Hey Everyone…
I’ve always wondered if a capacitor could somehow be used to smooth out the visible PWM that some/many drivers use for dimming? I dont know much about electronics, but I have a feeling its not that simple. I’m curious about a technical reason why this would work or would not work…
Thanks!
a simular thread is already made, maybe take a look there. if i understand correctly it would work but the size of those capacitors would be to big to fit a flashlight.
Thanks!
How large the cap is would depend upon the pulse frequency. The requirement is the time constant be long relative to the inverse of the pulse frequency.
If the PWM operates at 15Khz, C would not have to be particularily large. Consider this for a moment. If the PWM operates at 15kHz, then the time
constant needes to be long relative to 1/15,000 or about 67 microseconds. Long is generaly considered to be on the order of magnitude of 10 times, So the RC time constant would need to be on the order of 700 microseconds. If you assume the effective load from the LED is approximately 1 ohm (not far from the truth is we are talking about an XM-L2 driven to 1000 lumens), then we would need a capacitor of roughly 1/1500 farads, or about 700 microfarads. At 6 volts such a cap would in fact be quite small and would easily fit inside a flashlight.
Unfortunately that is not the only issue, the Cap will in fact present a very low impedance to the driver, so it might substantiallly shorten the life
of the driver.
mattheww’s correct. But then again, 15KHz PWM wouldn’t be visible anyway, so there’s no point in adding a cap.
To smoothen out 200-1000Hz PWM, would be a different story.
No.
The reason PWM works is that the voltage of the pulse is above the VF of the diode.
Adding a capacitor to average the voltage would lower it below the operating voltage of the diode.
That whole RMS thingy with AC (or pulsating DC) thing
gerald, contradicting what you say, a 1000 micro Farad capacitor changed the output of an annoying light I had to: not so annoying without reducing the output in any way.
I did a test already with a Electrolyte Cap ( a 10V/250UF ) on the PWM of a Black Shadow Padme and it increased the Low and Mid to a much higher output close to High, ( though it did “smooth” out the PWM, but the modes were to bright.
Even at the same junction temperature, LEDs put out more light per current at lower current. So the efficiency would be improved. Also, because of the lower impedance seen by the driver, it would stay in regulation longer in the lower modes.
The capacitor discharges only through the LED. So, unless there is leakage LED current below minimum Vf, the capacitor will charge until it generates light. You may be right for moonlight modes.
In general incorrect. Below Vf, the LED is a much higher resistance device, so the time constant would be much longer.
Effectively once the first PWM pulse is delivered, the cap would be charged to Vf very quickly, and thereafter remain at
essentially at or above Vf for all subsequent pulses. For all practical puposes the device would work exactly as desired.
That sounds right to me. The problem is finding a capacitor with a high enough value, low enough resistance and that fits in.
Reducing visible PWM would take a much higher valued capacitor than smoothing it out to improve efficiency for a higher frequency PWM.