Equipment / lighting

Driver and Efficiency of Caving Headlamps

Technical guide to caving headlamp efficiency: 1S to 4S battery architectures, buck/boost/linear drivers, and their real impact on runtime and carried weight.

  • driver
  • efficiency
  • runtime
  • battery
  • comparison

Why efficiency matters in real use

Higher efficiency means more useful light for the same onboard energy. In practice, it avoids carrying “dead” battery mass and extends useful runtime on levels that are actually used underground.

Typical efficiency by driver technology

When architecture allows it, pure buck or pure boost usually provides the best peak efficiency. Buck-boost is very useful over wide voltage ranges, but often loses a few points at the best operating point.

Driver technologyTypical useful range in headlampsPossible maximum (best point)Current behaviorPractical reading
Pure buck~90-96%up to ~98%Low: fixed losses stand out. Mid: best efficiency zone. High: thermal and conduction losses rise.Excellent when battery voltage stays above LED forward voltage plus margin.
Pure boost~88-95%up to ~97%Low: fixed losses hurt efficiency. Mid: often very good. High: thermal and conduction losses increase.Very good when battery voltage is below target LED voltage.
Buck-boost (non-inverting)~85-94%up to ~97%Low: efficiency is more modest. Mid: good compromise. High: may drop earlier than pure buck/boost depending on design.Very flexible across voltage, usually a bit less efficient than a well-sized pure buck/boost in daily use.
Linear~65-90%close to Vf/VbattEfficiency mostly follows Vf/Vbatt ratio. As current and heat rise, losses become critical.Simple and clean regulation, but losses rise quickly with a larger battery/LED voltage gap.

Battery architecture and efficiency impact

Key point: increasing cell count in series at equal output lowers current. Lower current cuts losses across the chain (battery, cables, connectors, driver, thermal path), improving global efficiency and stability.

Battery architectureTypical matching driversTypical efficiency rangeWhat changes underground
1S (1 cell: 18650/21700)Boost, buck-boost, sometimes linear per channel~75-95%Simple logistics. Good weight/runtime compromise when regulation is well designed.
2S (2 cells in series or equivalent pack)Mainly buck, sometimes linear~80-96% in switching, more variable in linearLower battery current at equal output: fewer losses and easier thermal management.
3S (3 cells in series)Mainly buck~85-96%Even lower current, lower I2R losses system-wide. Effective for sustained high levels.
4S / multi-cell packBuck, more integrated architecture~86-96% when well implementedMinimum current-related losses at equal output, excellent long-run stability; tradeoff is pack weight/cost.

LED count impact on efficiency

Optical/electrical choiceEfficiency effectField consequence
Single LED driven hardLm/W drops quickly at high LED current, then drops again with heat.Faster stepdown and shorter useful runtime at high mode.
Two LEDs sharing current (flood/spot or mixed)Each LED runs at lower current: better lm/W and slower thermal degradation.Real gain: better sustained intermediate levels and more stable runtime.
Well-driven multi-LED setupCan deliver best system efficiency if channels and cooling are well controlled.Excellent versatility and global efficiency, otherwise gains shrink when driver/thermal design is limited.

Model details (comparison driver and efficiency column)

ModelPack / battery architectureDriver (current state)Retained efficiency rangeTechnical rationaleRuntime / weight impact
Stoots Yeti1x21700 (1S)Regulated switching (manufacturer claim)eta ~85-92%Stoots publishes ~85% (PWM boost) and up to 92% on combined architecture. Lower bound kept to cover non-optimal operating points.Very good useful-energy to carried-mass ratio.
Scurion 1500 caving4-cell pack (multi-cell)Switching (published as “up to”)eta ~88-94%Scurion publishes >94% (upper bound). Lower bound estimated for real operating points away from optimum.Very energy-efficient, but heavier total system.
Meandre Prowide 4.53 cells in series (3S)Likely switchingeta ~85-92%3S pack favors efficient buck regulation, but no detailed published driver curve.Good sustained output, noticeable system weight.
Petzl DUO RLProprietary multi-cell packLikely switching regulated drivereta ~78-90%Petzl documents regulated output but does not publish detailed per-mode driver efficiency: conservative estimated range.Strong runtime, with pack weight and cost to account for.
Armytek Wizard C2 Pro Max LR1x21700 (1S)Likely switchingeta ~75-88%No detailed manufacturer efficiency curve on the product page: estimated range from regulated 1S architecture.Good global efficiency/price, low carried mass.
Phaeton Dual2x18650 (often 2S depending on setup)Linear (retained main case)eta ~65-90%With linear regulation, efficiency mostly follows Vf/Vbatt, so it varies with mode and state of charge.Can regulate smoothly, but thermal losses are higher at strong output.
Argolamp 2.02x21700 (2-cell pack)Linear (retained main case)eta ~65-90%Same logic as Phaeton in linear mode: efficiency depends on pack voltage vs LED voltage.Good capacity runtime, but real efficiency varies with selected mode.
Sofirn HS201x18650 (1S)Mixed architecture depending on channels (order of magnitude)eta ~60-80%Budget multi-channel positioning: efficiency varies more with mode and thermal behavior.Light and affordable, but useful runtime at high mode is more limited.