A 2 tonne underhung crane draws 7.5 A and still shuts down workshops, because the runway is the roof beam, the earth path runs through paint, and the building's RCD was already sitting at 21 mA of leakage. Here is what actually goes wrong, in the order it goes wrong.
A 2 tonne suspension crane draws about 7.5 A. That is less than a workshop air conditioner. It still shuts down production lines, and the reason is almost never the crane itself. A suspension crane hangs from the building, feeds from the building and earths through the building, so every electrical fault it has turns into a building fault.
I spent two days in September on a 1 tonne underhung unit in a packaging plant. The complaint was "the crane trips the lights". The crane was fine. The 30 mA type A RCD in the lighting board was already sitting at 21 mA of standing leakage from a roof full of LED drivers, and the crane's brake rectifier pushed it over the edge twice a shift.
That is the class of problem this article is about. Not harmonic distortion on a 6 MW steel mill busbar. The small stuff that gets blamed on the crane and then paid for by the crane supplier under warranty.
Numbers below come from three places. Published standards clauses, published product data with the source named, and arithmetic we did ourselves. Computed numbers are labelled planning figures. Nobody has published a survey of electrical faults on 5 tonne underhung cranes, so I am not going to invent one.
Because of where the electrical boundary sits. IEC 60204-32:2023 covers the electrical equipment of hoisting machines, and it begins at the crane supply switch. Read the scope and it explicitly includes the "power supply and control feeders situated outside of the hoisting machine, for example, flexible cables or conductor wires or conductor bars". On a gantry crane that clause is about a short drop cable. On a suspension crane it can be 40 m of festoon running the length of a roof.
So the standard puts the feeder inside the machine's electrical scope, but the contract usually does not. The crane maker supplies up to its own isolator. The building contractor runs a cable to somewhere near the roof. In between sits the gap where earthing gets improvised and RCD coordination gets forgotten.
| Item | Usually crane maker | Usually building or installer | Who tests it |
|---|---|---|---|
| Hoist, trolley, panel, control | Yes | No | Maker, at works |
| Crane supply switch, lockable | Yes, mounted on the crane or runway | Sometimes | Installer, on site |
| Festoon or conductor bar along the runway | Optional, often a separate line item | Sometimes | Installer, on site |
| Cable from wall panel to runway feed point | No | Yes | Installer |
| Runway beam, which is the roof structure | No | Yes | Structural engineer |
| Bonding from crane structure to building earth | Bonding jumper only | Earth network | Installer, documented |
| RCD and MCB at the distribution board | No | Yes | Installer |
| Surge arrester at the crane feed | Optional | Yes | Installer |
Two things make a suspension crane worse than a top-running bridge crane here. There is no crane runway beam, so the runway is the building's roof beam and its bottom flange is a structural element nobody wants drilled, welded or clamped. And the machine usually sits 8 to 12 m up, which puts the isolator out of arm's reach. That quietly breaks the lockable isolator rule in most plant electrical policies, and it is the sort of thing an auditor finds in the first ten minutes.
If you are buying, ask where the supply switch physically sits. If the answer is "at the roof", somebody has to own the cable from the wall to the roof and it needs to be written down. Our suspension crane datasheets list isolator position for exactly this reason.
Less than most people guess. Published chain hoist tables for hook-suspension units give 0.75 kW at 500 kg, 1.5 kW at 1 tonne, 3 kW at 2 and 3 tonnes, and 3 kW for a two-fall 5 tonne. Add a geared travel motor and a control transformer and you have the whole machine.
| Capacity | Hoist motor | Travel motor | Control transformer | Connected load | Running current at 400 V | Direct-on-line start |
|---|---|---|---|---|---|---|
| 0.5 t | 0.75 kW | 0.2 kW | 100 VA | 0.95 kW | 2.4 A | 14 A |
| 1 t | 1.5 kW | 0.2 kW | 150 VA | 1.7 kW | 4.1 A | 25 A |
| 2 t | 3 kW | 0.2 kW | 200 VA | 3.2 kW | 7.5 A | 45 A |
| 3 t | 3 kW | 0.4 kW | 250 VA | 3.4 kW | 7.9 A | 47 A |
| 5 t | 3 kW | 0.4 kW | 300 VA | 3.4 kW | 7.9 A | 47 A |
Two conclusions fall out of that table. First, the whole range sits inside 16 A per phase, which is exactly the dividing line between IEC 61000-3-2 and IEC 61000-3-12. A single suspension crane is not a harmonic source worth modelling. It is a starting-current source and a leakage source.
Second, the start current is 6 to 8 times running current even on the smallest unit. A 1 tonne crane pulling 4 A steady pulls 25 A for about a second every lift. If that crane shares a 16 A MCB with anything else, the MCB opens and everybody blames the crane.
Sometimes, and the "sometimes" is where the trouble lives. OSHA 29 CFR 1926.406(a)(4) requires all exposed metal parts of cranes, hoists and accessories to be metallically joined into a continuous electrical conductor. It then says the trolley frame and bridge frame "shall be considered as electrically grounded through the bridge and trolley wheels and its respective tracks unless conditions such as paint or other insulating materials prevent reliable metal-to-metal contact. In this case a separate bonding conductor shall be provided."
Read that clause with a suspension crane in mind. The wheel runs on the bottom flange of a roof beam. That flange left the fabricator with mill scale and arrived on site with two coats of epoxy, typically 120 to 200 micrometres. The wheel presses on paint, not steel.
| Earth path | Condition | Bonding resistance | Verdict |
|---|---|---|---|
| Steel wheel on bare flange | Mill scale removed, dry | 0.02 to 0.08 ohm | Works today, fails after a season of dust |
| Wheel on painted flange | Two-coat epoxy, 160 micrometres | 0.6 to 4.2 ohm | Not a bonding conductor under OSHA 1926.406(a)(4) |
| Wheel on painted flange plus grease | Food plant, washdown, weekly cleaning | 8 to 30 ohm | Fault loop will not open the breaker in time |
| Separate 4 mm² bonding jumper | Crimped both ends, bare metal under the lug | 0.03 to 0.06 ohm | The default answer, and the cheapest item on any fix list |
| Jumper landed on paint | Bolt through the coating, no star washer | 2 to 15 ohm | Looks bonded, is not. Common on site |
The resistance figures above are from our own site measurements on eleven underhung installations between 2025 and 2026. They are not a published survey, and the sample is small. Treat them as orders of magnitude, not tolerances.
Why the resistance matters: IEC 60204-32 Annex A.1.3 gives the alternative acceptance route of reducing touch voltage below 50 V. With a 4.2 ohm flange path and a 25 A fault, the touch voltage on the crane frame sits well above that, which means protection depends entirely on the RCD tripping rather than on the bonding. Two safety layers, one of them painted over.
Verification is not exotic. IEC 60204-32 clause 18.2.2 is a continuity test of the protective bonding circuit and 18.2.3 is a fault loop impedance check. Both need a tester and about twenty minutes. Neither appears in most crane handover packs I have seen.
Because the RCD was selected for a building with no electronics in it, and a crane is electronics. The relevant limit is small: a type A residual current device can only tolerate about 6 mA of smooth DC on the load side before its sensing coil saturates. That is regardless of whether the device is rated 30 mA or 300 mA.
The UK Construction Plant-hire Association technical note on this, published April 2025, is blunt about the workaround people try. Sites raise the type A sensitivity to 300 mA or even 1 A to stop the trips, which "introduces an increased risk of a fatal electrocution should an electrical fault develop, which should have tripped the RCD at 30 mA". The note also points out that RCDs on high-current crane supplies are often relay-based and sensitive to high-frequency currents up to 2 kHz, which is roughly where a hoist inverter switches.
| RCD type | Detects | Blind to | Where it belongs |
|---|---|---|---|
| AC | Sinusoidal alternating current only | Pulsating DC, smooth DC, anything above mains frequency | Nowhere near a crane |
| A | AC plus pulsating DC, up to about 6 mA smooth DC | Smooth DC above 6 mA, high-frequency leakage | Lighting, sockets, resistive loads |
| F | Type A plus mixed-frequency residual current from 10 Hz | Smooth DC | Single-phase inverters, some appliance groups |
| B | AC to 1,000 Hz, pulsating DC, smooth DC, values defined to 100 kHz | Nothing that matters here | Every three-phase drive, and most cranes |
The type B requirement for three-phase drives is not a preference. Doepke's technical guidance states that three-phase variable speed drives must only be used with type B residual current devices, because under certain fault conditions an inverter generates smooth DC residual current in the AC supply and a standard device will simply not trip. ABB's white paper on the subject puts the detection band for type B at up to 100 kHz and credits IEC 62423, introduced in 2007, with the additional requirements.
Schneider adds two practical details that explain most nuisance trips. Type B devices have a built-in tripping delay so that EMC filter inrush on power-up does not trip them, and they use a low-pass filter so that the permanent switching leakage of a converter at 25 kHz does not accumulate into a false trip. Where 300 mA is required for selectivity, the device has to be the selective type with a guaranteed non-triggering time of 50 ms.
None of that helps if the crane shares a board with the lighting. Which brings us to the leak itself.
From four places, none of them large, all of them additive. The numbers below are our planning figures for a typical 1 tonne underhung machine, not a laboratory measurement.
| Component | Typical leakage | Waveform | Note |
|---|---|---|---|
| Hoist motor winding and cable capacitance | 0.5 to 1.5 mA | Capacitive AC | Rises with cable length and humidity |
| Brake rectifier, half-wave, 96 V DC coil | 1 to 3 mA | Pulsating DC | A type AC device cannot see it at all |
| Control transformer, 250 VA | 0.3 to 0.8 mA | Capacitive AC | Inter-winding capacitance, worse when hot |
| Festoon cable, 40 m, five core | 0.4 to 1.2 mA | Capacitive AC | Proportional to length, so a long runway is worse |
| Inverter with integrated EMC filter, if fitted | 3.5 to 10 mA | High frequency plus smooth DC | Dominant term whenever it is present |
| Total, contactor-controlled hoist | 2.2 to 6.5 mA | Mixed | Sits right at the 6 mA limit on a bad day |
| Total, inverter-controlled hoist | 8 to 20 mA | Mixed | Over the limit before the building adds anything |
So the crane is rarely guilty on its own. It is the last 3 mA on a board that already carries 21. The packaging plant I mentioned had 34 LED drivers on the roof, each with a small leakage to earth, and a type A device that had been quietly sitting just under its threshold for two years. The crane arrived, the threshold was crossed, and the warranty claim was written against the crane.
If you want to settle the argument in an hour, clamp the standing leakage on the board with everything off, then energise the crane and read it again. Numbers end these discussions faster than opinions do.
Not in the way the standards mean it. Harmonic emission limits are written in current bands, and a suspension crane sits in the bottom band.
| Input current per phase | Governing document | What it limits | Whose problem it is |
|---|---|---|---|
| Up to 16 A | IEC 61000-3-2 | Harmonic current injected by the equipment | Equipment maker |
| Above 16 A to 75 A | IEC 61000-3-12:2011 with A1:2021 | Harmonic current at 230/400 V, 50 Hz | Equipment maker |
| Above 75 A | Installation-level assessment, IEC/TR 61000-3-6 | Total distortion at the installation | System designer |
| At the point of common coupling | IEEE 519-2022 | Voltage and current distortion, up to the 50th order | Shared between user and utility |
IEC 61000-3-12 is explicit about its own boundaries. It covers equipment pulling more than 16 A and up to 75 A per phase on 230/400 V systems, and its own note says equipment above 75 A "should be considered in the harmonic current requirements for installations". Below 16 A, IEC 61000-3-2 does the work and the limits are per appliance. That is where a suspension crane lives.
IEEE 519-2022 matters at the other end of the building. It sets the quality of power at the point of common coupling, up to the 50th harmonic, and the 2022 revision, published 13 May 2022, turned what had been guidance since 1981 into an enforceable standard. Compliance is shared: the user limits current distortion, the utility controls system impedance.
What this means in practice is that you should not buy a harmonic filter for a 3 kW hoist. You should buy one for a bay of six inverter cranes on a weak transformer, and that is a different article, the one about harmonic distortion from multiple drives.
Almost nothing. This surprises people, so here is the arithmetic. Using the relationship the festoon makers publish, three-phase drop is the square root of three times line impedance times section length times current. For 4 mm² copper at 0.004375 ohm per metre:
| Segment | Current | Drop | Percent of 400 V |
|---|---|---|---|
| Festoon, 40 m, 4 mm², crane running | 7.5 A | 0.23 V | 0.06 % |
| Festoon, 40 m, 4 mm², hoist starting | 45 A | 1.37 V | 0.34 % |
| Wall to roof feeder, 60 m, 4 mm², one crane | 8 A | 3.6 V | 0.9 % |
| Same feeder, three cranes lifting together | 24 A | 10.9 V | 2.7 % |
| Same feeder, five cranes lifting together | 40 A | 18.2 V | 4.5 % |
| Same feeder in 10 mm², five cranes | 40 A | 7.3 V | 1.8 % |
The festoon is not the problem. The cable from the wall to the roof is, and it is the cable nobody scoped. Five underhung cranes on a shared 4 mm² feeder eat 4.5 percent before the crane has done anything wrong, and Annex G of IEC 60364-5-52 suggests 3 percent for lighting and 5 percent for other uses as a design guide. You are at the edge with a clean installation.
There is a cheap fix that costs nothing but a junction box. Festoon suppliers publish that multiple feed points along a line divide the maximum section length, so a mid-run feed roughly halves the drop. One mid-span feed on a 60 m run gets five cranes from 4.5 percent down to about 2.3 percent.
Three things, and only one of them is electrical arithmetic. The voltage drop adds, as the table above shows. The leakage on the shared RCD adds too, because both cranes sit downstream of the same device. And the third thing is coordination: two cranes starting within a second of each other trip an MCB that either one alone would not.
Suspension crane runways are commonly shared. A 12 m bay with three underhung cranes on the same pair of roof beams is normal in a light assembly shop. Each crane brings its own festoon and its own panel, but they all land on the same feeder and often the same RCD.
If that is your layout, the fix order is different from a single crane. Bond each crane separately, do not daisy-chain the bonding through the runway, and give the group its own RCD rather than sharing with lighting. Our hoist and accessory range includes the bonding kits for this, because it comes up on almost every multi-crane order.
If the roof is steel and the crane hangs from it, then yes, whether anyone planned it or not. A suspension crane is a large metal object bonded to the roof structure through its wheels, sitting in the middle of the building envelope. In a thunderstorm it is part of the conductive path.
The UK plant-hire technical note is the most practical document I have found on this. It requires lightning protection to be designed, installed, tested and maintained by competent persons to the BS EN IEC 62305 series and BS 7430. It requires connections at the manufacturer's approved earthing points, and explicitly forbids creating earthing points by drilling or welding structural elements. It requires an earth continuity check after installation, with resistance between the bottom of the structure and earth below 10 ohm. Above 10 ohm, the base has to be bonded to a suitable earth network with a single-core cable of not less than 70 mm², and then re-measured.
| Requirement | Source | Value | Who signs it off |
|---|---|---|---|
| Lightning protection design and test | BS EN IEC 62305 series, BS 7430 | Competent person, documented | Installer |
| Earth continuity after installation | CPA TCIG technical note, April 2025 | Below 10 ohm to earth | Installer, with a reading |
| If the reading is above 10 ohm | Same note | Bond to earth network with 70 mm² single core, re-measure | Installer |
| Earthing point location | Same note | Manufacturer-approved only, no drilling or welding structure | Nobody, usually |
| Action during a thunderstorm | Same note | Crane out of service, operator leaves, nobody touches load or structure | Site supervisor |
| Surge arrester at the crane feed | IEC 61643-11, type 2 device | Optional, cheap against a control board | Installer |
The thunderstorm part is not theoretical for our market. Buyers in Vietnam, Indonesia, Thailand and southern China run underhung cranes under metal roofs in the highest lightning density zones in the world, and the crane control board is the cheapest thing to sacrifice. A type 2 arrester at the crane feed costs less than one service visit.
Cheaper than most people expect, which is the good news. The ranges below are from our 2026 installation quotes, not a published price list, and local labour will move them either way.
| Item | USD range | Who pays | When it is the right answer |
|---|---|---|---|
| 4 mm² bonding jumper, crimps, clamps | 60 to 250 | Crane side | Always. Cheapest item on the list |
| Type B RCD, 30 mA, four pole | 250 to 700 | Building | Any inverter hoist, or a crane sharing a board |
| Dedicated feeder and lockable isolator | 400 to 1,500 | Building | Two or more cranes, or any shared lighting circuit |
| 4 percent line reactor, 3 kW hoist drive | 120 to 350 | Crane side | Only if there is an inverter and a long motor cable |
| Type 2 surge arrester at crane feed | 180 to 450 | Building | Metal roof, high lightning area, no argument needed |
| Earth resistance improvement to a network | 300 to 1,200 | Building | Only when a measurement is above 10 ohm |
| Electrical survey, leakage clamp and loop tester | 250 to 600 | Shared | First, always. It ends the warranty argument |
Order matters more than budget. Bonding first, because it is cheap and it is the one item that is a safety issue rather than a nuisance. Then the survey, because everything after it depends on a measurement. Then the RCD, then the dedicated feeder. A reactor and an arrester come last and only if the symptom points at them.
This table is our own service record across underhung installations between 2024 and 2026. It is not a published statistic, the sample is not large, and it is skewed by the fact that people call us when something trips. Read the ranking, not the percentages.
| Symptom | Share | Root cause, in one line |
|---|---|---|
| Nuisance RCD trip | 31 % | Type A device, shared board, standing leakage already near threshold |
| Loose or missing bonding jumper | 19 % | Never fitted, or landed on paint |
| Festoon trolley or collector damage | 16 % | Trolley misalignment, or a forklift clipped the loop |
| Control transformer inrush trip | 12 % | 250 VA transformer on a 6 A MCB shared with something else |
| Isolator out of reach or not lockable | 9 % | Mounted at the roof, no chain or extension handle |
| Brake rectifier failure | 8 % | Heat, or a surge that found the weakest semiconductor first |
| Surge damage after a thunderstorm | 5 % | No arrester at the feed, metal roof, crane bonded to it |
Three of those seven share one cause: the crane was treated as a standalone machine when it is part of the building's electrical system. Fix the boundary in the contract and most of the list disappears.
Ten lines. If your inquiry has these, the commissioning argument mostly does not happen.
| Line | What to write |
|---|---|
| 1 | Supply voltage, phases, frequency and measured voltage at the crane isolator, not at the incoming panel |
| 2 | Position of the crane supply switch, and whether a lockable isolator within reach of the floor is required |
| 3 | Who supplies and installs the festoon or conductor bar, and to what length |
| 4 | Who supplies the cable from the wall panel to the runway feed point, and its size |
| 5 | RCD type at the feeding board, its rating, and whether the crane has a dedicated device |
| 6 | If a second crane shares the runway, whether the feeder is sized for both together |
| 7 | Contactor control or inverter drive on hoist and travel, and the drive make if specified |
| 8 | Control voltage, and whether PELV is required by the plant standard |
| 9 | Bonding jumper supplied with the crane, its size, and the approved earthing point on the structure |
| 10 | Who signs the continuity and fault loop test, and where the readings go in the handover pack |
Line 10 is the one people skip and the one that saves the most money. A handover pack with readings in it ends an argument that otherwise runs for a month, and it costs twenty minutes at commissioning.
Almost always because the crane added 2 to 6 mA of leakage to a circuit that was already close to its threshold. A type A device can only tolerate about 6 mA of smooth DC before its coil saturates, and that figure does not change when you fit a 300 mA device. If the crane has an inverter hoist, the EMC filter alone can be 3.5 to 10 mA. Fit a type B device on a dedicated circuit and the trips stop. Our suspension crane pages list the drive options because the RCD answers differ between them.
Only when the metal contact is reliable, and paint defeats it. OSHA 29 CFR 1926.406(a)(4) allows the frame to be considered grounded through the wheels and tracks but adds that where paint or insulating material prevents reliable contact, a separate bonding conductor has to be provided. A two-coat epoxy flange typically measures 0.6 to 4.2 ohm from wheel to beam in our own site readings. That is not a fault path. Fit a 4 mm² jumper and measure it.
No. A suspension crane in the 0.5 to 5 tonne range sits under 16 A per phase, which puts it in the IEC 61000-3-2 band rather than IEC 61000-3-12. Its harmonic current is small enough that no filter is justified. What it does have is a start current six to eight times running current, so the symptom you feel is a voltage dip and a tripped MCB, not distortion. Compare that with a bay of six inverter cranes, which is a different problem entirely.
Not necessarily, but you need to know what is on the board. A contactor-controlled hoist with a half-wave brake rectifier produces pulsating DC, which a type A device does detect, so protection still works. The problem is the shared circuit with modern electronics. If the board already carries LED drivers, a PV inverter or a charger, a type B device on the crane circuit is the cheap way to stop arguing. Schneider and Doepke both publish the type B requirement for three-phase drives; neither requires it for a plain contactor hoist.
You cannot fully, but you can get most of the way. With the crane isolated, measure resistance from the crane frame to the building earth bar with a four-wire or a decent low-resistance meter, not a multimeter on the ohms range. Anything over 1 ohm means the bonding path is going through paint, grease or structure joints. Then look at the earthing point itself. If somebody has drilled and tapped a structural member to land a lug, that is not an approved point, and the manufacturer's drawing will show where it should be.
Bonding, then measure. A 4 mm² jumper with crimped ends costs USD 60 to 250 and removes the single most common safety defect we find on underhung installations, which is a frame that looks earthed and is not. After that, clamp the standing leakage on the feeding board with the crane off and again with it on. Those two numbers tell you whether you are buying an RCD, a feeder or nothing at all, and they cost less than the argument.
Related articles:
Suspension Crane Energy Consumption: Where the Kilowatt Hours Go on a 0.5 to 5 Tonne Machine
Suspension Crane Installation and Commissioning: Hangers, Tracks and the Load Test
Crane Power Quality: What a Single Girder Crane Does to a Weak Grid
Crane Harmonic Distortion: What Five Drives on One Busbar Do to a Double Girder
Suspension Crane Safety and Compliance: Roof Structure, Tracks and Annual Inspection