AWG Selection: A Decision Tree With Voltage Drop, Ampacity, and Bundle De-Rating Math

AWG selection looks like a single-table problem until you stack ampacity, voltage drop, ambient temperature and bundle de-rating on top of each other. Get any one of them wrong and the harness either over-heats in service or over-costs at quote. Here is the decision tree we walk every customer through during DFM review.

The four inputs that matter, in priority order:

1. Continuous current (steady-state amps, not inrush).
2. Allowable voltage drop at the load (typically 3% on signal, 5% on power per NEC recommendation, 10% on automotive 12 V systems).
3. Ambient temperature at the harness routing, including any nearby heat sources.
4. Bundle count — how many current-carrying conductors share the same loom or conduit.

Step 1 — Establish the base ampacity

NEC 2023 Table 310.16 is the canonical reference for 60/75/90 °C insulated copper conductor ampacity in free air or in conduit, derated for up to three current-carrying conductors. The relevant rows for harness work:

For most harness work, we look up the 90 °C column (this is what TE Raychem 55A wire and most automotive grade wire use). For UL 1015 and UL 1007 hookup wire, the 80 °C or 60 °C column is the right reference depending on the insulation grade.

Step 2 — Calculate voltage drop

Voltage drop per the standard one-way DC formula:

V_drop = 2 × L × R × I

Where L is one-way length in meters, R is the wire resistance in Ω/km divided by 1000 to get Ω/m, and I is the current in amps. The factor of 2 accounts for the return path (assumes return wire is the same gauge).

Worked example: AWG 14 carrying 15 A over a 10-foot one-way run (3.05 m).

• R = 8.28 Ω/km = 0.00828 Ω/m
• V_drop = 2 × 3.05 m × 0.00828 Ω/m × 15 A = 0.758 V

On a 12 V system this is 6.3% drop — within the 10% NEC ceiling but above the 5% recommendation. On a 5 V system this is 15% drop and disqualifies AWG 14 immediately. For 5 V at 15 A over 10 ft, the buyer needs AWG 10 at minimum (R = 3.28 Ω/km gives V_drop = 0.30 V or 6%).

Step 3 — Apply temperature de-rating

NEC 2023 Table 310.15(B)(2)(a) gives the temperature correction factor. For 90 °C-rated wire run in 40 °C ambient, the factor is 0.91. For 50 °C ambient, the factor drops to 0.82. For 70 °C ambient (engine bay automotive), the factor is 0.58 — meaning AWG 14 90 °C wire that would carry 25 A in a 30 °C lab carries only 14.5 A in a 70 °C engine bay.

Step 4 — Apply bundle de-rating

UL 758 and NEC 310.15(C) both require de-rating when more than three current-carrying conductors share a bundle. The de-rating factors:

Critical clarification: the count is current-carrying conductors only, not total wires. Ground and neutral conductors that are not normally carrying current do not count. On a typical 32-circuit signal harness with 4 power/ground pairs, the bundle count is 8 (the 4 power conductors plus 4 returns).

Worked example — full decision tree

An EV charging gun harness, 12-circuit, with three pairs of high-current power conductors, three pairs of CAN signaling, and three pairs of pilot/proximity signaling. The harness is 2 m long, runs through a sealed strain-relief overmold rated to 90 °C insulation. Worst-case service ambient is 60 °C inside the gun handle.

Power pair (HV+ / HV-) at 30 A continuous, 5% drop allowed on a 400 V DC system (= 20 V budget):

1. Base ampacity at 90 °C: AWG 8 carries 55 A.
2. Temperature de-rating @ 60 °C ambient with 90 °C wire: × 0.71. New ampacity: 39 A.
3. Bundle de-rating (12 conductors total but 6 current-carrying): × 0.80. New ampacity: 31 A.
4. Voltage drop: 2 × 2 m × (2.06 Ω/km / 1000) × 30 A = 0.247 V (well under the 20 V budget).
5. Verdict: AWG 8 is correct. AWG 10 would still pass voltage drop (0.39 V) but would fail the de-rated ampacity check (40 × 0.71 × 0.80 = 22.7 A < 30 A required).

For the high-voltage harness specifics, see the High Voltage Harness page where we walk through the full HiPot and insulation-resistance test protocol.

Common DFM call-outs at this stage

• Up-gauge return conductor. If the return path is shared across multiple loads, sum the currents and pick the gauge for the worst case. Many drawings spec same-gauge return as outgoing — fine for single-load, wrong for shared-return.
• Round up only one step. Going from AWG 22 to AWG 20 adds 60% copper cost on the wire but saves significant inspection rework. Skipping to AWG 18 wastes copper.
• Watch the connector ampacity. Many connectors have lower ampacity than the wire. A TE Deutsch DT 0.4 mm² contact is rated 13 A — putting AWG 16 wire on it (15 A wire ampacity) means the connector is the limiting factor.
• Account for the splice tap. Y-splice taps de-rate by an additional 0.85 factor on the upstream leg.

FAQ

Why does NEC matter for a harness shipped overseas?

It does not, formally — but most international harness specs (IEC 60364, JIS C 8201, GB 3667) cite NEC as one of the reference tables. UL 758, which is the dominant insulation construction standard for harnesses sold into the US, requires the de-rating tables to match NEC 310. So in practice, NEC 2023 is the lingua franca.

How does AWG translate to mm² for European customers?

AWG 22 ≈ 0.34 mm². AWG 20 ≈ 0.5 mm². AWG 18 ≈ 0.75 mm². AWG 16 ≈ 1.0 mm². AWG 14 ≈ 2.0 mm². AWG 12 ≈ 4.0 mm². The IEC 60228 metric is what drives most EU automotive harness drawings; we accept either spec on the RFQ.

What about flex life — does AWG selection change for high-flex applications?

Yes. JIS C 3406 and UL 1581 both require finer-strand construction (105 strand vs 65 strand on a typical 18 AWG) for cables rated for > 5 million flex cycles. The ampacity is identical to standard stranding, but the wire diameter is slightly larger and the cost is 25–40% higher. For robotic and AGV applications, see the Robotics & Automation page where we cover the JIS C 3406 selection.

How does this all change at 1,000 V DC?

Above 600 V the ampacity tables shift to UL 4703 and IEC 60364-7-712 (PV systems). The de-rating factors are similar, but the insulation-rating column moves up — most 1,000 V applications run at 90–105 °C insulation rating. We covered the high-voltage application in the High Voltage Harness page.

One more lever — the connector matters

The fifth input that most AWG decision trees ignore: the connector contact ampacity. The wire is rarely the binding constraint on a real harness; the contact is. Quick reference for common contact families we use in Cavite:

If your harness drawing specs AWG 16 wire (15 A wire ampacity) on a TE Deutsch DT 0.4 mm² contact (13 A contact ampacity), the contact is the binding constraint. The wire is over-spec by 15% and the connector is at 100% of its rated ampacity. Either up-spec the connector to a 1.0 mm² contact or down-spec the wire to AWG 18.

This is one of the eighteen items on our DFM checklist for harness RFQs and one of the most common cost-down opportunities in a typical 24-circuit industrial harness.

Closing

The two AWG mistakes we see most often on inbound RFQs: undersized return conductors on shared-return architectures, and missed bundle de-rating on harnesses with more than 6 current-carrying conductors. Both add 6–18% latent cost in the form of warranty failures or field over-temp callbacks. Both are catchable in a 30-minute DFM review.

If you have a drawing under review and want a fresh AWG verification, send it to the RFQ desk and we will mark up the gauge selection in red within 24 hours, free, before any quoting starts. For deeper conversation on connector ampacity matching, see the Crimping capability overview and the Custom Wire Harness page.

Sources

• NEC 2023 Tables 310.15 and 310.16 — National Electrical Code, NFPA 70-2023.
• UL 758 — Standard for Appliance Wiring Material, 16th edition, 2023.
• UL 1581 — Reference Standard for Electrical Wires, Cables, and Flexible Cords.
• IEC 60228 — Conductors of Insulated Cables, Edition 3.0, 2004 (current with 2014 amendments).
• JIS C 3406 — Cables for Industrial Automation, 2017 revision.