Kitamura Mycenter 0 Yasnac MX3: Phase Power, 4th Axis, Retrofits

Tom Garrett17 min read
Motion ControlTechnical ReferenceYaskawa
Licensed PE Working through this on a live machine? A Maine-licensed engineer can take it from here — included with IMD hardware, by the hour for everything else. Book an engineer

Overview

The Kitamura Mycenter 0 is a compact vertical machining center originally shipped with the Yaskawa Yasnac MX3 control. It is mechanically simple relative to its larger Mycenter siblings: the machine does not require a separate hydraulic power unit, which makes it an attractive candidate for low-volume shops, prototyping cells, and home/hobby CNC installations where three-phase utility service may not be available. When paired with a static or rotary phase converter, the Mycenter 0 can be brought up on derived three-phase power with only modest electrical rework.

This reference consolidates the field experience with the Yasnac MX3 on the Mycenter 0 platform: the wiring conventions for a wild-leg phase converter, the procedure for adding a 4th-axis rotary table, the RS-232-based file transfer path used for program and parameter backup, the supported memory upgrades, and the most common MX3 board-level failure modes. The Mycenter 0 is a legacy machine; the OEM has long since discontinued new-build support, and most parts pass through industrial surplus channels rather than authorized distributors.

Power Topology and Single-Phase Operation

The Mycenter 0 lacks a dedicated hydraulic unit. The spindle, axis servos, tool changer, and way-lube pump are all driven either by direct AC motors or by integrated lubrication/ATC mechanisms. The practical consequence is that the machine has no large three-phase inductive load that requires careful phase-angle balancing beyond the spindle motor. This is what allows a sharp technician to bring the machine up on a single-phase 240 V supply, or on a derived three-phase source from a phase converter, without the full reconfiguration a hydraulic CNC would demand.

Before connecting any phase converter, identify the machine's main disconnect and confirm the nameplate voltage. Most Mycenter 0 units sold in North America are 200/220 V three-phase, 60 Hz. The control cabinet will have a clearly labeled three-phase input terminal block (typically L1, L2, L3) and a protective earth stud. The wild-leg (sometimes called the "stinger" or "generated" leg) of a phase converter must land on a specific phase, and Yaskawa controls in this era are no exception: the spindle drive and axis servo amplifiers expect a fixed phase rotation, and the wild leg must not be assigned to a servo input terminal at random.

Phase Converter Wiring With a Wild Leg

A static phase converter (SPC) or rotary phase converter (RPC) generates a third leg by either capacitor-derived phase shift (SPC, no rotating mass) or by a motor-generator set (RPC). The generated leg is at a different phase angle than the two utility legs and is typically within 5-10 degrees of 120 degrees on a rotary, or further off on a static. Yaskawa spindle drives from the MX3 era are tolerant of the small phase-angle deviation produced by a rotary converter, but the rule of thumb is: never land the wild leg on L1 of the CNC's main input if the spindle drive is the largest single load. The wild leg should drive the lowest-impedance branch of the system relative to phase rotation.

Recommended wiring convention for the Mycenter 0 on a phase converter:

  1. Verify the rotary phase converter is sized for at least 1.5x the running load of the machine; a 10 HP RPC is a practical minimum for a Mycenter 0 with a 5-7.5 HP spindle motor, accounting for inrush during spindle accel/decel.
  2. Bond the phase converter chassis and the CNC cabinet to a single earth ground with a #6 AWG or larger conductor, routed separately from the neutral.
  3. Connect the two utility legs (L1, L2) from the converter output to the corresponding terminals at the CNC disconnect. Connect the generated (wild) leg to L3 at the CNC disconnect.
  4. Check phase rotation at the CNC input with a phase sequence indicator. The expected rotation is ABC. If reversed, swap the two utility legs at the converter output; do not move the wild leg.
  5. Power the control with the door interlocks open and a multimeter on the input side. Confirm leg-to-leg voltage is within 5% of the nameplate (e.g., 220-230 V between L1-L2, L2-L3, L3-L1) and that no leg drops more than 8% under the spindle's no-load spin-up.
  6. If a voltage imbalance greater than 5% is observed, the phase converter is undersized or the run capacitors (on a rotary) need adjustment. Rebalance before proceeding.
If the machine's main contactor or any input filter module is documented as sensitive to phase sequence, the wild leg must always be terminated on the same conductor that the factory schematic calls L3. Reversing utility legs on an RPC can present a 240 V single-phase condition across the coil of a three-phase contactor if the wild leg is left floating during a transition; never interrupt a single utility leg while the converter is loaded.

Yasnac MX3 Control Architecture

The Yasnac MX3 is a Yaskawa CNC in the MX series (predecessor: MX1, successor: i80/MXi series). The control is built around a multi-board backplane with dedicated function boards: a CPU/NC board, a servo interface board, a digital I/O board, an operator panel interface, and a power supply board. The architecture is modular in the sense that a single failed board does not require control replacement, but in practice, the closed documentation and obsolete parts inventory make board-level diagnosis the bottleneck rather than the repair itself.

Board Function Common Failure Indicator
Power supply board +5 V, ±12 V, +24 V DC distribution Control will not power up; relays click repeatedly; display backlight fails to initialize
CPU/NC board Interpreter, part program storage, M-code handling Post displays or alarm 7xx series; no response to MDI
Servo interface board PWM command to spindle and axis drives Axis runaway on power-up; SV alarm class
Digital I/O board Limit switches, ATC solenoids, coolant, lube M-functions do not execute; spindle does not start despite spindle command
Operator panel board Display, keys, mode select Segmented LCD row failures; key chatter

The first thing a technician should do on a newly acquired Mycenter 0 is perform a complete parameter and part-program backup. The MX3 stores NC parameters in a battery-backed SRAM that is vulnerable to lithium-thionyl chloride cell depletion (typical cell life 5-8 years). When the battery dies, all G54-G59 work offsets, NC parameters, and pitch-error compensation tables can be lost. The backup path is RS-232 only on most MX3 units, and the procedure is described in a later section.

Fixture Offsets: G54-G59 Procedure (Known Quirk)

Setting work offsets on the Yasnac MX3 is a manual procedure. Unlike later controls that accept edge-finder or probe-driven offset values with a single keystroke, the MX3 requires the operator to type in the six-digit coordinate of each axis (X, Y, Z) by hand for each fixture offset register. The format is signed, in the active programming units (typically microns in metric mode, ten-thousandths in inch mode), with leading zeros.

Worked example for a typical mill orientation, inch mode, X at -12.4500, Y at -7.6250, Z at 0.0000:


G54 X-12.4500 Y-7.6250 Z0.0000

The six-digit, all-axes-every-time requirement is the most-cited workflow friction on the MX3. When CAM post-processors generate G54-G59 with high precision (e.g., -12.450123), operators often have to round to the MX3's stored precision. Document the offset values in a hard-copy log taped to the inside of the electrical cabinet door, because if the SRAM battery dies, the in-control values will revert to zero on next power-up.

Replace the SRAM backup battery on a fixed schedule, not on failure. Use a Yaskawa-specified cell (typically a 3.6 V lithium, 2/3 AA form factor) and a clip that matches the original. Reverse-polarity installation will destroy the SRAM in seconds.

G41/G42 Cutter Compensation: Quadrant-Sensitive Behavior

Cutter compensation on the MX3 is functional but exhibits a documented sensitivity to the approach quadrant. When activating G41 (left-of-path) or G42 (right-of-path), the required lead-in move length and direction depend on the direction from which the tool approaches the first compensated move. The compensation vector is applied at the lead-in move, and if the lead-in is too short or oriented in a way that the compensation vector is not yet fully resolved, the first cut can be over- or undersized.

Field-proven conventions for the MX3:

  • Use a lead-in length of at least 1.5x the cutter diameter before the first G1 move that engages material.
  • Avoid activating compensation within a G02/G03 arc. Always activate on a G1 lead-in.
  • In profile milling, choose the approach quadrant such that the compensation vector resolves along a tangent to the path. For external (G41) profiling, approach the start point so the lead-in is roughly tangent to the profile's starting direction.
  • When in doubt, dry-run the program with the tool raised above the workpiece and single-block through the first compensation block. Verify the computed offset position before lowering the tool.

This quadrant behavior is less of a concern when programs are generated by CAM, because the post-processor inserts an explicit lead-in move before the first compensated block. Hand-written programs are where this quirk shows up most often.

4th Axis Compatibility and Installation

The Mycenter 0 was ordered from the factory with a 4th-axis option. The factory-installed units used a 5-inch (130 mm) faceplate rotary table with a Yaskawa AC servo and a pneumatic collet closer. Units that did not ship with the option can still be retrofitted, but the upgrade is non-trivial. The retrofit requires:

  1. The mechanical rotary table itself (Trum, Tsudakoma, or Yaskawa-branded units were common); mounting hardware matched to the Mycenter 0 table T-slot pattern; and a tailstock if long parts will be run.
  2. A fourth axis servo amplifier card (or external drive, depending on the production year of the control) installed in the cabinet, with the matching M-code and M-function finishing relays added to the I/O board.
  3. Connection of the rotary encoder feedback to the servo interface board, including wiring the encoder power, A/B quadrature, and marker (Z) channel through the existing cable conduit.
  4. NC parameter updates to enable the fourth axis: pitch-error compensation, in-position window, following-error limits, and the interlock that prevents simultaneous 3rd and 4th axis motion beyond machine envelope.
  5. PLC ladder updates to handle the 4th-axis clamp/unclamp, the rotary table index-complete signal, and the tool-length offset referenced to the part centerline rather than the spindle face.
  6. Firmware verification: some MX3 units did not have 4th-axis firmware enabled at the factory. If a parameter write returns a non-acknowledgment or the new axis is not recognized in the parameter screen, a firmware PROM swap is required. The firmware is generally NLA (No Longer Available) from Yaskawa for end-user purchase; it is sometimes sourced from parted-out machines on the surplus market.
Without 4th-axis firmware, no amount of amplifier or parameter work will make the control recognize the additional axis. Verify firmware revision before committing to the mechanical installation; a surplus Yaskawa MX3 CPU board sourced from a 4th-axis-equipped machine is the most reliable path.

The PLC ladder modifications for the 4th axis are not user-editable through the standard MX3 operator interface in most shipped versions. They are written through a Yaskawa service tool (typically a Yaskawa PLC programming pendant or a PC running legacy Yaskawa PLC software over RS-232). End-user access to the PLC source is restricted; the service technician must have the original ladder printed or the OEM's encrypted disk image.

File Transfer: USB to RS-232 on the MX3

The Yasnac MX3 does not have a native USB port. It uses a 25-pin female D-sub on the back of the operator panel for RS-232C serial communication at 4800 to 19200 baud, with XON/XOFF or hardware flow control. Part programs and NC parameters are sent and received in ASCII, with a Yaskawa-specific header and terminator that is documented in the Yasnac MX3 operator's manual.

Modern shop PCs rarely have a true RS-232 port, so the common path is a USB-to-RS-232 DB9 adapter, then a DB9-to-DB25 null-modem cable (or a direct USB-to-DB25 adapter rated for the MX3's signaling levels). Compatible USB-to-serial chipsets include FTDI FT232 and the SiLabs CP2102; both are widely available in industrial-grade isolator housings. Avoid unbranded adapters based on the Prolific PL2303 chipset, which has known driver-stability issues on modern Windows versions.

Recommended transfer procedure:

  1. Set the MX3 I/O channel to RS-232 (parameter selection in the operator menu under I/O or by the dedicated channel-select softkey on later MX3 firmware).
  2. Set the baud rate to match the sending PC: 4800, 9600, or 19200 are the practical choices. 19200 is reliable on FTDI-based adapters with shielded cabling under 3 m.
  3. From the receiving PC, open a terminal in Yaskawa drip-feed mode or use a Yaskawa-aware DNC software (e.g., CIMCO Edit with the Yaskawa post, or a generic ASCII drip-feed tool configured for XON/XOFF). The MX3 expects the file in plain ASCII with a Yaskawa header.
  4. Trigger RECEIVE on the MX3 operator panel, then transmit from the PC. The display will show the incoming program name and byte count as it loads.
  5. When transferring NC parameters, use the dedicated parameter send/receive function on the MX3, not the part-program channel. Parameters are sent as a block-by-block ASCII stream and must be received with the control in a specific state (typically MDI or the parameter I/O mode).
Always do a full parameter receive and then power-cycle the control before running a part. The MX3 does not validate parameter ranges on receive; an out-of-range value can cause an SV alarm on the next axis motion attempt.

Memory and Storage Upgrades

The MX3 stores part programs on battery-backed SRAM (typically 64 KB to 256 KB, depending on the option board fitted at the factory) and NC parameters in a separate battery-backed parameter SRAM. The on-board SRAM is not user-upgradeable in the field in the conventional sense; the SRAM chip is soldered to the CPU/NC board and the battery is socketed. Practical upgrade paths:

  • Battery replacement: swap the lithium cell on a fixed schedule. The cell is the only consumable on the CPU/NC board.
  • Board-level memory upgrade: the most reliable approach is to source a higher-capacity CPU/NC board from a parted-out MX3 or MX4 control. Verify the firmware revision matches; the same firmware can run on a higher-capacity board, and the part-program storage scales accordingly.
  • External DNC drip-feed: the most common modern workaround. Run long programs from a connected PC in drip-feed mode rather than storing them in on-board SRAM. This is the only practical way to run programs larger than ~50 KB on a stock MX3.

For shops that must run programs larger than the on-board SRAM, drip-feed over RS-232 at 19200 baud is more than adequate for programs that fit comfortably in the operator's review window.

Known Board Failure Modes

The Mycenter 0 with the Yasnac MX3 has a handful of well-documented failure modes that, in the field, are more common than catastrophic control failure. Diagnosis is straightforward once the symptom is recognized.

Symptom Likely Board Verification Repair Path
Control will not power up; relays chatter Power supply board Measure +5 V at the backplane; if absent or sagging below 4.75 V, replace the supply Board-level recap (electrolytic age-out) or full board replacement from surplus
Axis drifts or runs away on enable Servo interface board Check encoder feedback with scope; verify with known-good amplifier Board swap from a known-good donor; check for cold solder joints at the backplane connector
One M-function does not execute; others work Digital I/O board Check the output transistor for the specific M-code; test with a known M-code pair Replace the output driver section; the affected M-code's relay driver IC is a known weak part
Operator panel keys do not register; LCD segments dead Operator panel board Visible segment failure; sticky keys after long storage Panel board replacement; membrane key repair is unreliable, full board swap is preferred
SRAM loses parameters on power-down Battery / CPU board Measure battery voltage; replace if below 3.0 V Restore from backup; replace battery; reload parameters
Persistent fuse blow on initial power-up Power supply input filter Disconnect downstream of supply; if fuse still blows, supply is shorted Supply board repair or replacement
End-user board-level repair is uncommon on the MX3. The schematics were never released for general distribution, and most repair shops in the secondary market will require the entire board to be sent in. Avoid the temptation to recap the power supply in place without a service manual: the supply includes a 5 V sense line that must be re-trimmed after recap, and a poorly trimmed supply will manifest as intermittent SV alarms weeks after the repair.

Service Documentation and Support

Yaskawa's official support for the MX series has been substantially reduced since the 2010s. Main-office engineering contact (historically the Chicago office) is the most reliable path for genuine engineering support; regional parts-and-service channels may push replacement-machine sales over parts supply for end users. The original machine documentation (operator's manual, parameter list, PLC ladder printout, wiring diagram) is the most valuable asset on a used Mycenter 0. If the machine is acquired without the documentation, contact the seller before parting out and request the full document set; missing documentation can reduce the machine's recoverable value by an order of magnitude relative to the same machine with a complete binder.

For board-level repair, the practical options are: a known-good donor board from a parted-out MX3, a Yaskawa-authorized drive-and-control repair house, or a CNC-specific repair shop that has worked on Yaskawa controls in the past. Avoid generic industrial electronic repair shops that quote a flat diagnostic fee and then return the board unrepaired; the MX3 documentation gap is a known reason these shops decline the work.

Commissioning Checklist

  1. Verify incoming voltage, phase rotation, and ground integrity at the main disconnect before energizing the cabinet.
  2. Replace the SRAM backup battery on a used machine as a matter of course, regardless of measured voltage.
  3. Perform a full NC parameter and PLC parameter receive from the existing control (if it boots) or load from a known-good backup.
  4. Verify pitch-error compensation tables for each axis. The Mycenter 0's ballscrews are not precision-ground and require compensation every 25-50 mm for production work.
  5. Test all M-codes from MDI: spindle start/stop, tool change, coolant, way lube, and (if equipped) 4th-axis clamp/unclamp.
  6. Run a warm-up program in air at 30%, 60%, and 100% rapid traverse to thermally stabilize the servos before any precision cut.
  7. Document every parameter and every wiring change made during commissioning. Tape a copy of the parameter printout inside the cabinet door.

FAQ

Can a Kitamura Mycenter 0 run on single-phase power without a phase converter?

Practically, no. The spindle motor and servo amplifiers are three-phase. The Mycenter 0's lack of a hydraulic unit makes it a relatively friendly conversion, but a static or rotary phase converter is required to derive the third leg. A 10 HP rotary is the typical minimum for a 5-7.5 HP spindle.

Where should the wild leg land on the Mycenter 0 main input?

The wild leg should land on the same L3 terminal that the factory schematic calls for. Always verify phase rotation with a phase sequence indicator and never swap the wild leg to fix a rotation error; instead, swap the two utility legs at the phase converter output.

Can I add a 4th axis to a Mycenter 0 that was not factory-equipped?

Yes, but the upgrade requires a 4th-axis servo amplifier, NC parameter changes, PLC ladder updates, and in most cases a firmware PROM swap. Firmware is generally NLA from Yaskawa for end users; sourcing a CPU/NC board from a parted-out 4th-axis MX3 is the most reliable path.

What is the most common cause of parameter loss on the Yasnac MX3?

The lithium backup battery on the CPU/NC board. When the cell voltage drops below approximately 3.0 V, the SRAM contents (G54-G59, NC parameters, pitch-error tables) are no longer retained across power cycles. Replace the battery on a fixed schedule, not on failure.

How do I back up part programs and parameters to a modern PC?

Use a USB-to-RS-232 adapter (FTDI FT232-based is recommended) and a null-modem cable to the MX3's 25-pin serial port. Set 19200 baud with XON/XOFF, trigger RECEIVE on the control, and transmit from a Yaskawa-aware DNC client. Parameters must be transferred through the parameter I/O function, not the part-program channel.

Why does the control blow its main fuse the moment power is applied?

The most common cause is a shorted power supply board, typically from aged electrolytic capacitors on the input or 5 V rail. Disconnect the supply from the backplane; if the fuse still blows, the supply is internally shorted and must be replaced or repaired. Avoid bypassing the fuse with a higher rating; the wiring downstream is sized for the original fuse's let-through energy.

Back to blog