LED Flash light ..Pulsed LED, underdamped oscillations

[williB]

here is what i am thinking of doing ,IF i could use PWM in such a way as to supply a constant average current , keeping the led lit brightly , no matter if the cap is 30V or 4 V .. , and it should be possible to NOT use a current limiting resistor , wasting power in a resistor bugs me..
and if i could recycle the current used by the diode ala ,a few oscillations in a LC network , i might be able to do it..
a lot of ' IFs ' there but it might be possible, maybe..

 

[audioguru]

Hi Willi,
Your voltage is far too high for an LED. With a lower voltage then the generator's current will be higher. Then it can fully charge the capacitor much quicker or you could use a bigger capacitor for the LED to light longer.

PWM changes the duty-cycle of the percentage of time that the LED receives full power. If the average power is low then its brightness is also low.

I have seen IR LEDs with a momentary max current of 10A, but my ordinary bright LEDs are 200mA and my ultra-bright LEDs are only 100mA. Both lose efficiency at high current.
So you must use a current-limiting resistor, and its power waste is reduced with less voltage across it.

You can replace the current-limiting resistor with a constant current source circuit to reduce the dimming of the LED as the cap's voltage discharges.

 

[Someone Electro]

You cod use an cap over the LED so it wont take high curent pulses.

So wen the PWM is on it will charge the cap and lit the led wen its off the cap will lit the LED.

might work.

 

[williB]

I recently bought a white LED reading light , battery powered that uses three AG-13 button cells..
It doesnt limit the current to the diode at all.. why doesnt it burn the diode out??
Something is limiting the current to 30mA , but what ? is it possible that the LED itself has current limiting built in??
the no load voltage on the cells is 4.6V , when it is on it drops to 3.9V

 

[williB]

You cod use an cap over the LED so it wont take high curent pulses.

So wen the PWM is on it will charge the cap and lit the led wen its off the cap will lit the LED.

might work.

the power for PWM circuitry will have to come from the caps ,eventually, if it is going to be a real flash light..

 

[audioguru]

Hi Willi,
Those tiny battery cells have internal resistance which limits the current in your simple reading light. Look at how their voltage drops when loaded. :lol:

 

[williB]

ok thanks , i knew something was limiting the current ..
i am going to use the White LED in the light for MY flashlight..
so i need to supply 30mA for twenty min or so..
i was also thinking about a super cap.

 

[williB]

change that..i just took out one cell and it put out usefull light at 4.6 mA & 3V. so if i shoot for say 3.2 V at 5or 6 mA for 20 min that would be a reasonable goal to shoot for..
EDIT : thats 19.2 mW

 

[audioguru]

Hi Willi,
A 2.4F super capacitor should power a 3.2V LED at 6mA for about 20 minutes if you use a current source circuit to keep the current constant during the discharge.
How many hours of shaking would your generator take to charge 2.4F? I figure about 5 hours. :lol: :lol:

 

[williB]

https://www.electro-tech-online.com/custompdfs/2005/05/517.pdf
they sell a 10 F 2.5V super cap with an internal resistance of probably 1.0 ohms..
if i get two maybe i can charge them one at a time..
LOL we shall see about how long it actually takes to charge..
last night i was looking at the scope while shaking a (lol) 1.0 ohms resistor and peak to peak the wave was slightly over 1 div at 1 volt per div..
i just have to transfer that power without loss to the super cap..

 

[williB]

I have thought about using Mosfets , but this is a simple configuration that you really cant beat .
i have been able to get 143 mA out of it , with R1 being a 1.0 ohms , and measuring voltage across it..

 

[Someone Electro]

You can use an PIC for PWM.You run it on its internal 32KHz oscilator and it shodnt take much. (It can run for more than a minute on an 1000uF)

Ufcurse the PIC has to have an PWM module (doing PWM by software wod have to use an higher clock and therfore more power)

You cod also put an low value resistor in the circuit to firstly limit the peaks of power and can be used to meshure curent whith the PICs ADC.And an 1000uF filter cap before the didodes so it reuces those high curent peaks that cod damege the LEDs

It may get complicated but it wod certanly waste les power.

 

[williB]

You are talking about using PWM to drive the Led right ?
I could measure the cap voltage , with the pic ADC and adjust the PWM accordingly , this is true..
There are two sections to this problem
1) charging cap or supercap as efficiently as possible .
2) using limited amt. of power to drive LED for twenty min. or so, also as efficiently as physically possible..

 

[Someone Electro]

I seen an litle circuit that ajusts the curent acording to the voltage it can even run from an lover voltage than the LEDs forvared breakdown (It can drive an white LED from an empty 1.5V button cell)

But i dont think it works at such high powers for an LED light

 

[williB]

i wonder if i could use PWM to turn a constant current source on and off ??
Blinking the LED would extend the power used , for sure..

 

[Someone Electro]

That was an simple 1 transistors circut

Im thinking of an PIC runing at 32Khz that uses PWM to drive the LED,also an 1000uF cap right before the LED so the LED dosent get high curent pulses that cod damege it,and and small resistor (meaby 1Ohm) on the negative side and ADC before it so it can meshure how much curent the LED it taking.

This shod add about 0.1mA to the total pover usege.

The 1Ohm rersistor wodnt have a lot of voltage acros it so it wodnt heat up much.

Shod waste les power

 

[williB]

Last night i got the idea to use a switch when charging , and discharging ..
Switch one way to charge , other way to discharge..
this 1000 uF cap would be in series with LED?

 

[Someone Electro]

Your using PWM and thats making high curent peaks that cod damage the LEDs so if you put an 1000uF cap in paralel whith the LED the some of the high curent peak will go in the cap and some in the LED wen the PWM is off the cap will then feed it to the LED and this way reducing the peak that cod damage the LED whithout losing power.


yea and switch is a good idea

 

[williB]

https://www.electro-tech-online.com/custompdfs/2005/05/PSSI2021SAY_2.pdf
wow this seems perfect..

 

[Someone Electro]

but it also waists power as heat!