How much separation should I leave between control and power cables?
Leave at least about 20 cm on parallel routes, cross at 90 degrees, and do not carry power and signal in one trunking; use trunking with a metal divider.
Leave at least about 20 cm between power cables and control and signal cables that run in parallel. Make any unavoidable crossings at a right angle of 90 degrees. Do not carry the two in the same cable trunking; if a single run is unavoidable, use trunking with a metal divider and ground the divider.
Why distance is so effective
The coupling between a power cable and a signal cable arises through two mechanisms: inductive coupling through the magnetic field and capacitive coupling through the electric field. Both weaken rapidly as the distance increases; that is why cable separation distance is the cheapest EMC measure. The size of the coupling is also proportional to the length over which the two cables run side by side: an approach of a few centimetres causes no trouble, whereas a route that runs in the same trunking all the way from the panel to the machine produces interference over that whole distance.
Rules for practice
- At least ≈20 cm on a parallel route. If the motor cable is fed from an inverter, increase this distance as far as you can.
- Cross at a right angle. At 90 degrees the coupling area is at its minimum.
- Separate trunking. Use separate cable trunking for power, control and signal. If a single run is unavoidable, choose trunking with a metal divider and ground the divider.
- Twisted pair. Carry the control signals in twisted pair; the twist cancels the effect of inductive coupling by reversing its sign at every twist.
- Short routes. Run the cables by the shortest path; avoid long routes parallel to large metal surfaces.
- Inside the panel. Keep the supply and output cables of the receiver away from the area of the inverter, the contactors and the braking resistor.
Which pair, how much separation
| Cable pair | Same trunking? | Practice |
|---|---|---|
| Inverter motor output – RF receiver signal | No | The greatest separation possible, preferably a different route |
| Contactor coil line – control signal | No | ≈20 cm; trunking with a divider is acceptable |
| Lighting or socket line – control signal | Yes in divided trunking | Keep ≈20 cm of separation |
| Short crossing points | — | 90 degree right angle, pass without running parallel |
When separation alone is not enough
- If the source is very strong, distance is not enough. With interference originating from an inverter, screen the inverter output cable first; terminate the screen properly to the motor body and to the inverter chassis.
- If you are using screened cable, which end you ground the screen at determines the result; connecting it at both ends can create an ground loop.
- If common mode noise remains, fit the ferrite ring at the right point; a ferrite fitted in the wrong place changes nothing.
- If the problem is seen only while the welding machine is running, the coupling may be arriving directly at the receiver housing rather than over the cable; moving the receiver gives a faster result.
The three most frequent mistakes on site are these: taking the signal cable of the receiver inside the same spiral conduit as the motor cable, running the loom that goes to the panel door right over the inverter, and lining the power and signal ends up next to each other in the terminal row. All three comply with the distance rule on paper and not in practice. The diagnosis is simple: stop the inverter and try the remote control from the same spot. If the complaint disappears the source is the inverter; correct the route first and think about a filter afterwards.
Making this separation at the installation stage is both cheaper and more lasting than trying to make up for it with a filter and a ferrite later. On a new panel, draw the route in advance and separate power from signal from the outset. On an existing panel, try a single step first: take the signal and supply cable of the receiver out of the inverter area. Most complaints are closed by that one change.
The full topic: 433 MHz or 2.4 GHz?