:
    113
    City:
    Sandy
    State:
    Oregon

    -Return to Top-

    Basically you would run the C axis at some RPM by a M3 Sxxxx, then slave the Z axis to the C axis at some ratio that would move the Z axis the proper amount per revolution to generate threads. This would be commanded by the G-code with a G76, but all of the real programming to make this happen is internal in the motion controller.
     
  17. barnbwt

    barnbwt United States Active Member Active Member

    Likes Received:
    43
    Trophy Points:
    28
    City:
    Greenville
    State:
    Texas

    -Return to Top-

    Okay, if you can switch between drive modes on the fly like shifting gears, then things make a bit more sense. So for threading, the servo drive will be looking for an analog spindle velocity/torque signal on the 10V input line from the motion controller (both with encoder feedback), which will direct a simultaneous step/dir signal output from the controller to the stepper driver running the relevant axis at the proper ratio. For milling or engraving, the servo drive will make the desired finite spindle moves based on step/dir inputs off the motion board (both with encoder feedback) independent of the other axes.

    That does seem more efficient from a data standpoint, than "go-go-go-go-go-go-go-go-go-go-go-go-go-go..." for the entire spindle operation ;). Is velocity control generally 'smoother' than positional control for a servo, and more like a conventional VFD variable speed control (only precise)? Or perhaps the electronics in the driver can 'see ahead' to the incoming code and interpolate between steps on the fly to smooth them out microstep-style so there isn't much functional difference, apart from the bandwidth usage? I suppose it doesn't matter a huge amount, if both modes are accessible, since I can simply choose whichever yields the best result.

    I guess my thought was for something like single-cut threading with a number of passes, the step based approach made more sense due to the number of spindle velocity changes involved (and also precise positioning). I guess what you're saying is the accuracy of either servo control method is comparable due to the encoder feedback, and whether slaved or independent, the two axes involved in the thread will be synched by the same motion control board that compensates for their latency (with at least the spindle running closed-loop on its encoder). Do you think there's much to be gained by putting an encoder on the Z-axis early on? I planned to do this later, as a means to incrementally improve upon hopefully already-acceptable performance.

    TCB
     

Share This Page