Continuous Discharge Current on Battery Protection Boards: Definition, Calculation Inputs and Reading the Rating

Why the continuous discharge rating needs a definition before it needs a number

Every protected lithium pack carries a headline figure such as "40 A continuous". That number is not a property of the board alone. It is the outcome of a thermal judgement: which component gets hottest first, at what ambient temperature, with how much copper around it, and against which temperature limit. Two boards with identical MOSFETs and the same current printed on the label can legitimately carry very different currents if one is tested free-standing in still air and the other is bolted to a metal enclosure with a heatsink pad.

This article explains what the continuous discharge current specification means, which inputs actually drive it, how to build an estimate from a bill of materials, and how to read a vendor rating without assuming it transfers to your assembly. It is a measurement-and-reading guide, not a claim about any particular product.

Parameter definition: continuous versus peak

Three ratings are routinely confused on protection board (PCM/BMS) datasheets and listings:

  • Continuous discharge current — the current the board can pass indefinitely while every component stays inside its own long-term temperature limit. "Indefinitely" is the operative word: the board must reach thermal equilibrium, not just survive a few seconds.
  • Peak / instantaneous discharge current — a short-duration level tied to a protection delay. A board quoting 120 A peak usually means the protection circuit will not trip before a delay window expires; it does not mean the MOSFETs can dissipate that power for minutes.
  • Overcurrent protection (OCD) trip current — a threshold in the protection IC, deliberately set above the continuous rating, with its own tolerance and delay. It is a safety limit, not a working current.

The gap between the continuous rating and the trip threshold is intentional headroom. A load that draws the continuous rating should never trip protection; a genuine fault should always trip it.

Units and test conditions that belong with the number

Amperes alone are incomplete. A usable rating statement also names:

  • Ambient temperature at which the rating holds — commonly 25 °C, sometimes 20 °C or 40 °C. Ratings without an ambient are marketing numbers.
  • Temperature limit used — junction temperature of the switching MOSFETs, case temperature of the shunt, or PCB copper temperature.
  • Thermal environment — free air, still air inside an enclosure, forced air, or a metal chassis acting as a heatsink.
  • Mounting and copper area — board area and copper thickness materially change the achievable continuous current because they set the thermal resistance from the die to ambient.
  • Load duration and duty cycle — continuous means steady state; pulsed loads are treated with an RMS (root-mean-square) calculation instead.
  • Cell count / series configuration — the same board family may be offered in 3S, 4S or more, and higher series counts raise total dissipation for the same current.

The four limits behind the rating

Estimating a continuous current is a chain of "what fails first" checks. Each candidate limit yields its own maximum current; the smallest of them governs.

1. MOSFET conduction loss and junction temperature

The main switching devices dissipate I² × RDS(on). Conduction resistance rises with temperature, so the calculation is circular and must be iterated: guess a junction temperature, look up the corresponding on-resistance, compute loss, compute the temperature rise, and repeat until it settles. As a first-order approximation, RDS(on) at temperature T can be taken as RDS(on) at 25 °C multiplied by a factor of (1 + α × (T − 25 °C)), where α is the datasheet temperature coefficient; TI's application brief SDAA117 uses exactly this linear form, and its measured figures show a single FET's on-resistance rising by roughly a factor of 2.5 between −40 °C and 125 °C. That factor of two to three is why a 25 °C number cannot be used directly.

2. Shunt resistor (current-sense) power rating

Boards that sense current with a dedicated shunt dissipate I² × R across it as well. The shunt datasheet gives a rated power at a specified case temperature and a derating curve above it. A shunt rated at a given power at 70 °C may carry only a fraction of that at 100 °C, and the current scales with the square root of the derating factor, so a 30 % reduction in allowable power costs about 16 % of current.

3. PCB copper and connector limits

Copper traces, via count and terminal blocks have their own current-carrying capability, and connector contact resistance adds series loss. On compact boards this is frequently the binding constraint even when the semiconductors are generously sized.

4. Protection threshold and delay

The OCD threshold must sit above the continuous rating with margin for tolerance and for on-resistance spread. Where protection is derived from MOSFET voltage drop rather than a shunt, the trip point drifts with temperature: in the TI example, a design targeting a 15 A threshold behaved as roughly 23 A at −40 °C and 8 A at 125 °C without compensation. If the drift is not characterised, the "continuous" rating has to be pulled down to keep the low-temperature trip point above it.

How to read the specification on a datasheet

  1. Find the rating table and note which of the three ratings above is being quoted. If only one number appears with no qualifier, assume it is the most favourable reading and discount accordingly.
  2. Look for the footnotes: ambient temperature, thermal condition, and whether the number applies to the bare board or an installed assembly.
  3. Check whether the current is stated per board or per parallel string, and how many devices are fitted in parallel.
  4. Cross-check against the protection thresholds. A continuous rating within 20–30 % of the trip threshold leaves little room for tolerance and is a warning sign that the two numbers were generated separately.
  5. Check the cell configuration the rating belongs to. A rating published for a 3S configuration should not be carried over to 6S without revisiting dissipation.
  6. Note the date and revision of the datasheet. Ratings are revised when components are substituted; an undated listing cannot be relied on.

Estimate the continuous current from a bill of materials

A defensible first-pass calculation, using only published component values, follows this sequence:

  1. Set the allowed temperature rise. Choose the maximum junction temperature you accept (for example, a conservative margin below the datasheet maximum) and the ambient inside your enclosure.
  2. Get RDS(on) at that junction temperature. Start from the 25 °C maximum value, apply the datasheet temperature coefficient or the normalised curve, and use the worst-case (maximum) column rather than typical.
  3. Compute allowed dissipation. Divide the temperature rise by the junction-to-ambient thermal resistance of the actual assembly (die-to-case plus case-to-ambient including copper and enclosure). If the thermal resistance is unknown, treat the result as an upper bound only.
  4. Convert to current. For a single conduction path, I = √(P / R). With n devices in parallel, total resistance is R / n, assuming balanced sharing, so the current scales as √n—doubling MOSFETs raises the thermal current limit by about 41 %, not 100 %.
  5. Repeat for the shunt and the copper. Take the smallest of the three results.
  6. Apply the protection check. Confirm the low-temperature trip threshold stays above the candidate current with margin.

Worked illustration with published behaviour

Take a single MOSFET whose 25 °C maximum RDS(on) is 3 mΩ, a temperature coefficient that takes it to about 7.5 mΩ near 125 °C (the factor-of-2.5 behaviour reported in TI's SDAA117), an accepted junction temperature of 125 °C, a 60 °C ambient inside the enclosure, and an assembly thermal resistance of 20 °C/W from junction to ambient.

  • Allowed rise: 125 − 60 = 65 °C.
  • Allowed dissipation: 65 °C ÷ 20 °C/W = 3.25 W.
  • Current: I = √(3.25 W ÷ 0.0075 Ω) ≈ 20.8 A.
  • The same board tested on a bench at 25 °C with 10 °C/W would allow 125 − 25 = 100 °C, i.e. 10 W, and about 36.5 A.

The identical hardware therefore supports roughly 21 A inside the enclosure and roughly 36 A on the bench. Neither number is wrong; they describe different thermal conditions. This is precisely why a vendor's figure must be matched to the conditions of your build before it is adopted.

Comparison caveats when two boards are side by side

  • Same current, different conditions. Compare ambient, thermal mounting and temperature limit before comparing amperes.
  • Typical versus maximum RDS(on). A rating computed from typical values can be 20–30 % optimistic relative to worst case.
  • One MOSFET versus a parallel bank. Board area and device count both change the achievable current; a compact board with many devices may still be thermally worse than a larger board with fewer.
  • Shunt versus shunt-less sensing. Shunt-less designs save the shunt's I²R loss but inherit temperature-dependent trip thresholds, which must be compensated or de-rated.
  • Peak figures quoted as continuous. A pulse-rated number presented without a duty cycle is not comparable to a steady-state rating.

Compatibility, limits and safety notes

The continuous rating must be read together with the pack's other constraints: cell maximum continuous discharge current, fuse or disconnect device rating, connector and wire gauge, and the load's own duty profile. The lowest of these governs the system, not the board. Where a load is pulsed, use RMS current over the realistic duty cycle rather than the peak. For battery packs in regulated products, the applicable safety standards and transport rules impose their own verification; this article does not substitute for that testing or for the cell manufacturer's data.

Source and version record

  • Texas Instruments, application brief Improving Accuracy of Shunt-Less Short Circuit Protection in Wide Temperature Range Using LM74912-Q1, SDAA117, September 2025 — source of the linear RDS(on) temperature model, the measured factor of about 2.5 between −40 °C and 125 °C for the cited FET, and the measured threshold drift (approximately 23 A at −40 °C and 8 A at 125 °C for a 15 A design without compensation).
  • Component-level figures used in the illustration are chosen to demonstrate the arithmetic and are not specifications of any commercial protection board.
  • Ratings quoted by board vendors vary in conditions; always record the datasheet revision, ambient temperature and thermal mounting with any figure you adopt.

Before publishing a number for your own assembly, re-measure it: run the board at the candidate current inside the intended enclosure until temperatures stop rising, and confirm the protection thresholds at the temperature extremes the product will see.