Subpanel Feeder Runs Warm While Main Stays Cool Decision Tree

Why this matters

A subpanel feeder that runs noticeably warm while the upstream main breaker and service conductors stay cool is reporting that the heat is being generated locally, in the feeder run or its terminations, not by a service-wide overload. Three causes produce this signature, and they carry very different urgency: a loose lug at either feeder end (a glowing-connection fire path), an undersized or over-loaded feeder (a thermal-rating violation), or a feeder neutral carrying unexpected return current from an improper neutral-ground bond at the subpanel (an objectionable-current and shock issue). A tech who feels a warm feeder and just retorques the obvious lug without checking the other two causes can miss the actual fault. This tree forces a measurement-driven split.

The decision flow at a glance:

  Subpanel feeder warm, main cool - why?
  |
  +-- 1. Drop across one lug? ----------> LOOSE /
  |                                       CORRODED LUG
  |
  +-- 2. Conductor warm along length? --> AMPACITY /
  |                                       OVERLOAD
  |
  +-- 3. Aluminum creep at lug? --------> ANTIOXIDANT +
  |                                       TORQUE
  |
  +-- 4. Neutral bonded at subpanel? ---> REMOVE BOND
  |
  +-- 5. Current over ampacity? --------> UPSIZE / SHED
  |                                       LOAD

Symptom presentation

The customer reports warmth, a faint burning smell, or discoloration at a subpanel feed, sometimes a breaker that trips on hot days, while the main panel feels normal. Ask whether the subpanel was added or extended recently, what loads it serves (a shop, an EV charger, an addition), and whether the feeder is aluminum (far more prone to lug creep) or copper. Establish the feeder conductor size and the feeder breaker rating before touching anything; an undersized feeder for the connected load is a paperwork-and-measurement finding, not a guess.

Before the cover comes off

This tree splits into two kinds of work with two different rules, and the split is not optional.

The measurements are energized work. Clamping a feeder and reading drop across a lug can only be done with load on it, which is the narrow case 29 CFR 1910.333(a)(1) allows. Work it inside the arc flash boundary and in the PPE your employer's electrical safety program establishes from the risk assessments at NFPA 70E-2021, 130.5 and 130.7, sized to the available fault energy at that panel rather than to the 240 volts you are reading. A feeder lug in a panel with a hot spot behind it is the exact geometry that produces an arcing fault when a probe slips.

Everything after the measurement is de-energized work. Opening a termination, re-torquing a lug, cutting a conductor back, pulling a bonding screw, and swapping a breaker are all done with the feeder isolated, locked and tagged under 29 CFR 1910.333(b)(2), or 29 CFR 1926.417 on construction work, and proved dead with the live-dead-live sequence at NFPA 70E-2021, 120.5 on both ends of the feeder. Both ends. Killing the feeder breaker leaves the main-panel end of that same feeder lug energized on its line side, and that end is half the terminations this tree asks you to open.

Quick checks before isolation

With representative load running, clamp the feeder hots at the subpanel and confirm actual current against the feeder breaker and conductor ampacity (NEC 310.16). Read voltage drop across each feeder lug (probe both sides of the termination) at the main and at the subpanel; a good lug drops near 0V, a loose one drops volts under load and is your hot spot. Thermal-scan or back-of-hand check each lug. Confirm the subpanel neutral and ground are separated (four-wire feeder, isolated neutral bar, bonding screw removed) per NEC 250.32 - a bonded subpanel neutral routes return current onto the EGC and feeder, heating it.

Isolation tree

Step 1 - Loose connection versus distributed heat. Measure voltage drop across each of the four feeder terminations (two hots, neutral, ground) at both ends under load. If one termination shows a multi-volt drop and a hot spot, the fault is that loose or corroded lug - go to Step 2. If all terminations read near 0V drop but the conductor itself runs warm along its length, the heat is distributed and points to an ampacity problem - go to Step 3.

Step 2 - Loose lug or aluminum creep. Open the hot termination. Aluminum feeders commonly loosen from thermal cycling (cold flow) and need antioxidant and re-torque to the listed value; a discolored or pitted lug must be cleaned or the conductor end cut back to clean metal and re-landed. If the lug or busbar shows arcing damage, the panel lug or the breaker may need replacement. Re-torque to the manufacturer's marked value with a calibrated torque tool per NEC 110.14(D), never by feel.

Step 3 - Undersized or overloaded feeder. Compare clamp-meter current to the conductor ampacity at the correct temperature column and the feeder breaker rating. If load is at or near the conductor's ampacity, the feeder is simply working hard within rating and warmth is expected; if measured current exceeds the conductor ampacity, the feeder is overloaded - either the breaker is oversized for the conductor (a protection violation per NEC 240.4) or the connected load grew past the feeder design. Recalculate the subpanel load per Article 220 and upsize feeder and breaker, or shed load.

Step 4 - Neutral-ground bond at the subpanel. Open the subpanel. If the neutral bar is bonded to the enclosure (bonding screw or strap in place) on a feeder-fed subpanel, normal neutral current splits between the feeder neutral and the EGC, heating both and creating objectionable current. Remove the bond, isolate the neutral bar, and verify the ground bar is bonded to the can instead. Clamp the EGC afterward; it should read near zero current under load.

Confirming the diagnosis

A confirmed loose lug shows a localized hot spot and a measurable voltage drop across that one termination that disappears after re-termination, with the feeder then running cool under the same load. A confirmed ampacity problem shows distributed warmth with all lugs tight and clamp current at or above the conductor rating, resolving only with an upsize or load reduction. A confirmed subpanel bond fault shows measurable current on the EGC under load that drops to zero once the neutral bar is isolated and the bonding screw removed, with feeder warmth easing accordingly.

Remediation

Re-terminate loose lugs to the marked torque value with antioxidant on aluminum and clean conductor ends; replace arc-damaged lugs, breakers, or busbars. For an undersized or overloaded feeder, recalculate per Article 220 and upsize conductor and overcurrent device together so the breaker never exceeds conductor ampacity (NEC 240.4). For a subpanel neutral-ground bond, remove the bonding means, install or confirm an isolated neutral bar and a separate bonded ground bar, and run a proper EGC if the feeder is only three-wire. Know which section is behind that requirement, because it depends on where the subpanel is. For a subpanel in the same building, the rule is that the grounded conductor is not re-connected to equipment or to grounding conductors anywhere on the load side of the service disconnect, which lives in NEC Article 250, with the panelboard side of it in Article 408. For a subpanel in a separate building or structure, NEC 250.32 is the section, and it is the one that ended the old three-wire-to-the-garage practice. Re-clamp and re-scan under load to verify the heat is gone.

References

• NEC 2023, Section 250.32 - Buildings or structures supplied by a feeder, separate neutral and ground
• NEC 2023, Article 250 Part I and Part V, and Article 408 - the grounded conductor is not re-bonded on the load side of the service disconnect, and panelboard neutral and equipment grounding terminal bars are kept separate except at service equipment
• NEC 2023, Section 110.14(D) - terminations tightened with a calibrated torque tool where a torque value is given
• NEC 2023, Section 240.4 - protection of conductors, overcurrent device not to exceed conductor ampacity
• NEC 2023, Table 310.16 - ampacities of insulated conductors
• NEC 2023, Section 250.6 - objectionable current over grounding conductors
• NEC 2023, Article 220 - feeder and subpanel load calculation
• 29 CFR 1910.333(a)(1) and (b)(2); 29 CFR 1926.417; NFPA 70E-2021, 120.5, 130.5 and 130.7