Configuring Mach3 Output #3 for M8 Plasma Torch Relay

What the Screen Is Telling You

Two observations narrow this to configuration, not a hardware fault. The relay changes state the instant Mach3 takes the parallel port, so the path from DB25 pin 14 to the board's relay driver is intact and the port is being written. And the Output Signals table shows pin 14 sitting on the Output #1 row while Output #3 carries no port, no pin, and no enable flag. M8 drives Output #3. Output #3 goes nowhere.

The tag is right; the binding is wrong. An M-code reaches a connector pin through two independent configuration stages in Mach3, and both have to agree:

Why Does M8 Not Reach Pin 14?

M8 is not bonded to a pin number. It asserts a logical output number set on the Spindle Setup tab, and that number then has to be resolved to a physical pin by an enabled row in Output Signals. Mapping pin 14 onto a row that no M-code addresses, or addressing a row that has no pin, breaks the chain in the middle. That is exactly the state in the screenshot: pin 14 on Output #1, Output #3 empty.

Three gates upstream of the pin will also swallow the command silently. Disable Flood/Mist Relays blocks M8 and M9 regardless of any pin assignment; Disable Spindle Relays does the same for M3 and M4. Mach3 holds the spindle and coolant outputs deasserted while Reset is flashing, so an M8 typed into the MDI line with the machine in E-Stop produces nothing. And two enabled output rows pointed at the same pin fight each other — one deasserts what the other sets. Check that the Charge Pump row is not also parked on pin 14.

M8 on Output #3 or M3 on Output #1?

Both mappings drive the same relay. The criterion is what your CAM post emits and how you want the torch to turn off.

Use M3/M5 for arc-on if you control the post. The torch then falls off with the same logic that stops the spindle — program end, Stop button, Reset — instead of depending on the operator or the post remembering an M9. Keep M8 for the air solenoid. If the post is fixed and emits M8, take the M8 route and configure Output #3 properly; there is no performance difference at the pin.

Configuring the Output

1. Open Config > Ports and Pins > Port Setup. Confirm Port #1 is enabled and its base address matches the LPT resource shown in Device Manager (commonly ). Set Kernel Speed to 25000 Hz to start.
2. Go to the Output Signals tab. Enable exactly one row for the torch — Output #3 for the M8 route, Output #1 for the M3 route — with Port # = 1 and Pin Number = 14. Leave Active Low unchecked for now.
3. Set every other row that references pin 14 to disabled, including Charge Pump if the board does not use it. One pin, one enabled signal.
4. Open the Spindle Setup tab. For M8: uncheck Disable Flood/Mist Relays and set the Flood M8 output number to 3. For M3: uncheck Disable Spindle Relays and set the Clockwise (M3) output number to 1.
5. Apply, OK, then close and restart Mach3 so the driver reloads the pin map.
6. Clear Reset. Outputs stay deasserted while the E-Stop condition is active.
7. Type M8 (or M3 S1) in the MDI line and watch the corresponding LED on the Diagnostics screen together with the relay.

Why Is the Relay Closed With the PC Powered Down?

Pin 14 is Auto Feed, one of the four control-register lines on the DB25 (pins 1, 14, 16, 17). With the PC off, that line is not driven; the board's input sees its unpowered rest level and the relay driver follows it. As soon as Mach3 initializes the port, the line is driven to its deasserted level and the relay flips. That is the polarity mismatch talking, and it is fixed with the Active Low checkbox on the output row, not with a wiring change.

Before touching that checkbox, settle which contact you are wired to. A relay reading closed with the PC off may simply be de-energized on its NC contact — in which case Mach3 has been energizing the coil at idle all along, the arc-start command adds nothing visible, and the fix is to move to the NO contact or invert Active Low. Listen for the coil click while toggling M8/M9; a clicking relay with no change at the load contact means the wrong contact.

Either way, an arc-start relay that closes whenever the PC is off or Windows is booting is not an acceptable idle state on a plasma table. Break the torch trigger circuit through the E-Stop chain or a charge-pump-driven enable relay so the arc cannot be commanded while the port is unpowered or Mach3 sits in Reset. Route the trigger pair away from the parallel cable and keep the table grounded — high-frequency arc start couples into an unshielded LPT run and will lock the port mid-cut.

Verifying at the DB25

Put a meter on DB25 pin 14 referenced to a ground pin (18–25) with the torch lead disconnected from the plasma unit, and step through the states:

Branch on what you measure. No level change between idle and asserted means the software path is still broken — recheck the output number in Spindle Setup, the Enabled flag on the output row, and the port base address. Level changes but the relay does not move means the fault is board-side: verify the relay coil supply, any enable or charge-pump jumper, and the board manual's contact rating before you hang the torch trigger on it.

Close it out with a dry run: load the cut file with the plasma unit's torch lead unplugged, and confirm the relay closes at the pierce line and drops at the end of every cut segment before you strike a real arc.

FAQ