Short answer: you have two separate failures in that cabinet. Insulation that cracks and crumbles right at the lug, while the same cable stays flexible in your other panels, means that joint ran hot — and the fix is the termination and the protection settings. The droplets on the walls are a separate dew-point problem a small heater solves. Fix the joints first; the heater stops the rust, not the cooking.
Why is the insulation crumbling only at the lugs?
Because the joint is the weak point: a loose, corroded or mismatched lug raises resistance at one spot, the heat starts exactly there, and it works back through the first stretch of insulation while the cable body stays sound.
A cable that survives elsewhere with the same marking tells you the conductor is fine and the terminations are not. The usual sequence: the lug palm rocks on the pad instead of sitting flat, the hardware was never torqued, or an aluminium conductor landed on a copper-only lug. Contact resistance climbs, and the heat embrittles the insulation nearest the joint. Most building-wire insulation is rated for 90 C, and weeks of running hotter than that strip out its flexibility until it cracks at a touch. Two things hide it from you: these joints show little discoloration, and the panel reads cool the moment the load drops.
Overload protection is the other half of the cause. The NEC sizes overcurrent protection to the ampacity of the conductor, so a setting parked above the cable's ampacity lets a mild, sustained overload do the same damage further along the conductor. Check the setting against the cable, and against the load history, before you clear the panel.
Is the damage from heat or from water?
Heat, when the damage stops at the lugs: water moves down and along surfaces and leaves deposits, rust streaks and tracking marks, while joint heat stays concentrated at the termination and kills only the insulation nearest it.
- Cracking at the lugs, flexible cable body — termination heating or a past overload; go to the joint resistance and the protection settings.
- One lug hotter than its twin at the same current — a high-resistance joint; scan it live, then measure it dead.
- Stiffness along the whole conductor, past the first bend — sustained overcurrent; log the load against the cable's ampacity.
- Droplets, rust streaks, deposits, tracking across plastics — condensation or ingress building leakage paths; treat as a dew-point fault.
- Every plastic and rubber part in the box brittle — heat well beyond one lug, or a chemical in the air; sample the atmosphere.
The same cable running cool in your other panels is the comparison that convicts the lug. Trace the water path separately with water inside a sealed enclosure and the condensation guide, and treat the two faults as unrelated until your checks link them.
How much heat does a loose lug make at 200 A?
Forty watts at a single terminal: a sound bolted lug around 50 microohms dissipates 2 W at that current, and that same joint loosened to 1 milliohm dissipates 40 W into a few grams of metal.
The arithmetic, with P = I² × R at I = 200 A, so I² = 40,000 A²:
- Sound joint: 40,000 × 0.00005 Ω = 2 W
- Loose joint: 40,000 × 0.001 Ω = 40 W
Forty watts is a soldering-iron element clamped to the insulation, running every shift. Week after week, the joint runs far past the cable's temperature rating long before anything smokes — field thermography regularly catches defective connections sitting 45 C [81 F] above their neighbours with the current barely moved.
That is also why your breaker stayed quiet. Protection sees amps, and a good joint reads exactly the same as a bad one carrying the same current; the heat lives at the joint, invisible to the trip curve. A mis-set overload and a loose lug produce the same cable, and both can be present in the same panel.
How do you check the joints safely?
Scan the panel live at real load with an infrared camera first, then lock it out and measure the flagged joints dead with a micro-ohmmeter — under 100 microohms at a bolted lug on a feeder that size is what you want.
- At full load, compare each lug with its twin at the same current. More than 10-15 C [18-27 F] hotter marks a high-resistance joint; photograph it for the record.
- Lock the cabinet out and prove absence of voltage before you open a single joint — this is the one step in the job that can kill you.
- Open the flagged joints. Look for a palm that rocks on the pad, oxide in the barrel, aluminium on a copper-only lug, and hardware that spins without seating.
- Re-make with a lug matched to the conductor and a filled barrel, and torque to the value marked on it — about 45 lb-in [5 N·m] on 1/4 in hardware where nothing is marked.
- Measure the finished joint and log the microohms as the reference value for next year's scan.
Why is water forming on the inside walls?
Because a surface inside the box falls to the dew point of the air around it: air at 25 C [77 F] and 80% RH condenses on any surface at or below 21 C [70 F], and walls, incoming cable and chilled pipework are those surfaces.
The worst hour is after shutdown. While the drives run, the equipment holds the inside warm; the moment the load drops, the walls and metalwork cool fastest while the air is still damp. The droplets appear at night and dry before your morning walk. Where cooling water runs through the panel, isolate it when the machine is idle if the machine maker allows it — otherwise it keeps a condensation surface cold with nothing left to offset it.
To prove the path, log three numbers through a shutdown: inside air temperature, relative humidity, and the surface temperature of the coldest wall or pipe. Compare that surface with the dew point computed from the first two; a surface crossing it during the problem hours is your answer. Droplets that gather under one gland rather than evenly across the walls are ingress instead — how to check an enclosure after water testing works that path.
What size heater stops the condensation?
About 25 W for a 20 × 16 × 8 in box: holding the inside 5 C [9 F] above the coldest surface loses roughly 21 W through the walls, and the next size up covers it with margin.
- Wall area: 2 × (20 × 16 + 16 × 8 + 20 × 8) = 1216 in² ≈ 0.78 m²
- Loss at a 5 C rise, still air each side of a steel wall, U ≈ 5.5 W/(m²·K): 0.78 × 5.5 × 5 ≈ 21 W
- Fit a 25 W anti-condensation heater
Double the heater size for outdoor or windy sites, which roughly double the loss. Control it with a humidistat set to 60-65% RH in hot, humid plants; in temperate ones a thermostat a few degrees above the coldest night works. Small enclosure heater selection works the sizing in detail, and when an outdoor enclosure needs a heater covers the outdoor case.
How do you repair the cable and set the protection?
Cut back in 2 in [50 mm] steps until the insulation bends without cracking — expect the first 2 ft [600 mm] to be dead — then fit a conductor-matched lug torqued to its mark, and set the overload protection to the cable's ampacity.
If the damage runs past the slack you have, pull a new conductor; a buried splice inside a cabinet rarely ends well. Before you re-energize, dry the panel and insulation-test the repaired circuits at 500 V DC, with 1 MΩ as the floor for dry low-voltage wiring. Once the repair has settled, re-scan the lugs at load — a healthy one sits within a few degrees of its twin — and keep doing it on your normal shutdown schedule.
When should you replace the whole enclosure?
When the box itself is beaten: a pitted gasket face, a seized hinge, rust holes at the fixings, or a panel that keeps sweating because it is the coldest object in the room; if only the joints were bad, the enclosure can stay.
Rust streaks from years of condensation are cosmetic on 304 until they reach the gasket face. Near washdown, salt spray or chlorine aerosols, move to 316L. A 16 × 16 × 6 in 316L wall-mount enclosure with a hinged gasketed door and a removable mounting panel is the drop-in shape for a panel this size; it is rated NEMA 4X / IP66 by the manufacturer and not UL listed. On a custom box you set every cutout — gland plate, drain, heater mount — and it is delivered in 10 business days anywhere in the world, counted from your approval of the drawing for custom work. Browse 316L wall-mount enclosures or start from the enclosure selector if the environment around this panel has changed.
Enclosures in this article
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 16 × 16 × 6 in | 316L stainless | 16 x 16 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Does aluminium cable need a different lug than copper?
Yes. Use a bimetallic or aluminium-rated lug and fill the barrel, because a plain copper-only lug on an aluminium conductor builds a high-resistance joint. Aluminium also cold-flows under the hardware, so re-check the torque after the first month of service and then on the same cycle as your thermography scans.
Will an anti-condensation heater dry out a cabinet that already got wet?
No. A heater holds surfaces above the dew point; it will not pull standing water out of a lug barrel. Mop the box dry, run warm air through the open door for 4-6 h, then watch it through the next 24 h of the worst night cycle before you call it dry and close it up.
Where should the heater sit inside the enclosure?
Low in the box and at least 3 in [75 mm] clear of cable insulation and plastic parts, mounted below the electronics so convection carries the warm air up past them. In a tall cabinet one small heater at the bottom beats a big one at the top, because the warm air stratifies where you need it.
How often should the lugs be re-scanned after the repair?
Once after the first week at full load, while the new joints settle through their first thermal cycles, then every 6-12 months with a logged image of each connection. Keep the comparison honest: same load, same camera angle, same twin joint as the reference.
