FAQ
TL;DR: USB 5V can dip to 4.75V, and LM317L dropout is ~1.6–1.9V; “Looks like part of your problem might be Dropout Voltage.” [Elektroda, Steve Lawson, post #21669399]
Why it matters: If you power LM317L from USB, dropout plus USB tolerance can cap output near 3.0V, not 3.3V.
- USB 5V tolerance is ±5%; designing for 3.3V needs ≤1.45V dropout from a 4.75V minimum. [Elektroda, Steve Lawson, post #21669399]
- 18Ω+18Ω divider at 3.3V draws about 92 mA; it consumes most of a 100 mA LM317L. [Elektroda, Steve Lawson, post #21669400]
- LM317L dropout typically ~1.5–1.9V near 100 mA; verify by measuring Vin–Vout. [Elektroda, Steve Lawson, post #21669399]
- Overheating can trigger thermal limiting and sag the output voltage. [Elektroda, Steve Lawson, post #21669399]
- Prefer an LDO or separate fixed 3.3V/1.6V regulators for better headroom. [Elektroda, Steve Lawson, post #21669409]
Quick Facts
- USB 5V tolerance is ±5%; designing for 3.3V needs ≤1.45V dropout from a 4.75V minimum. [Elektroda, Steve Lawson, post #21669399]
- 18Ω+18Ω divider at 3.3V draws about 92 mA; it consumes most of a 100 mA LM317L. [Elektroda, Steve Lawson, post #21669400]
- LM317L dropout typically ~1.5–1.9V near 100 mA; verify by measuring Vin–Vout. [Elektroda, Steve Lawson, post #21669399]
- Overheating can trigger thermal limiting and sag the output voltage. [Elektroda, Steve Lawson, post #21669399]
- Prefer an LDO or separate fixed 3.3V/1.6V regulators for better headroom. [Elektroda, Steve Lawson, post #21669409]
Why does my USB-powered LM317L only output ~3.0V instead of 3.3V?
Because USB can be as low as 4.75V and LM317L’s dropout near 100 mA is ~1.6–1.9V. That leaves only ~3.15V headroom, so the regulator cannot reach 3.3V. Measure Vin–Vout to confirm dropout on your board. “Looks like part of your problem might be Dropout Voltage.” [Elektroda, Steve Lawson, post #21669399]
How much current do two 18Ω resistors in series (as a divider from 3.3V) draw?
About 92 mA: I = 3.3V / (18Ω + 18Ω) = 0.0917 A. That quiescent load already uses most of an LM317L’s 100 mA capability, leaving little for your actual circuit. This heavy divider also increases dropout and heat. [Elektroda, Steve Lawson, post #21669400]
What is dropout voltage in plain terms?
Dropout is the minimum difference between input and output for a regulator to regulate. With USB min at 4.75V, a 3.3V rail needs ≤1.45V dropout. LM317L often needs ~1.6–1.9V near 100 mA, so it falls out of regulation from USB. [Elektroda, Steve Lawson, post #21669399]
How do I measure dropout on my existing PCB?
Load the 3.3V rail as in use, then measure Vin and Vout pins. Dropout equals Vin minus Vout. If this value rises near 1.6–1.9V at your load, the LM317L can’t hold 3.3V from USB’s 4.75–5.25V. [Elektroda, Steve Lawson, post #21669399]
Why did the LM317L get painfully hot when powered from 7V?
Linear regulators burn the voltage difference as heat: P ≈ (Vin−Vout)×I. Heat can trigger internal thermal limiting, which lowers output voltage until the die cools. That looks like random sag but is self-protection. [Elektroda, Steve Lawson, post #21669399]
Should I reduce R1 from 1.2 kΩ to 470 Ω on the LM317L?
Lowering R1 can improve reference behavior at low input headroom, aligning with datasheet guidance. Using 470 Ω is a common starting value for LM317 reference current. Recalculate R2 to set your target voltage afterward. [Elektroda, Malcolm Whinfield, post #21669396]
Is a resistive divider a good way to derive 1.6V from 3.3V?
Not with low resistor values. Your 36Ω divider wastes ~92 mA and starves the regulator and load. Use a dedicated 1.6V regulator or redesign the divider with high values if the 1.6V node draws almost no current. [Elektroda, Rohit Dubla, post #21669397]
What regulators should I pick for 3.3V and 1.6V rails?
Use an LDO or fixed 3.3V regulator, and a separate adjustable/fixed LDO for 1.6V. Options suggested include 78L33A, MCP1702, LP2950-33, or LP2951-ADJ; check capacitor requirements and current limits. [Elektroda, Steve Lawson, post #21669409]
Can I trust Multisim or other simulators for regulator behavior?
Simulators won’t enforce real device limits like dropout, thermal limits, and max current. Always read the datasheet and breadboard critical cases. “The importance of going through the datasheet cannot be understated.” [Elektroda, Rohit Dubla, post #21669404]
How do I redesign using two regulators from a 9V, 700 mA adapter?
Feed both regulators from 9V, one set to 3.3V and the other to 1.6V. Ensure the adapter’s current covers both loads with margin, and that each IC’s current limit exceeds its rail draw. [Elektroda, Rohit Dubla, post #21669408]
What current does the ‘black box’ actually draw at 3.3V and 1.6V?
Measured values were Icc ≈ 73 mA at 3.3V and Iac ≈ 2.17 mA at 1.6V. That implies about 45Ω and 737Ω effective resistances respectively (V/I). Size regulators and thermal design for these loads. [Elektroda, John S, post #21669405]
Quick 3‑step: how do I diagnose low LM317L output from USB?
- Measure USB input under load; note the minimum it reaches.
- Measure Vin and Vout at the regulator; compute dropout.
- Measure load current and regulator temperature; watch for thermal limiting. [Elektroda, Steve Lawson, post #21669399]
Will a linear regulator here be inefficient? Does it matter?
Linear efficiency is limited by (Vout/Vin). Large Vin−Vout becomes heat, so heatsinking or lower Vin helps. If efficiency is not a requirement, simplicity may win; otherwise consider LDOs or switching options. [Elektroda, Malcolm Whinfield, post #21669410]
Could thermal limiting be the hidden failure mode on my board?
Yes. If the LM317L overheats, it reduces output current or voltage to protect itself. This looks like droop under load and recovers as temperature falls. Improve cooling or reduce Vin−Vout and load. [Elektroda, Steve Lawson, post #21669399]
