Lithium Battery for Film Lighting: Flicker-Free Power Budget
On a documentary shoot in Oman we lost forty minutes to a fault that was neither the fixture nor the camera. Six 300 W LED panels were pulling from a 26 V cart pack, and every time the gaffer brought them up together the whole rig blinked and the BMS latched off. On paper it was correct: 2 kWh, 70 A continuous, above the 69 A steady draw. What it had never been sized for was the first 200 milliseconds. Lighting is a fundamentally different load than a camera body, and most lithium battery mistakes on set come from treating the two as interchangeable. Here is how I specify a lithium battery for film lighting and portable rig work.
Why Lighting Loads Are Not Camera Loads
A camera draws in bursts. A body idles at 12 to 25 W, jumps to 40 or 60 W while recording and transmitting, then falls back. Between takes it recovers, so the average load over a shoot day is far below the peak. An LED panel does the opposite. Once the driver is regulating, it holds a near-constant current for hours, and the I2R heating inside the pack never gets a gap to dissipate.
That constant-current profile changes every sizing decision. Take a 300 W fixture. LED driver efficiency sits around 88 to 93 percent, so the pack must deliver roughly 333 W at the terminals. On a 14.4 V single-block lithium battery that is 23 A; on a 26 V two-block bus it is 12.8 A. Put that 300 W fixture on a 98 Wh pack and you get about 17 minutes of runtime, not the two hours the marketing sheet implies.
The energy budget, written out
- Fixture nameplate power is array power, not system power: add 8 to 12 percent for driver loss.
- Runtime equals usable watt-hours divided by actual draw, where usable is nameplate minus the 10 to 15 percent the BMS withholds at each end.
- A six-hour day with four 300 W panels is 7.2 kWh of fixture draw, about 8 kWh from the pack. That is a cart, not a shoulder bag.
- Duty cycle on set is often 60 to 70 percent, because lights come down between setups. Budget 70 percent and treat the rest as headroom for reshoots.
Flicker, Ripple and High-Frame-Rate Banding
Flicker complaints almost always land on the battery first, and they are almost never the battery. There are two independent sources, and it helps to separate them before you change hardware.
The first source is the fixture. Almost every LED panel dims by pulse-width modulation. At full output the duty cycle is near 100 percent and pulses are long, so a camera at 24 fps never sees them. Dim to 10 percent and the pulse collapses to a narrow slice of the switching period. Shoot that same fixture at 240 fps with a 180 degree shutter, an exposure of about 1/480 second, and two or three pulses can fall inside a single frame. The result is banding that scrolls. No battery fixes a PWM mixing problem with the shutter; flicker-free dimming modes and a camera test at your delivery frame rate do.
What the pack genuinely contributes
The battery contributes a slow variable, not a fast one. Pack voltage sags over seconds as state of charge falls and as the cells warm. A constant-current driver rejects input variation by 60 to 80 dB, so a 5 percent bus sag is invisible at the LED. Trouble starts only when the input approaches the driver dropout threshold. Then regulation fails, output dims abruptly, the driver may restart, and the symptoms look exactly like flicker to a gaffer.
Two habits prevent this. First, keep worst-case terminal voltage at least 3 to 4 V above the fixture dropout spec, measured under full load, not at rest. Second, check DC ripple at the fixture input rather than at the pack. Ripple above roughly 200 mV peak-to-peak, usually from a cheap downstream DC converter or a BMS with aggressive balancing pulses, can appear as faint banding on a high-speed shot. Diagnose with an oscilloscope on the DC lead and a photodiode at the panel, never a light meter, which averages away exactly the artefact you are chasing.
Cold-Start Inrush and the First 200 Milliseconds
This is the failure mode from the opening paragraph.
LED drivers place bulk capacitance directly across their DC input, typically 100 to 470 microfarads for a 100 to 300 W panel. A discharged capacitor is a short circuit. For the first few milliseconds after you close the switch, the pack sees an impedance limited only by cable resistance and cell internal resistance, and it sees a current five to ten times the steady value. Six 300 W panels on a 26 V bus draw 69 A steady. Energise them together cold and the instantaneous peak can reach 300 to 400 A for 10 to 20 milliseconds. A BMS with a fixed 100 A overcurrent trip reads that as a short and opens the contactor. The rig goes dark and the pack looks faulty.
Four ways out, in order of preference
- Stagger the power-up. Two seconds between fixtures is free, needs no hardware, and drops the peak to a single fixture’s inrush. Train the crew to obey it.
- Soft-start at the distribution point. An NTC inrush limiter or a DC-DC stage with a controlled ramp moves the peak off the battery entirely.
- Specify a two-stage BMS. Moderate overcurrent should hold for 500 milliseconds to a second before tripping; only a genuine short should open instantly. Ask for the trip curve, not the trip value.
- Size the pack for the peak, not the average. Keep the 10 millisecond peak under 60 percent of the instantaneous trip threshold.
Then verify it cold. Cell internal resistance roughly doubles at 0 degrees Celsius, so a pack that passes at 25 degrees Celsius can trip on a winter exterior. Run the chamber test at 0 degrees Celsius with the real fixture complement and cable lengths.
14.4 V, 26 V or 48 V: Bus Choice and Cable Loss
Voltage selection is a cable-loss decision disguised as a camera-accessory decision. Copper does not care about tradition.
Use 2.5 square millimetre cable and a four metre run, so eight metres of conductor, at 7.41 milliohms per metre. That is 0.059 ohms round trip. Now compare a 300 W fixture across three buses.
- 14.4 V: 20.8 A, 1.23 V dropped, 8.6 percent of the bus lost as heat in the cable. Twenty-six watts warming the copper.
- 26 V: 11.5 A, 0.68 V dropped, 2.6 percent lost.
- 48 V: 6.25 A, 0.37 V dropped, 0.8 percent lost.
Move to 4 square millimetre cable and the 14.4 V case improves to 5.3 percent, still worse than 26 V on thin cable. My rule: keep bus loss below 5 percent, which at 14.4 V means runs under 2.5 m with 4 square millimetre.
Connectors are the quieter bottleneck
Contact resistance of 5 to 10 milliohms per pin is normal, and it is a heater at current. A D-Tap or P-Tap is intended for roughly 8 to 10 A, so a 300 W fixture at 14.4 V needs 20.8 A and is well outside its comfort zone. Move high-draw panels to XT90, powerCON TRUE1 or a locking two-pin industrial connector rated above 30 A continuous, and keep D-Tap for monitors and small on-camera fixtures. Distribute from a central bus bar in a star, not a daisy chain; the last fixture in a chain sees the accumulated drop of every upstream joint plus its neighbours’ inrush.
Heat, Fan Noise and Weather on Location
Set life punishes batteries in ways a lab never does. A black anodised pack on a light stand in direct sun absorbs roughly 1000 W per square metre, and surface temperature can reach 70 to 75 degrees Celsius on a 40 degree day. Cell ageing roughly doubles for every 10 degrees of sustained temperature rise, and most BMS designs derate continuous current above 45 degrees, some as early as 40.
Mitigation is cheap. Shade the packs, keep them off hot ground and reflective surfaces, stand them vertically so convection works, and derate your continuous current expectation by 30 to 40 percent above 45 degrees ambient. If a fixture is 300 W and the day is brutal, budget the pack for 450 W.
Cold, rain and condensation
- At minus 10 degrees Celsius expect 20 to 30 percent less delivered capacity and two to three times the internal resistance. Precondition packs indoors to about 10 degrees before the call time.
- Never charge a lithium battery below 0 degrees Celsius. Charging cold is the fastest route to lithium plating and permanent capacity loss.
- For exterior night work, IP54 splash protection with capped connectors and a drip loop is usually enough. IP54 is not immersion; do not leave packs on a wet deck overnight.
- Moving from a cold truck into a warm stage creates condensation inside connector shells. Let packs equalise for 20 to 30 minutes before powering up.
Sound stage means no fan
On a sound stage the ambient target is often NC 20 to 25, and a 30 dBA fan at one metre is clearly audible on the boom. Specify fanless packs with conduction cooling and thermal mass, or accept that the pack must live off-set with a cable run. On a lithium battery pack carrying a continuous 300 W load, fanless means derating the continuous current and accepting a warmer case. Make that trade at specification time, not on the day.
Air Travel, Certification and Fleet Rotation
Lithium batteries for lighting travel constantly, and the watt-hour thresholds are what actually constrain pack design. Under IATA rules, spare batteries up to 100 Wh go in carry-on without airline approval; between 100 and 160 Wh they require approval and are limited to two spares per passenger. Standalone lithium battery cells and batteries shipped as air cargo, UN 3480, must be at no more than 30 percent state of charge. Most 98 Wh mount packs exist precisely because they sit just under the 100 Wh line, and 150 Wh blocks exist to sit under the 160 Wh line with an approval letter.
Documentation matters as much as the number. You need a UN 38.3 test summary covering the T1 to T8 series, and for these portable packs IEC 62133-2 is the right standard, since the batteries are portable and not traction packs. Keep terminals taped or capped in transit, and carry the test summary in the case, not in an email thread someone has to find at the check-in desk.
The rotation programme that ends surprise failures
- Label every pack with a serial and a commissioning date. No label, no rotation, no warranty claim.
- Capacity-test quarterly, or every 100 cycles, whichever comes first, and retire at 80 percent of nameplate.
- Assign the newest packs to the highest-drain fixtures. Wear tracks load, so match them deliberately.
- Store at 30 to 50 percent state of charge between 15 and 25 degrees Celsius and recheck every three months.
- Watch dark current. A smart pack with a display drawing 5 mA burns about 1.7 Wh per day, which is a 98 Wh pack flat in under two months on the shelf.
None of this is exotic. It is the ordinary discipline of matching a lithium battery pack to a load that never stops, on a day that will not wait. Get the inrush, the bus voltage and the thermal derating right at specification time, and the set stops being a test bench.
Frequently Asked Questions
Why does my LED fixture flicker on battery but not on mains?
It is usually the dimmer, not the pack. Pulse-width dimming at low output shortens the on-pulse until a high-speed camera can resolve individual pulses. A stiff mains supply hides nothing here either; the difference is often a different frame rate than you used in the studio. Test at your delivery frame rate before blaming the battery.
Can I run a 300 W fixture from a single 98 Wh mount pack?
Only for about 17 minutes, and probably not at all in practice. A 300 W panel needs roughly 333 W from the pack, which is 23 A at 14.4 V on a 98 Wh block, well beyond a D-Tap and close to the continuous limit of many small BMS designs. Two packs in series on a 26 V bus halves the current and is the honest answer.
What state of charge should lighting batteries be at for air travel?
Thirty percent or below for batteries shipped as standalone air cargo under UN 3480; this is a regulatory limit, not a suggestion. For spares in carry-on, capacity limit rules apply instead, but many airlines still ask for the lower state of charge. Store and travel at 30 to 50 percent anyway, because it is also the best state for cell life.
Is a D-Tap enough for a small LED panel?
For panels up to roughly 100 W, yes, with a good quality connector. Above that the maths turns against you: 150 W at 14.4 V is 10.4 A, already at the top of the D-Tap range, and contact resistance turns that into heat inside the shell. Use a locking high-current connector for anything above 100 W and keep D-Tap for monitors and camera accessories.
How do I stop the BMS tripping when I switch six fixtures on at once?
Stagger the power-up two seconds per fixture, and if that is unacceptable, specify a BMS with a two-stage trip curve that tolerates 150 percent of rated current for half a second. Then test the whole array cold, at 0 degrees Celsius, because cell internal resistance roughly doubles in the cold and that is when marginal designs fail.
How often should I capacity-test on-set batteries?
Quarterly, or every 100 cycles, whichever arrives first. Retire packs at 80 percent of nameplate capacity, because internal resistance climbs before capacity visibly falls and a pack that still holds charge can still fail to hold voltage under a 300 W load. Keep the test log with the serial number so the trend, not a single reading, drives the retirement decision.
Further Reading
References
