584,846 active members*
4,180 visitors online*
Register for free
Login
IndustryArena Forum > CNC Electronics > CNC Machine Related Electronics > Replacing tool turret servo system
Results 1 to 6 of 6
  1. #1
    Join Date
    Apr 2013
    Posts
    8

    Replacing tool turret servo system

    I would appreciate a practical viewpoint in a problem
    involved in turret servo repair of Femco HL15
    CNC lathe.
    Original Mitsubishi Melservo MR-J2-A-S24 stopped
    working obviously because of failed encoder. (It runs perfectly
    when the encoder is borrowed from another identical machine for
    testing) According to Mitsubishi policy the entire servo unit is
    meant to be replaced. This brings us to the idea of repairing this
    machine in practical and independently sustainable way. And further...
    in a way that serves CNC-lathe owners in general.
    My idea is to rebuild this turret servo using GD servo
    amplifier with suitable and readily available encoder. Within my knowledge
    and documentation of interface between servo system and CNC controller
    (FANUC Oi-T) I see two main problems to be solved.

    #1. Absolute positioning system.
    - Original system comprises complicated absolute positioning memory.
    Battery back-up for SRAM in servo, battery line for encoder where there
    is an additional ultracapasitor to allow it to be disconnected from
    servo amplifier for on hour or two. By the battery going flat within
    intervals of few years there is a procedure to regain absolute
    positioning.
    - According to my understanding this positioning memory is needed
    even thought the encoder is of absolute type. This is because mechanical
    transmission ratio (1:9) between turret and motor. So it retains the track
    of motor revolutions that the encoder cannot tell after power has been
    down.

    #2. Lack of I/O:s available.
    - Communication between servo system and CNC-controller is simple I/O.
    No sophisticated protocol is needed because amplifier firmware is hacked
    directly to make 12 positions. Four lines are used to command position,
    another four indicates the same back to controller (as feedback?) when the
    positioning is attained. There are several I/O:s more to accomplish the
    whole thing. Simple and well documented but takes so many lines.... So the
    question is how we could interface a dozen of I/O lines to be operated
    with the servo amplifier (Argon...?).

    Ideas for the solution...

    -Absolute positioning could also be attained at start-up by driving the carrier to
    safe position and rotating the turret to find an index. How to make an index pulse
    system is another story, there are several, feasible, ways to do it anyway.
    No positioning memory, no absolute encoder, simple and robust approach.

    -If demand for numerous I/O:s cannot be met by servo amplifier itself so one
    possibility could be to use some microcontroller board as a front-end system
    for this. But would it then be natural to give all interfacing tasks (indexing
    also) to it... so servo amplifier just takes step and direction pulses in and
    returns signal to tell positioning being attained.

    - If this is the way to make an option to turret drives, and if this ever be
    realized... the next question is how general it is and how general it could be
    to serve every CNC-lathe owner. So you people out there, if you have knowledge
    of tool turret control systems, please comment.

  2. #2
    Join Date
    Sep 2004
    Posts
    1207
    Interesting retrofit plan! I also see that in the original machine there are some solutions that I would do differently from the beginning at least with today's tech.

    Absolute positioning can indeed be achieved by doing homing before using axis for indexing. Simplest way is to use hard stop detection which requires no sensors or switches. Drive simply drives gently at the end of mechanical travel and detects end position. VSD and Argon drives can be configured to do that automatically after power on. That method is also more accurate than typical mechanical home switch. If index pulse is available, it can be additionally searched to achieve highest possible homing accuracy.

    If we count all usable inputs of Argon, we have 8 digital inputs that may be read from user programmable firmware. However, some of them should be used for enable & reset fault signals. One quick'n'dirty option is to make resistor ladder to convert multiple digital signals to analog voltage and read voltage with one of the analog inputs:
    Resistor ladder - Wikipedia, the free encyclopedia

  3. #3
    Join Date
    Apr 2013
    Posts
    8
    @Xerxes, about hard stop detection... it seems not only solve
    positioning reference problem but it is present already!.
    The turret lock...

    I had yet another discussion with the owner of that
    particular lathe. He suggested using tool position no:1 as a
    general reference point. The idea is to always leave turret to that
    position at shut-down and to "command" referencing at power-up.
    Commanding reference simply means pressing a button to aknowledge
    servo system that the turret is in reference. We assume now that the
    turret lock is (and has been) engaged during power down. If, for some
    reason, this was not the case and the turret might have lost its position,
    the operator shall manually drive it into its correct position according
    visual markings. (Yes, he explained this to be well acceptable. Also in
    original system if repositioning is needed (flat back-up battery etc.)
    this is accomplished by applying external force to turret.) Here we must
    remember that minor unaccuracy in positioning is adjusted (=forced) by
    locking device anyway. There is a friction clutch to allow for remaining
    movement without mechanical damage.

    To estimate accuracy needed, as an exampel, I found that
    a parameter to define limits to an acceptance of positioning was 400 steps.
    (one revolution of servo motor is 80000 steps respectively) Within these
    limits servo amplifier is allowed to signal CNC-controller of succesful
    positioning. Locking of turret is then carried out. It seemed to me that
    during a split second needed for engagement the positioning was further
    improved to few steps.

    About resistor ladder as an input....
    When it comes to inputs only, yes, it it obvious that this solves
    the problem. Without further combinatorial evaluation we can say that
    if one analog input receives rotational information, then only few more
    (arbitrary) input signal (and lines) combinations are needed.

    Some output signals are needed also. Is there information
    available about Argons interface... and how to use its I/O lines?

    If in here or in some other cases a vast amount of I/O:s
    are needed then maybe we could approach this like some Raspberry
    people has...

    <http://www.element14.com/community/thread/18681>

    using I/O expanders MCP23017/MCP23S17 by Microchip...

    <http://www.microchip.com/wwwproducts/Devices.aspx?dDocName=en023500>
    <http://www.microchip.com/wwwproducts/Devices.aspx?dDocName=en023499>

    To save output lines I believe we could "send back" rotational
    bit combination to satisfy CNC-controller. Perhaps I missed something
    but I cannot understand why it needs that same information it just send
    out. Or maybe it really doesn't. But this is what wiring diagram tells..?

  4. #4
    Join Date
    Dec 2003
    Posts
    24216
    Are you committed to using the present command & protocol as it is now from the Fanuc?
    Al.
    CNC, Mechatronics Integration and Custom Machine Design

    “Logic will get you from A to B. Imagination will take you everywhere.”
    Albert E.

  5. #5
    Join Date
    Apr 2013
    Posts
    8
    @Al_The_Man. Yes, the idea behind is not to "confuse" Fanuc, or any other CNC-controller.
    Rather keep this limited and clearly defined retrofit to tool turret only. And, at least
    for the beginning, to only those machines where there is a hard wired, simple I/O
    communication to/from servo amplifier.
    Certainly it would be interesting to retrofit any servosystem in any machine
    if needed but I understood that typical communication interface is via serial protocol
    which we do not know in detail. If we did, we could do anything.
    Retrofitting the whole machine system (by LinuxCNC or something) is another, interesting,
    story but I cannot see options between these extremes.
    Now that we have (at least one to my knowing) an open firmware servo amplifier
    available it would be great to collect information about command & protocol used by
    CNC-controllers. Worth to its own (sub)forum somewhere here...?. People out there, do you
    see it feasible? Please comment.

    Mjerica

  6. #6
    Join Date
    Sep 2004
    Posts
    1207
    Hi Mjerica & co. Sorry for the long absence from this forum. Anway, now I'm back!

    Thanks for descriptions! I now understand the goal better. It certainly would work that drive is powered on and off at the correct position. Some drifting may happen as these positions won't be exactly same on each cycle but could be compensated manually.

    About I/O's: it certainly would be possible to add more IO to Argon drive. In the drive there is a expansion connector that has user programmable UART which could be used to connect to an IO expander.

    One solution is to control the drive over SimpleMotion bus from the beginning, so no direct IO's on drive needed. One may control drive from RS485 based SimpleMotion bus that may be driven for example by MCU or Raspberry. Another way is to use MCU to generate rotation by sending step/direction pulses based on which position is being requested by Fanuc. The latter may be easier to do.

    The source code of Argon ARM CPU and SimpleMotion lib are currently under finalization/clean up and soon available online.

Similar Threads

  1. Replacing Fanuc 6M Servo Drives
    By barchdesigns in forum Servo Motors / Drives
    Replies: 9
    Last Post: 09-29-2012, 07:41 PM
  2. Replacing Fanuc 6M Servo Drives
    By barchdesigns in forum Stepper Motors / Drives
    Replies: 5
    Last Post: 09-25-2012, 12:52 PM
  3. Need Help: Replacing Stepper with DC Brush Servo Motor
    By Marty_Escarcega in forum Servo Motors / Drives
    Replies: 13
    Last Post: 09-10-2012, 03:56 PM
  4. Replacing box mount receptacle connector on servo motor
    By gandp in forum Servo Motors / Drives
    Replies: 2
    Last Post: 02-27-2011, 08:34 PM
  5. Tool Turret Fault Alarm on TL-1 with no Turret?
    By Edster in forum Haas Lathes
    Replies: 10
    Last Post: 02-08-2010, 03:35 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •