The UC3843BN is a current-mode PWM controller for switching power supplies, packaged in a Minidip-8, that drives an N-channel MOSFET directly from a totem-pole output stage built to source or sink high peak current. Its defining characteristic within the UC384xB family is a low undervoltage-lockout window: it starts at a typical 8.4 V (min 7.8 V, max 9.0 V) and keeps running down to a typical 7.6 V after turn-on, in contrast to the UC3842B/UC3844B siblings whose 16 V-on/10 V-off window the datasheet flags as ideally suited to off-line applications. That makes the UC3843BN the natural pick for DC-DC converters fed from low or intermediate voltage buses. It also runs to duty cycles approaching 100% (max duty typ 96%, min 94%), with no internal toggle flip-flop capping it near 48% like the 3844B/3845B variants. The oscillator is guaranteed at 250 kHz, with current-mode operation specified to 500 kHz. If you're building a boost, flyback, or forward converter where the supply rail is modest and you need full duty-cycle range, this is the part of the family aimed at you.
The rest of this post is structured as the questions you'd actually ask when bringing the UC3843BN into a design: what it's for, how to wire it, what will bite you, how to prove it works on the bench, and why it might be misbehaving.
What is the UC3843BN actually for, versus its siblings? (Which use cases is the UC3843BN intended for?)
The datasheet grades the UC384xB family explicitly, and the UC3843BN sits in a specific slot:
- Off-line versus low-voltage grading. The UC3842B/UC3844B parts have a 16 V-on/10 V-off UVLO window that the datasheet describes as "ideally suited off-line applications." The UC3843B/UC3845B grade — which the UC3843BN belongs to — uses 8.5 V on / 7.9 V off instead, which is what makes it the low-start-voltage choice for converters supplied from lower-voltage buses rather than rectified mains. Note that these 8.5 V / 7.9 V figures come from the datasheet's narrative DESCRIPTION section, whereas the electrical-characteristics table quotes the same thresholds as typ 8.4 V (min 7.8 V, max 9.0 V) on and typ 7.6 V off; the datasheet presents both sets in different sections without explicitly reconciling the rounded versus typical values, so treat 8.5/7.9 as the rounded description and 8.4/7.6 as the table's typical figures for the same window.
- Full-duty-cycle versus toggle-limited. The UC3843B can run duty cycles approaching 100% (typ 96%, min 94% max duty). The 3844B/3845B variants insert an internal toggle flip-flop that caps duty around 48%. If your topology needs high duty — a boost converter stepping up aggressively, for instance — the UC3843BN is the right sibling; the toggle parts are not.
- Direct N-channel MOSFET drive. The totem-pole output is described as designed to source or sink high peak current, low in the off-state, and suitable for driving N-channel MOSFETs. The current-sense comparator provides cycle-by-cycle current limiting, so peak-current-mode control comes built in.
One practical note the datasheet asserts less explicitly: the low 8.4 V start threshold is what practically suits this part to DC-DC converters on low/intermediate buses, while the off-line statement is reserved for the 16 V parts.
How do I integrate the UC3843BN into a design? (Wiring, oscillator, and bias supply)
Three design areas matter most: the oscillator timing, the start-up bias, and the gate-drive interface.
Oscillator: pick RT/CT from the part's own curves. The datasheet's Figure 2 (Timing Resistor vs. Oscillator Frequency, Vi = 15 V, TA = 25 °C) gives concrete anchors:
- 100 kHz → RT ≈ 10 kΩ with CT = 1 nF
- 50 kHz → RT ≈ 20 kΩ with CT = 1 nF
- 100 kHz → RT ≈ 1 kΩ with CT = 5 nF
- 500 kHz → RT ≈ 1 kΩ with CT = 200 pF
- As CT increases, the RT needed for a given frequency decreases.
In short, you have two knobs and one frequency target, and the trade-off between them also sets your maximum duty cycle: the Maximum Output Duty Cycle vs. Timing Resistor curve (Figure 5) is where you confirm the duty headroom at your chosen RT. The Current Sense Threshold vs. Error Amp Output curve (Figure 7) shows the same relationship from the other side — for example, RT = 2 kΩ gives D_max ≈ 60–70% and RT = 5 kΩ gives ≈ 82–92%, depending on the 7.5 vs 8.8 mA discharge-current curve (Vi = 15 V, CT = 3.3 nF, TA = 25 °C). Check both before committing to a layout.
Start-up bias: budget around the UVLO window and the supply currents. The electrical-characteristics table gives you the numbers that shape the VCC start-up network:
- Start threshold: typ 8.4 V, min 7.8 V, max 9.0 V
- Minimum operating voltage after turn-on: typ 7.6 V, min 7.0 V, max 8.2 V
- Start-up current: typ 0.3 mA, max 0.5 mA (tested at Vi = 6.5 V)
- Operating supply current: typ 12 mA, max 17 mA (VPIN2 = VPIN3 = 0 V)
- Internal zener: 30–36 V
The design consequence: during the pre-start phase the part draws well under 0.5 mA, so a start-up resistor feeding the VCC capacitor can be large enough to charge to the 8.4 V threshold without burning much idle power. Once switching begins, the auxiliary winding must supply the 12–17 mA operating current and hold VCC above the ~7.6 V turn-off point. Size the hold-up so transients on the aux rail don't dip below that.
Gate drive: respect the output stage ratings. The output delivers VOH typ 13.5 V (min 13 V) at 20 mA source and typ 13.5 V (min 12 V) at 200 mA source; VOL is typ 0.1 V / max 0.4 V at 20 mA sink and typ 1.6 V / max 2.2 V at 200 mA sink. Rise and fall times are typ 50 ns (max 150 ns) with CL = 1 nF. Size the external gate resistor so peak gate-charge current stays within the 200 mA-class source/sink ratings given those ~50 ns edges — a direct-drive MOSFET with a large gate charge will pull harder than the output is rated for, and the energy delivered into a capacitive load is separately capped (see the next section).
What will bite me? (Critical limits and parameter sensitivities of the UC3843BN)
The absolute maximum ratings — all voltages referenced to pin 5 — are the hard walls:
- Supply voltage (Vi): 30 V max from a low-impedance source. The alternate rating says supply with Ii < 30 mA is "self-limiting" — a resistor-fed VCC self-limits, but a stiff rail must never exceed 30 V.
- Output current (Io): ±1 A max.
- Output energy into capacitive load (Eo): 5 µJ max — a caution against dumping excessive gate-capacitance energy into the totem-pole output.
- Analog inputs, pins 2 and 3 (error amp / current sense): −0.3 to 5.5 V. Both input pins must stay inside this window.
- Error amplifier output sink current: 10 mA max.
- Power dissipation (Ptot) at Tamb ≤ 25 °C: 1.25 W for the Minidip (the UC3843BN package); 800 mW for SO8.
- Temperatures: storage −65 to 150 °C; junction operating −40 to 150 °C; lead soldering (10 s) 300 °C.
The headline numbers to carry in your head: VCC on a stiff rail stays under 30 V, the pin 2/3 analog inputs stay within −0.3 to 5.5 V, output current within ±1 A, and the Minidip thermal budget is 1.25 W.
Beyond the absolute limits, two parameter sensitivities deserve bench attention across your full operating range:
- Switching frequency is temperature-dependent. The oscillator discharge current is ≈ 8.4 mA from −25 to 0 °C and falls to ≈ 7.5 mA at 100 °C (Figure 6). Since that current sets the oscillator discharge rate, your switching frequency drifts with temperature — verify frequency at your hot-end temperature, not just at 25 °C.
- Output saturation at heavy load. At 600 mA sink the output saturates around 3.5 V down from the rail (25 °C), and source saturation is ≈ −2.2 V (Figure 12). If you drive a heavy gate, confirm the resulting on/off drive levels still meet your MOSFET's threshold margin.
How do I verify the UC3843BN on the bench? (Testing and verification procedure)
A short, ordered bench sequence proves the part is what you think it is and behaves as specified.
1. Verify the UVLO thresholds — the signature check. With pin 5 grounded and pins 2 and 3 held at 0 V, slowly ramp VCC (pin 7) from 0 V up through ~10 V while watching the output. The part must stay off until VCC crosses the ~8.4 V start threshold (output low, saturating near 0.1 V / 1.1 V max at 1 mA sink per the VOLS spec at VCC = 6 V), then begin switching. Ramp VCC back down: it must keep running until ~7.6 V, confirming the hysteresis. Measure both transition points and confirm they sit inside the min/max windows (7.8–9.0 V start, 7.0–8.2 V turn-off). This is the single most effective "is this really a UC3843BN and not a 3842/3844" sanity check, since the 842B/844B parts start at ~16 V.
2. Verify frequency and duty-cycle capability. Measure the output frequency against your RT/CT target, and confirm maximum duty approaches the ~96% typical figure rather than the ~48% you'd see on a toggle-limited 3844B/3845B variant.
3. Check slope compensation at high duty. Because this part runs above ~50% duty, current-mode control needs slope compensation to stay stable — which is precisely why the datasheet documents an application circuit for it: the "Isolated MOSFET Drive and Current Transformer Sensing" circuit (Figure 19) shows two slope-compensation variants feeding pin 3 (ISENSE), one with an external slope resistor and one with an internal transistor for automatic slope compensation. Use that topology as your starting point; the exact slope magnitude is a loop-design value, not a datasheet constant.
4. Verify cycle-by-cycle overcurrent behavior. The part provides "latching PWM for cycle-by-cycle current limiting," and the current-sense input on pin 3 terminates the output switch conduction; the Current Sense Circuit figure (Figure 17) documents a sense resistor Rs monitoring inductor current feeding the comparator. On the bench, drive the output into a short and verify each PWM cycle terminates when the pin-3 sense voltage reaches the comparator threshold — the pulse width pulls back cycle-by-cycle rather than the part latching off and staying off (which would instead suggest UVLO collapse). Confirm it returns to normal regulation once the fault is removed.
5. Watch supply current throughout. A healthy unit draws less than 0.5 mA in UVLO lockout and 12–17 mA once switching. An out-of-family reading is a quick flag that something is wrong before you dig deeper.
Known limitation: the exact current-sense comparator trip threshold voltage is not stated in the datasheet segments covered here, so size Rs against the comparator threshold using the datasheet's Figure 17 setup rather than a quoted number from this post.
Why is my UC3843BN circuit misbehaving? (Troubleshooting)
When a UC3843BN converter misbehaves, the failure usually traces back to one of the limits or sensitivities above. Work through these in order:
- Part won't start. Measure VCC at pin 7 while the supply ramps. If it never crosses ~8.4 V, your start-up resistor is too large for the leakage path or the VCC cap is loading it excessively — remember the part only draws < 0.5 mA pre-start, so the resistor has headroom, but the threshold is a hard gate. If VCC starts the part but then collapses, your auxiliary winding isn't supplying the 12–17 mA operating current before the hold-up cap discharges to the ~7.6 V turn-off point.
- Part starts and immediately shuts down. Check whether VCC is dipping below the 7.6 V minimum operating voltage when switching begins — the operating current jumps from sub-milliamp to 12–17 mA the instant it starts, and an undersized aux supply will starve it. This is distinct from an overcurrent latch; watch VCC and the output together to tell them apart.
- Frequency is off, or drifts as the board heats up. Re-measure at your hot-end temperature. The oscillator discharge current falls from ≈ 8.4 mA to ≈ 7.5 mA between 0 °C and 100 °C, so some frequency shift with temperature is expected behavior, not a fault. If the frequency is wrong even at 25 °C, re-check RT/CT against the Figure 2 curves.
- Duty cycle is capped lower than expected. Your RT choice sets maximum duty via the Figure 5 and Figure 7 curves (e.g. RT = 2 kΩ → ≈ 60–70%; RT = 5 kΩ → ≈ 82–92%). If you need more headroom, adjust RT — you are not fighting a 48% toggle limit on this part, so a low duty ceiling is a resistor choice, not a device property.
- Subharmonic oscillation or instability at high duty. This is the classic current-mode symptom above ~50% duty; add slope compensation per the datasheet's Figure 19 topology if you haven't already.
- MOSFET runs hot or doesn't fully enhance. Check the output saturation figures: at 600 mA sink the output sits ~3.5 V below the rail, and at 200 mA sink VOL can reach 2.2 V. A heavy gate load through a large gate resistor pushes you into these saturation regions — reduce gate charge demands or add a driver stage.
- Output stuck low with VCC healthy. Verify pins 2 and 3 are within the −0.3 to 5.5 V analog-input window. A voltage outside that range on the current-sense pin will terminate output conduction via the comparator; a pin driven negative can hold the part off entirely.
- Supply current way out of family. If I_CC is far from the < 0.5 mA lockout / 12–17 mA operating figures, suspect a damaged part or a shorted load on the output — both show up as abnormal supply current before anything else does.
Part page: UC3843BN.