The SX3 CNC spindle-control upgrade can provide computer-controlled start, stop, direction, and speed, but the evidence shows that compatibility depends on the mill’s board revision, supply-voltage version, and command interface. Commission the control in stages: verify the upgrade with a direct analog command first, then test the breakout board or motion controller separately.
Identify the Hardware Variant Before Wiring
Do not assume that every SX3 uses the same spindle-driver PCB. One 230 V machine purchased in May 2014 accepted the upgrade connector mechanically, although the fit was imperfect, but its start/stop conductor terminated at an unconnected PCB location. Another owner reported that machine-specific differences prevented the kit from working.
| Evidence | Engineering decision |
|---|---|
| Separate 110 V and 230 V installations were reported | Match the kit transformer and wiring to the machine supply before energizing. |
| A 110 V transformer was connected to 240 V and failed | Never infer transformer compatibility from connector fit. |
| At least one connector did not fit perfectly | Confirm pin alignment, conductor destinations, and seating without forcing the plug. |
| SX3 board differences were reported | Compare the installed driver PCB with the kit documentation instead of approving compatibility from the SX3 model name alone. |
Diagnose a Spindle That Will Not Stop
On one 230 V installation, the spindle ran and reversed correctly but ignored the start/stop input whether or not that input was connected to GND. Tracing the red start/stop conductor showed a signal changing from approximately -0.7 V to 4 V relative to a reference that might not have been true ground. The signal did not reach a pin on the ATMEGA device because the corresponding driver-board connection was open from the factory.
- De-energize the machine and verify absence of hazardous voltage before inspecting connectors or removing the driver PCB.
- Confirm that the start/stop conductor reaches the intended destination on the installed board; do not treat connector engagement as proof of electrical continuity.
- With suitable test equipment and a defined reference, check whether the start/stop state changes at the kit output and whether that change reaches the driver circuitry.
- If the signal exists at the kit but disappears across an open PCB location, document the board revision and obtain an approved repair decision before modifying the board.
That installation was made operational by soldering a 100 ohm resistor across the identified gap. Treat this only as a diagnosis from one 230 V board, not as a universal SX3 modification. Applying it to a different revision without tracing the circuit could bypass an intentional isolation or protection feature.
A separate installation stopped the spindle by reducing the speed-control voltage to zero; its start/stop input did not have to be grounded. This is a functional workaround for that configuration, but it provides less independent stop-state checking than a working discrete start/stop channel.
Separate the Upgrade Board from the Command Source
Reported speed errors were not consistently caused by the upgrade board. The board produced a linear response when driven from an external variable DC supply, while one G540 installation produced a nonlinear analog output and oscillated at some commands. Another G540 installation correlated five Mach3 commands—300, 500, 700, 1000, and 1700—with measured speed within approximately ±10 to 30 RPM.
- Disconnect the motion-controller speed output and apply a controlled DC command directly to the upgrade board.
- Measure command voltage and spindle speed across several points in the required operating range.
- If direct-voltage tracking is acceptable, reconnect the controller and repeat the same measurements at its analog output.
- If voltage or speed becomes nonlinear only after reconnection, troubleshoot PWM scaling, analog conversion, grounding, or the interface board rather than replacing the spindle-control kit.
One machine produced 1710 RPM at a 5 V command. Its factory display indicated 1800 RPM while an external tachometer measured 1710 RPM, although readings from 100 through 1700 RPM were otherwise described as close. Use an independent tachometer during calibration; do not calibrate solely from the machine display.
Handle Raw PWM as a Separate Interface Case
One LinuxCNC installation applied 100 Hz PWM and experienced oscillation at low speed commands. The same upgrade board behaved linearly from a DC power supply, isolating the problem to the raw PWM interface rather than the board’s steady analog response.
A low-pass filter eliminated the oscillation in that installation. During a recorded test, a 10 V PWM signal at 10% duty cycle produced approximately 1 V across a 1 kΩ dummy load after filtering. The response took about 0.5 seconds to rise to 1 V and retained minor ripple. A 12 V supply was used to obtain the full 10 V output.
The evidence does not provide the complete filter component values, optocoupler driver, output impedance, or input specification. Do not reproduce a filter from the timing and load data alone. Design or select the filter from the actual PWM frequency, acceptable ripple, required response time, and measured input impedance.
Commission and Verify the Complete Control Chain
- Confirm the machine voltage, kit voltage version, transformer rating, connector orientation, and driver-board revision before applying power.
- Use the kit’s simplest documented configuration to verify spindle rotation and direction independently of CNC software.
- Verify discrete start/stop operation. If it fails, trace continuity instead of assuming that software configuration is responsible.
- Drive the speed input from a controlled DC source and record command voltage against tachometer speed.
- Connect the breakout board or motion controller and repeat the voltage and speed measurements at the same command points.
- Test startup, software closure, controller disable, charge-pump loss, and emergency-stop states. Reports included unintended spindle startup when the computer was on before Mach3 opened and one isolated startup while Mach was closed and a G540 indicated a disabled state.
Do not accept a configuration merely because normal software commands work. The verified safe state must persist during boot, application shutdown, disabled outputs, and loss of the supervisory signal. Where the speed command itself controls running, evaluate whether a separate hardware stop path is required by the machine risk assessment.
FAQ
Why does an SX3 spindle keep running when start/stop is grounded?
On one 230 V SX3 purchased in May 2014, the start/stop connector pin ended at an open location on the driver PCB, so the signal never reached the ATMEGA circuitry. Trace the signal and board continuity before changing software or copying the reported 100 ohm repair.
Does the SX3 spindle-control board accept raw PWM?
A 100 Hz LinuxCNC PWM signal caused low-speed oscillation in one installation, while direct DC control was linear. A low-pass filter removed the oscillation, but the available evidence does not provide enough component data to prescribe that filter universally.
How do I calibrate SX3 commanded speed against actual RPM?
Apply a controlled DC command, measure the voltage at the board, and verify spindle speed with an independent tachometer before reconnecting the CNC interface. One machine produced 1710 RPM at 5 V even though its factory display indicated 1800 RPM.