Overview: What the J15/J16 Signal Actually Is
On a Hypertherm Powermax 1000 the J15/J16 terminals present raw tip voltage — roughly 100-200 VDC during a cut. This is the same potential you can measure on any plasma supply between the electrode lead and the work clamp (ground). Hypertherm brings it out on a defined machine-interface connector together with Torch Switch and Arc Good, so you do not have to tap the electrode lead or open the torch lead conduit.
Arc voltage is the primary feedback variable for torch height control (THC): standoff distance is approximately proportional to arc voltage at a fixed current and gas setting. Hold voltage constant with the Z axis and you hold standoff constant over warped plate.
Signal Characteristics and Safety Constraints
| Item | Value / Requirement |
|---|---|
| Source terminals | J15 / J16, machine interface |
| Signal | Raw tip voltage, electrode-to-work potential |
| Steady-state cutting range | 100-200 VDC (source-stated) |
| Open-circuit / pilot-arc transient | Higher than cutting voltage — verify the OCV figure in the Powermax 1000 operator manual before selecting component ratings |
| Other defined I/O on the same interface | Torch Switch (start command), Arc Good (arc-transfer confirmation) |
| Reference | Work clamp / workpiece potential, not the CNC chassis ground |
Designing the Voltage Divider
The divider does two jobs: bring 100-200 V into the input window of the THC electronics, and put a high series impedance between the plasma supply and the low-voltage side so a component failure cannot dump arc voltage into the controller.
Standard resistive divider:
Vout = Vin * R2 / (R1 + R2)
Ratio N = Vin / Vout = (R1 + R2) / R2
I_divider = Vin / (R1 + R2)
P_R1 = I_divider^2 * R1| Ratio | R1 (high side) | R2 (low side) | Vout @100 V | Vout @200 V | I @200 V | P in R1 @200 V |
|---|---|---|---|---|---|---|
| 50:1 | 4.9 MΩ | 100 kΩ | 2.00 V | 4.00 V | 40 µA | 7.8 mW |
| 20:1 | 1.9 MΩ | 100 kΩ | 5.00 V | 10.00 V | 100 µA | 19 mW |
| 40:1 | 3.9 MΩ | 100 kΩ | 2.50 V | 5.00 V | 50 µA | 9.8 mW |
Choose the ratio from the input range of the THC board you are feeding, not from a rule of thumb. A 0-10 V input board pairs naturally with 20:1 (200 V full scale); a 0-5 V input board pairs with 40:1.
Component selection rules
- Split the high side into a series string. Use five 1 MΩ or four 1 MΩ + 1 x 900 kΩ elements rather than a single resistor. Standard axial resistors are typically rated a few hundred volts of working voltage per element; a string distributes the arc voltage and the start transient across several bodies.
- Use 1% metal film, 100 ppm/°C or better. Divider tolerance is a direct scale error on the arc-voltage reading, which becomes a standoff error.
- Keep divider current low but not negligible. 40-100 µA keeps dissipation in the milliwatt range while staying far above the input bias current of any practical buffer stage.
- Buffer the tap. Feed the R2 node into a unity-gain op-amp or an isolation amplifier before it leaves the divider board. A 100 kΩ source impedance driving a long cable is an antenna for plasma-generated interference.
- Clamp the output. A low-capacitance TVS or Zener at the buffer input, sized just above the maximum expected divided voltage, protects the downstream board during start transients.
- Filter. A single-pole RC across R2 (for example 100 kΩ with a 0.1 µF film capacitor, fc ≈ 16 Hz) removes arc noise. Verify the corner frequency is still fast enough for your Z-axis correction rate; too much filtering makes the THC lag the plate.
Getting the Signal Into Mach3
Because Mach3 accepts no analog input, the height loop must close outside the PC. Three practical topologies:
| Topology | How Mach3 sees it | Notes |
|---|---|---|
| External THC board with UP/DOWN contacts | Two digital inputs mapped to THC Up and THC Down in Ports & Pins | Board compares divided arc voltage against a setpoint and pulses the direction lines; Mach3 nudges Z. Simplest wiring. |
| External THC board injecting step/dir into the Z channel | Nothing — the Z position is corrected downstream of the controller | Mach3's internal Z DRO drifts from true machine position; handle with a homing/re-zero routine after each cut. |
| Divider board with a serial/digital link to the motion software | Digital port data, not an analog level | Requires software that accepts the specific link; confirm compatibility before buying hardware. |
Interlock signals you should also wire
- Arc Good → digital input. Gate the entire THC loop on it. Without arc transfer confirmation, a THC reading 0 V will drive the torch down into the plate.
- Torch Switch → relay output from the CNC (M3/M5 or dedicated output). Use a dry contact; do not drive it from a logic pin directly.
- Anti-dive / delay: hold THC disabled for a configured delay after arc-good and again during corner slowdown, where reduced feed raises arc voltage and would otherwise command the torch upward.
Commissioning and Verification
- With the plasma OFF, verify divider continuity and ratio: inject a known DC voltage (a 24 V supply is convenient) across the divider input and measure the tap. At 20:1, 24 V in should read 1.20 V ±2%.
- Confirm the low-voltage side is fully isolated from the high side path to earth — ring out buffer output to J15 with a megohmmeter or DMM on the highest range, then verify no continuity path exists that bypasses the resistor string.
- Make a test cut on scrap. With THC in monitor-only mode, log the divided voltage and simultaneously measure electrode-to-work with a DMM set to a range above 200 VDC. Divided reading × ratio should match the DMM reading within a few percent.
- Record the arc voltage at your known-good manual standoff for the consumable, material thickness, and current in use. That number is the THC setpoint — there is no universal value; it is consumable- and current-specific.
- Enable closed-loop control at reduced Z speed. Watch for oscillation: reduce loop gain or increase the RC filter until the Z motion is a slow correction, not a hunt.
- Verify the abort path: break the arc mid-cut and confirm Arc Good drops and THC releases the Z axis without a dive.
Troubleshooting Table
| Symptom | Likely cause | Check |
|---|---|---|
| Torch dives into plate at arc start | THC enabled before arc transfer; no Arc Good gating | Confirm Arc Good input state and the post-transfer delay timer |
| Torch rises steadily during cut | Rising arc voltage from worn electrode/nozzle, or setpoint captured with fresh consumables | Replace consumables and re-capture the setpoint voltage |
| Z axis hunts continuously | Loop gain too high or insufficient filtering of a noisy divider tap | Add/verify the RC filter; reduce correction rate; check buffer is present |
| Reading is unstable or reads zero | Unbuffered high-impedance tap, or DMM/board input loading the divider | Measure at the buffer output, not the resistor node |
| Torch climbs at corners | Feed reduction raises arc voltage | Enable anti-dive/velocity-based THC lockout below a threshold feed |
| Breakout board damage after first cut | Arc voltage reached logic-level I/O | Verify series string integrity; add clamping; never route J15/J16 to the parallel port |
Design Checklist
- Divider ratio matched to the THC board's input window, not chosen arbitrarily.
- High side built as a series string of metal-film resistors, 1% tolerance.
- Buffered, filtered, and clamped output before any cable run.
- Divider board mounted in a grounded metal enclosure, close to the plasma supply; run the low-voltage output as shielded twisted pair with the shield grounded at one end only.
- Arc Good used as a hard enable for the height loop.
- Torch Switch driven through a relay contact, never a logic output.
- Setpoint documented per consumable set, current, gas pressure, and material thickness.
What voltage appears on Powermax 1000 terminals J15 and J16?
Raw tip voltage, approximately 100-200 VDC while cutting. It is the same electrode-to-work potential you would measure between the electrode lead and the work clamp on any plasma supply. Open-circuit and start transients exceed the cutting range, so verify OCV in the operator manual before rating components.
Can Mach3 read plasma arc voltage directly for torch height control?
No. Mach3 has no analog input support. You need external hardware that divides the arc voltage, compares it to a setpoint, and outputs either discrete THC Up/Down contacts mapped to Mach3 digital inputs or a step/direction correction injected into the Z axis.
What divider ratio should I use for arc voltage sensing?
Match it to the THC input range. A 0-10 V input needs 20:1 (R1 = 1.9 MΩ, R2 = 100 kΩ), giving 5.0 V at 100 V arc and 10.0 V at 200 V. A 0-5 V input needs 40:1. Divider current stays at 40-100 µA, so dissipation is under 20 mW.
Why does my torch dive into the plate when the arc starts?
The height loop is running before arc transfer, so it sees 0 V and commands maximum down travel. Gate THC on the Arc Good signal from the plasma machine interface and add a post-transfer delay before enabling closed-loop correction.
Why does arc voltage drift upward during a long cut?
Electrode and nozzle wear raise arc voltage at the same standoff. A THC holding a fixed voltage setpoint will steadily lift the torch. Re-capture the setpoint with fresh consumables and treat a persistent upward trend as a consumable-change indicator.
