My goal is to record a video with a varying intensity light source.

  • No flickering at any frame rate (frame rate 30 60 120 240) while watching frame by frame, not human eye.
  • Smooth dim level
  • Any type of lamp is OK. (DC powered is preferred)
  • Dimming by voltage controlling from signal generator (sine shape about 10-60 Hz or higher)

Many requirements make me confuse how to choose the right things to the right job.

I tried dimmable halogen capsule with 12VDC. It worked well when I powered it using large DC power supply. Varying voltage changes light intensity as I need and cause no flickering. But when I use the signal generator to smoothly dim the light, it cannot afford that high current.

I will try LED with 0-10V dimmable driver. No need high power source. But the price is expensive for me. I want someone to make sure it will be the right way or not? I am worried about flickering at high frame rate.

Could you please recommend the type of light source, driver (if needed)? Or I have to go back to halogen lamp and find high power oscillator?

  • \$\begingroup\$ What role does the signal generator play in this scenario? \$\endgroup\$
    – Majenko
    Aug 14, 2016 at 10:36
  • \$\begingroup\$ Sorry if I did not ask clearly. I would like to use the signal generator to control the light intensity of the lamp. If I feed a sine wave, it will smoothly dim the light. \$\endgroup\$
    – pakornosky
    Aug 14, 2016 at 10:40
  • \$\begingroup\$ I thought it was probably something like that. \$\endgroup\$
    – Majenko
    Aug 14, 2016 at 10:43
  • \$\begingroup\$ Describe the order of magnitude of brightness you need. A candle? 10 candles? A desk lamp? A spotlight? \$\endgroup\$ Aug 14, 2016 at 10:49
  • \$\begingroup\$ I did not consider the brightness much. I just need it change through time. But the wider range of brightness is better if possible. :) \$\endgroup\$
    – pakornosky
    Aug 14, 2016 at 11:14

2 Answers 2


unlike Majenko, I'd totally not say that incandescent is the only way to go:

Yes, many cheap LED dimmers will just PWM your LED; meaning it turns on and off faster than the eye can register – however, when catching that on a sensor/film, the shutter speed will make it very likely that you might see flickering, simply because frames might catch a varying amount of on/off.

So, true PWM directly applied to the LED is not an option – however, there's nothing wrong as using an LED as such. In fact, LEDs are very nice for applications like these – they can be brightness-adjusted without any risk of damage by varying the current flowing through them.

So, what you'd need is an adjustable constant current source – in the easiest case, that would just be an adjustable power supply in current mode, which internally is actually pretty much an amplifier (typically, class A), meaning that it also wastes quite some energy as heat – but that might not be your primary concern here.

Going one step further (and maybe saving you a lot of money on adjustable benchtop power supplies), there's a lot of ICs that actually are power supply controllers dedicated to dimming LEDs with adjustable current. They do control a MOSFET switch with a PWM, but they do it at a high rate, and use that PWM to drive an inductance, which, on time average, outputs a constant current. I've built such an adjustable LED supply with then ON Semi NCP3065, which enables me to drive a 39V max 200mA LED from a 12V battery. Ripple in current at relevant frequencies is very low – the switching frequency lies between 100 kHz and 250 kHz, so far above the shutter opening period that no camera sensor can catch the ripple, even if it was of significant height, which it isn't, because the PWM isn't directly changing the current flowing through the LED, but just the amount of energy stored in the inductivity.

Now, of course, my LED is a blindingly bright light source that is best comparable with a really, really bright flashlight; this might be far too small for studio usage. If you need to dim something like an LED flood light, inductivities might get larger than a single small <1€ IC can handle; then, you're suddenly the business of designing your own Switch mode power supply. Might not be the easiest choice.

  • \$\begingroup\$ "the PWM isn't directly changing the current flowing through the LED, but just the amount of energy stored in the inductivity." If I understand you correctly, PWM frequency will not affect the operating frequency. It controls MOSFET to change current continuously and nothing to consider with flickering. Am I misunderstood? What is the consequence if I change the PWM frequency? e.g. PWM 100Hz and 200Hz. \$\endgroup\$
    – pakornosky
    Aug 14, 2016 at 12:05
  • \$\begingroup\$ What I'm saying: There's no PWM in the frequency range as your camera operates. It is a switch mode power supply, so the current isn't completely constant over time, but that doesn't matter here, since the switching frequency is much higher than the shutter speed/frame rate. \$\endgroup\$ Aug 14, 2016 at 14:00
  • \$\begingroup\$ So the switching part turn on and off at very high frequency and the PWM does not matter with it. The PWM just control the current level during "on" cycle, nothing affect the switching frequency? \$\endgroup\$
    – pakornosky
    Aug 14, 2016 at 14:43
  • \$\begingroup\$ @pakornosky please read up on switch-mode power supplies. \$\endgroup\$ Aug 14, 2016 at 14:54
  • \$\begingroup\$ After reading, I mostly get how switching regulator works, but still have some question with my application. According to NCP3065 application AND8298 note, The note said the frequency range of PWM is basically 200Hz-1kHz. The frequency below 200Hz will cause flickering for human eye. This implies that high switching frequency does not matter, but the bottleneck is PWM frequency. With higher video frame rate will need higher PWM frequency to avoid flickering. Can I just use 100kHz PWM frequency? \$\endgroup\$
    – pakornosky
    Aug 14, 2016 at 19:54

DC Incandescent (halogen, xenon, etc) is really the only way to go. Pretty much everything else has situations where flickering can occur. LEDs are usually dimmed by PWM which introduces flickering (although limiting the current can also be used - check how your driver works), and all AC systems naturally flicker due to the alternating nature of the current.

Instead of using a signal generator (which isn't designed for driving heavy loads like incandescent bulbs) I would recommend using a programmable bench power supply. One that has a serial or USB connection to a computer and a simple to use protocol for controlling it (I use the TENMA 72-2535 - they do other versions too with different current and voltage - TENMA re-badge other peoples' products so you would likely find the same thing under different names). With that you can program the computer to set the output voltage to specific values at specific times and create the varying voltage output that you desire (note: that model can't do high frequencies, only slow changes, due to the nature or the serial protocol).

  • \$\begingroup\$ Not inherently true – there's plenty of adjustable constant current source with switching periods far smaller than shutter speed, and those will be just as flicker free as a thermally slow incandescent. \$\endgroup\$ Aug 14, 2016 at 10:50
  • \$\begingroup\$ Also, if recommending an adjustable bench supply: set it to constant power mode, add a Low Pass filter if necessary (it shouldn't be necessary if the supply isn't PWMing too slowly) and just drive an LED with adjustable current. \$\endgroup\$ Aug 14, 2016 at 10:51
  • \$\begingroup\$ There's "flicker free" and "no flicker". The OP asked for "no flicker" at "any" frame rate. That includes 1 frame per week. \$\endgroup\$
    – Majenko
    Aug 14, 2016 at 10:51
  • \$\begingroup\$ yes, so? That's easier than having 1000 frames per second without flicker! \$\endgroup\$ Aug 14, 2016 at 10:52
  • \$\begingroup\$ Interesting. I had never thought of the programmable power supply before. I do not know the maximum frequency that the supply can do. The technical datasheet did not give this information. Is it OK to generate the sinusoidal wave with 10-60 Hz? is it too fast? \$\endgroup\$
    – pakornosky
    Aug 14, 2016 at 11:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.