Crash course: Voltage, amperage, and resistance basics for solar installers (part 2)

Part 2: Beyond Ohm's law on the roof

Part 1 covered the basics: voltage pushes, current flows, resistance fights back. Part 2 gets into what actually bites you on a PV install. Temperature swings, voltage drop on long DC runs, conductor sizing under 690.8, and why your string voltage at sunrise on a 20F morning is not the voltage on the spec sheet.

If you size a system using STC numbers and forget the cold correction, you will trip an inverter or pop a fuse. Worse, you will fail an AHJ inspection and reroll the truck.

Cold temperature voltage correction

PV module Voc rises as temperature drops. NEC 690.7 requires you to correct maximum voltage for the lowest expected ambient temperature at the site. Two methods are allowed: the table in 690.7(A) or the manufacturer's temperature coefficient (typically around -0.27%/C to -0.35%/C for crystalline silicon).

For a 1000V system, blowing past the limit by even a few volts is a code violation and an insurance problem. Always run the math on the coldest record low for the ZIP code, not the average winter temp.

• Pull ASHRAE 2% extreme low or use NEC Table 690.7(A)
• Multiply Voc by the correction factor
• Confirm string Voc, corrected, stays under inverter max input and conductor/equipment rating
• Document the calc on the plan set, AHJs ask for it

If the inverter says 600V max and your corrected string Voc is 598V, you are not "fine." You are one cloudy edge-of-cloud irradiance spike away from a fault code. Leave headroom.

Voltage drop on DC and AC runs

NEC does not mandate a hard voltage drop limit, but Informational Note 4 to 210.19(A) and 215.2(A) recommend 3% on branch circuits and 5% total. For PV, most jurisdictions and good practice cap DC string drop at 2% and AC output at 1% to 2%, because every percent is lost production for 25 years.

Voltage drop is governed by Vd = 2 x I x R x L for single phase, where R is the conductor resistance per foot from NEC Chapter 9, Table 8 or Table 9. On long roof-to-inverter runs, upsize the conductor before you upsize the string.

1. Calculate one-way distance, then double it for round trip
2. Use Imp at STC for DC, not Isc, for realistic loss numbers
3. Account for conductor temperature, hot conduit on a roof runs 60C plus
4. Recheck after final routing, plans rarely match the real run

Conductor ampacity and the 690.8 stack

PV circuits use a different sizing stack than a typical branch circuit. NEC 690.8(A) defines the maximum circuit current as Isc x 1.25 (or a calculated value per 690.8(A)(1) Exception). Then 690.8(B) requires a second 1.25 multiplier for continuous duty when sizing overcurrent and conductors without temp/fill adjustments.

That stacks to 1.56 x Isc before you even touch ambient temp correction (310.15(B)(1)) or conduit fill (310.15(C)(1)). Skip a step and your THWN-2 is undersized in summer.

• Step 1: Imax = Isc x 1.25 per 690.8(A)
• Step 2: Conductor and OCPD = Imax x 1.25 per 690.8(B), unless using 100% rated equipment
• Step 3: Apply ambient and fill derates from 310.15
• Step 4: Confirm result is at or below the 75C or 90C column ampacity for the terminal rating

Resistance, connections, and heat

Most field failures are not modules or inverters. They are connections. A loose lug, a dirty MC4, or a backed-out terminal screw introduces resistance, and resistance under current means heat. Heat means oxidation, which means more resistance. The loop runs away until something melts.

Torque every termination to the manufacturer's spec, use a calibrated torque screwdriver, and mark it with a paint pen. NEC 110.14(D) requires terminations to be torqued per listing. Re-torque is not in the NEC, but most O and M agreements call for it at the one-year mark.

If a combiner box smells like burnt fish, that is melting plastic from a hot lug. Kill the string, open it cold, and never just retighten a discolored terminal, replace it.

Putting it together on a real string

Say you have a 12-module string, Voc 49.5V, Isc 11.2A, temp coefficient -0.28%/C, and the site's record low is -10C. Corrected Voc per module is 49.5 x (1 + (-0.28% x (-10 - 25))) = 49.5 x 1.098 = 54.4V. String Voc is 12 x 54.4 = 652V. That kills a 600V inverter, fits a 1000V inverter with margin.

For conductors: Imax = 11.2 x 1.25 = 14A, then 14 x 1.25 = 17.5A for sizing. In 40C ambient with three current-carrying conductors in conduit, 10 AWG THWN-2 at the 90C column derates plenty for that load. On a 200 foot run at 14A, voltage drop in 10 AWG is roughly 4.5V, about 0.8%, acceptable.

Run the numbers every time. The modules and inverter spec sheets are the source of truth, the NEC is the floor, and the AHJ has the final word.