← Back to: Battery Charging / Gauging / Protection / BMS
Problem & Scenario: Why Battery Needs Ideal Diode / Power Mux
This section focuses on battery + external power source coexistence, not a single-battery-to-DC/DC case. When an adapter or USB-C source is present, the system should draw from it; when it disappears, the system must fall back to the battery without brown-out; and at no time should the system rail backfeed into the adapter, automotive line, or solar input. This is a power-path arbitration problem, not a charging state-machine problem.
1) Adapter / USB-C first
When an external source is plugged in, the system must run from it, not from the battery. Battery stays chargeable / on standby.
2) Seamless fallback
If the external power disappears, the path must move to the battery fast enough so MCUs, radios, or storage do not reset.
3) No backfeed to the source
System or battery must not backfeed the adapter/vehicle/solar port. This is where reverse blocking becomes mandatory.
4) Charge while powering
Even when the battery is being charged, the system must still source from the adapter. Avoid the “charge–discharge ping-pong”.
Why classic Schottky OR-ing fails in battery systems
- High forward drop → heat & lost headroom: with a 3.7–4.2 V Li-ion, every 200–300 mV matters.
- No real priority control: whichever diode sees the higher voltage conducts; you can’t “force” adapter over battery.
- Backfeed still possible: in multi-rail or poorly sequenced designs, the return path can still sneak current backwards.
Why a simple load switch is not ideal
- Often one-way only: many load switches do not provide a real reverse-blocking path.
- No dual-input arbitration: when adapter and battery are present together, the switch doesn’t “choose”.
That gap is exactly where an ideal-diode controller or power-mux IC makes sense: it gives you low-drop conduction, direction awareness, and priority arbitration at the same time.
Topology Options for Battery Power-Path
This chapter compares legacy diode OR-ing, FET-based ideal-diode (controller + MOSFET), true two-input power-mux to a system bus, and back-to-back FET reverse-blocking for battery-centric designs. We only discuss the battery-side power mux, not USB data multiplexing, not Type-C PD policy engines, and not pack-level redundant FET trains.
1) Diode OR-ing (legacy)
Simple, cheap, but high drop and no real priority. OK for low-current backup batteries.
Reverse blocking: weak · Priority: none · Drop: high
2) FET OR-ing / Ideal-diode Ctrl
Controller senses ΔV and drives an external MOSFET to get very low loss.
Reverse blocking: good · Priority: voltage-based · Drop: low
3) Two-input Power Mux
True dual-source → one system rail. Perfect for adapter + battery coexistence.
Reverse blocking: good · Priority: fixed/EN · Drop: low
4) Back-to-back FET
Use when both sides must be protected from backfeed (system ↔ battery).
Reverse blocking: strong · Priority: external · Drop: depends on FET
Why battery-side mux must watch voltages
External adapters are often higher than the Li-ion pack, so “choose the higher source” works. But in solar or vehicle inputs, voltage may swing below the battery level. In that case, the mux or ideal-diode controller must compare voltages and hold the battery active until the external rail is truly higher or enabled.
Likewise, putting the ideal-diode on the battery side (not on the system side) keeps the charging path open; placing it wrongly may block legitimate charge current.
Priority Arbitration Strategies (Adapter > Battery > Others)
In this section, we explore various power-path arbitration strategies to determine which power source (adapter or battery) gets used first. We’ll cover three key strategies: Fixed-priority, Voltage-priority, and MCU/EN-controlled priority.
Fixed-priority (Adapter > Battery)
Used in typical consumer and industrial applications where external power (adapter/USB-C) must be prioritized over battery power to ensure the battery stays in charging or floating mode.
Voltage-priority (Higher VIN wins)
Used in systems with fluctuating input sources, such as solar or automotive power. External sources are preferred when their voltage is higher than the battery voltage.
MCU/EN-controlled priority
Suitable for automotive or more complex BMS systems, where an MCU or EN (enable) pin is used to control the priority of power sources based on system needs.
Why should external power always have higher priority in charging scenarios?
In charging applications, the system must ensure that the battery is kept in a charging or float mode, without being drained while charging. Allowing battery power to supply the system during charging would lead to inefficient charging cycles and potential damage to the battery.
By prioritizing the external source (like an adapter or USB-C), we can keep the system running without draining the battery, allowing the battery to focus on charging rather than discharging.
The Relationship Between Priority and VSYS/VBAT
When the external power is selected as the priority, the system voltage (VSYS) is more stable, allowing for accurate current control and proper charging of the battery. This ensures that the system can handle fluctuating input sources without compromising performance.
Reverse Blocking & Backfeed Protection in Battery Systems
In this section, we explain the causes of backfeed, three critical scenarios where reverse blocking is essential, and how ideal diodes and back-to-back FETs can prevent damage caused by backfeed. Additionally, we will explain how ideal diodes react to reverse voltage and ensure seamless operation.
Why Backfeed Happens
Backfeed occurs when the system or battery voltage exceeds the input voltage, or when the input power is disconnected but still connected to the system. In these cases, the system could inadvertently send power back into the input source, potentially damaging the upstream devices.
Three Critical Scenarios Needing Reverse Blocking
- USB/Adapter Source (from PC or Car): Backfeed can damage upstream devices like USB ports or car circuits.
- Solar Panel Disconnection: If solar power is disconnected, we must prevent battery voltage from backfeeding into the solar panel.
- Multi-board System: One board losing power must not supply voltage back into other boards to avoid system-wide failure.
Single FET vs Back-to-Back FET
A single FET can only conduct in one direction, meaning it cannot block reverse current. In contrast, back-to-back FETs are connected in a way that they block reverse current effectively, making them suitable for protecting both the system and the battery from backfeed damage.
How Ideal Diode Controllers Detect and Block Reverse Current
Ideal diode controllers monitor the voltage difference across the device. If reverse voltage is detected, the controller immediately turns off the MOSFET, thus preventing backfeed from reaching the system or the battery.
Low-Drop, Fast-Switch, and Thermal Considerations
In battery + adapter dual-source systems you cannot “waste” voltage on the power-path: the cell is already drooping from 4.2 V toward 3.0 V, so any series element must look like an ideal-diode-controlled FET, not a Schottky. That is why devices such as TPS2121, TPS2115A-Q1, and LM74700-Q1 regulate the forward drop to a few tens of millivolts when the FET is fully enhanced. :contentReference[oaicite:0]{index=0}
The real loss is ILOAD × RDS(on) of the pass FET. To keep the system rail alive at brown-out edges, pick parts that ① support your max current without going above your ΔV budget, ② have fast break-before-make switchover to the battery, and ③ specify thermal performance (θJA) for the package you actually solder. If the adapter is high and you didn’t put a front DC-DC, the mux will run warm — that’s normal, but it must be predictable. :contentReference[oaicite:1]{index=1}
A good engineering note to leave in the BOM is: “RDS(on) ≤ XX mΩ at VGS=5 V; do not replace with higher-drop variants or battery rail may brown-out during adapter removal.” That line is what stops purchasing from silently dropping in a cheaper, hotter switch.
Quick drop & thermal check
1) Voltage drop: ΔV = ILOAD × RDS(on)
2) Loss: P ≈ ILOAD2 × RDS(on)
3) Estimated rise: ΔT ≈ P × θJA
4) If ΔT too high → use external, lower-RDS(on) MOSFET, or parallel FETs, or step up the package.
Small-Batch Procurement & Cross-Brand Alternatives
For ideal-diode / power-mux ICs the distribution is unbalanced — TI and ST have the richest automotive and high-current portfolio (true priority mux, reverse-blocking controllers), NXP and onsemi have very strong USB/low-voltage switches that already integrate TRCB, and Renesas / Microchip cover the high-reliability OR-ing space with controller + external MOSFET recipes. Melexis doesn’t ship a ready-made power mux here, so we treat its reverse-polarity capable SBC / driver as the L3 fallback front-end. :contentReference[oaicite:2]{index=2}
TI (primary)
When you must do adapter > battery priority with low drop and automotive Q1:
- TPS2121 — 2.7–22 V, 4.5 A seamless power mux. :contentReference[oaicite:3]{index=3}
- TPS2120 — 3 A version, same family. :contentReference[oaicite:4]{index=4}
- TPS2115A-Q1 — 2.8–5.5 V auto-switching mux for adapter/battery. :contentReference[oaicite:5]{index=5}
- LM74700-Q1 — 3.2–65 V ideal diode controller, 20 mV fwd drop. :contentReference[oaicite:6]{index=6}
- LM7480-Q1 / LM74801-Q1 — back-to-back FET drive if you need reverse + OV. :contentReference[oaicite:7]{index=7}
ST
For 5 V / 12 V lines needing reverse-current blocking and eFuse-style control:
- STEF12 / STEF12S — 12 V electronic fuse with reverse current blocking. :contentReference[oaicite:8]{index=8}
- STPW12 — programmable power breaker, sits in series to the rail. :contentReference[oaicite:9]{index=9}
- Hot-swap / eFuse portfolio (STEF01, STEVAL-EFUSE…) for higher power. :contentReference[oaicite:10]{index=10}
NXP
USB-PD / Type-C class parts — many already have reverse current protection:
- NX5P2090 — 5.5 V / 2 A high-side with RCP and inrush control. :contentReference[oaicite:11]{index=11}
- NX5P3290 — adjustable CL, reverse current protect, 29 V tolerant. :contentReference[oaicite:12]{index=12}
- NX20P5090 — 5 A uni-directional PD switch, can be paralleled for dual input. :contentReference[oaicite:13]{index=13}
Renesas
Controller + external MOSFET path, good for higher currents:
- ISL6146 / ISL6146A — 1–18 V OR-ing controller for low-loss diode replacement. :contentReference[oaicite:14]{index=14}
- (pair with) automotive N-MOSFET <5 mΩ to hit 3–5 A battery rails.
onsemi
Portable / automotive DISO load switches with TRCB:
- FPF1048B — 1.5–5.5 V, 3 A, true reverse current blocking (TRCB). :contentReference[oaicite:15]{index=15}
- FPF1320 / FPF1321 — dual-input single-output power mux, break-before-make. :contentReference[oaicite:16]{index=16}
- NCV68061 — ideal-diode NMOS controller for automotive battery inputs. :contentReference[oaicite:17]{index=17}
Microchip
SiPoE / high-voltage ideal diode bridge → very good for “any polarity in” front-end:
- PD70224 — FET-based ideal diode bridge, up to 2 A. :contentReference[oaicite:18]{index=18}
- PD70288 — newer dual ideal-bridge, 100 V rating, PD types 1–4. :contentReference[oaicite:19]{index=19}
- Use with external N-MOSFET when battery current >2 A or you need VSYS priority.
Melexis (fallback)
Melexis is sensor / LIN SBC centric, not a power-mux vendor. Use its reverse-polarity-tolerant front-ends + external MOSFET for BMS demo boards:
- MLX92223 / MLX92242 — wide-range Hall switch, integrated reverse supply. :contentReference[oaicite:20]{index=20}
- MLX83100 + external FET stage — pre-driver with reverse-pol supply path. :contentReference[oaicite:21]{index=21}
- Note: mark BOM as “L3: Melexis + MOSFET, not true 2-input power mux”.
What to tell purchasing
Primary: TI / ST listed above, must support reverse blocking and seamless switchover.
Alternate: NXP / onsemi parts with TRCB and the same or lower RDS(on) at rated current.
Fallback: Renesas / Microchip controller + external MOSFET, or Melexis RP-front-end + MOSFET.
BOM remark: “Power path is critical for VSYS hold-up. Do not replace with high RDS(on) or non-TRCB parts. If TI/ST OOS, use NXP NX5P2090 / onsemi FPF1320 at equal or lower drop.”
Validation / Bench Test / BOM Remarks
In this section, we outline how to properly validate the ideal diode / power mux functionality of the device. This includes reverse blocking, switchover behavior, and thermal considerations under full load. After testing, we also discuss what to put in the BOM to ensure correct parts are procured and how to document the test results.
Bench Test Setup
The test setup consists of two adjustable power supplies and a system load (MCU/Wi-Fi or electronic load):
- PSU #1: Adapter (5–20 V adjustable)
- PSU #2: Battery Emulator (2.5–4.2 V adjustable)
- System Load: Electronic load or MCU/Wi-Fi module (to simulate typical system behavior)
- Measurement Equipment: Oscilloscope (to monitor VIN, VSYS, Gate/PG), Temperature Probe or Thermal Camera
**Test Sequence**:
1) Start with the Adapter power source and monitor voltage on the system rail (VSYS).
2) Enable Battery power and observe if the priority works as expected.
3) Simulate power loss on Adapter and observe if the system switches to battery without issue.
4) Inject reverse current and verify the system shuts down (reverse blocking test).
Reverse Injection Test
Inject reverse current at the system side (VSYS) and observe if the device immediately shuts down to prevent backfeed to the input (VIN). The test should pass if the reverse blocking occurs within a few microseconds.
Simultaneous Power-Up/Down Test
Test the behavior when both power sources (Adapter and Battery) are present. The system should prioritize the adapter, and if the adapter loses power, the system should seamlessly switch to the battery. The test should confirm that there is no power oscillation or glitching in the system voltage.
Full-Load Drop and Thermal Test
Under full load (2 A / 3 A / 5 A), measure the voltage drop across the power mux and ensure it is within acceptable limits (e.g., <50 mV). Additionally, monitor the temperature rise of the power mux and MOSFETs to confirm that they are not exceeding safe operating temperatures.
BOM Remarks for Procurement
Ensure that the BOM remarks clearly specify the essential requirements for the power-path device. Here are the necessary entries:
- Functional Requirement: “Power-path device must support reverse blocking from system to input.”
- Switchover Time: “Switchover time shall maintain system rail within ±X% at Y A load.”
- Substitute Requirement: “Substitutes must keep package and Rds(on) the same class; do not replace with load switch without TRCB.”
This will prevent the procurement team from mistakenly replacing the ideal-diode power mux with a simple load switch without reverse blocking.
Documenting Test Results
After bench testing, it is essential to document the results and indicate the tested parts in the form of:
- Tested with: “TI TPS2121, LM74700-Q1, or onsemi FPF1320”
- Result: “Pass at 3 A, ΔV < 40 mV, no backfeed"
- Note: “Layout near FET must keep short loop to minimize voltage drop.”
For components that failed to meet the requirements, mention the specific failure (e.g., “Too hot under full load, needs thermal redesign”).
Request a Quote
FAQ (12)
The following questions focus only on this page's topic — battery dual-source power-path using ideal-diode / power-mux ICs. They do not cover generic eFuse topics or full BMS AFEs so that intent does not overlap with sibling pages.
1. Why do I need an ideal diode instead of a Schottky for battery power-path?
Because the battery rail is only 4.2 → 3.0 V, losing 300–500 mV on a Schottky wastes the usable SOC window. A FET-based ideal diode can hold the drop to tens of millivolts and still give you reverse blocking. If you cannot buy the ideal-diode part, pick from Section 6 the NXP / onsemi TRCB switches that keep low Rds(on).
2. How does the power mux decide between adapter and battery when both are present?
It uses priority arbitration: fixed-priority (adapter > battery), voltage-priority (higher VIN wins), or MCU/EN override. For charging scenarios you must put the adapter at higher priority, or the system will keep pulling from the battery while charging. Devices like TI TPS2121 and onsemi FPF1320 support this behavior directly.
3. Can I use a simple load switch IC and still get reverse blocking?
Usually no. Most load switches are one-directional and do not stop backfeed from the system to the input. You must pick devices explicitly marked “true reverse current blocking / TRCB”. If procurement only finds a load switch, add an external back-to-back FET as shown in the topology chapter.
4. What is the typical forward drop I should target for a 1-cell Li-ion system?
For 1-cell systems, target <50 mV at the rated load (2–3 A typical). This keeps the battery usable down to around 3.3 V without forcing the system to brown-out. Write this requirement into the BOM: “Forward drop ≤ 50 mV @ I = … A.” That will stop purchasing from using higher-drop substitutes.
5. How to prevent the system rail from backfeeding the USB-C/adapter input?
Place the ideal-diode / reverse-blocking stage on the input side, not on the battery side, so that when VIN collapses the path is shut within a few microseconds. Back-to-back FETs or controller + external MOSFET (TI LM74700-Q1 / Renesas ISL6146 + FET) are the common ways to do this.
6. Does the ideal-diode controller need an external MOSFET for higher current?
Yes, when you go above about 3–4 A, internal FETs are often not low-enough in Rds(on) or not thermally robust. In that case pick a controller-only part (LM7480, ISL6146, NCV68061) and choose your own MOSFET footprint to hit the current and thermal target. This is why in Section 6 we put “controller + external MOS” as the L3 fallback.
7. How fast should switchover be to avoid MCU brown-out?
For MCU / wireless / storage rails, keep the switchover in the low tens of microseconds, and make sure the system has a small hold-up capacitor. If your device needs millisecond-level switchover, document that in the BOM as “Switchover time shall keep VSYS within ±X%” so purchasing doesn’t pick a slower part.
8. Can I mix a TI charger IC with an ST power mux?
Yes, as long as the charger exposes a stable VSYS path and the power mux is placed on the external input side. Make adapter the higher-priority source so the charger sees a predictable input. If the ST part does not have native reverse blocking, pair it with external FETs as shown in the reverse-blocking chapter.
9. What to put in the BOM note so purchasing won’t replace it with a non-blocking switch?
Use wording like: “Power-path device must support reverse blocking from system to input. Substitutes must keep package and Rds(on) in the same class. Do not replace with load switch without TRCB.” This exactly matches the validation flow in Chapter 7.
10. How to test reverse blocking on the bench safely?
Use two PSUs (adapter + battery), set the system side slightly higher, and monitor current on the input side. A real ideal-diode / power-mux should shut immediately when VSYS > VIN. Always test with a current-limited source first to avoid damaging the device under test.
11. Are automotive-grade ideal-diode / power-mux parts mandatory for 12 V battery systems?
For production automotive yes, for small-batch validation no. You can accept industrial-grade TI/ST/NXP/onsemi parts if you have validated reverse blocking and switchover on the bench. Make this explicit in the BOM: “Automotive preferred, industrial acceptable for validation lots.”
12. How to scale from 2 A to 5–6 A without redesigning the whole path?
Move to a controller + external MOSFET architecture (TI LM7480-Q1, Renesas ISL6146, onsemi NCV68061) and pick a MOSFET with lower Rds(on) and better thermal pad. Keep the arbitration logic unchanged so firmware and charging stay identical. This is the scaling route described in Section 6.
