LED torch/spotlight - Electronic Circuits Projects Diagrams Free

Da_vinci_cod · 14th February 2008, 12:18 PM

hello all,

I'm planning to make an LED torch. however, wanting to make it efficient, I did some "research" and unearthed these two articles: http://electronicdesign.com/Articles....html?feed=rss
and http://superpositioned.com/articles/...light-with-pvc

In these articles (which are on the same circuit design by the way) they say that they use Pulse Width Modulation in the circuits. They also use 7 LEDs in series. they both say that the design is effiecient.

I was wondering if it would be better if i connected the LEDs in parallel rather than in series, how to change this circuit so that the LEDS are in parallel, and if PWM is good to use in this case or not. can someone help me here?

**Nigel Goodwin** · 14th February 2008, 12:27 PM

The best way is in series, as you only have a single current limiter - in parallel you need one for every LED.

Hero999 · 15th February 2008, 04:18 PM

The beat way is to use a switching regulator to limit the current rather than a simple resistor.

PWM doesn't save you any power because the power dissipation in the resistor relative to the LED is always the same.

Da_vinci_cod · 3rd March 2008, 11:07 AM

thanks for the advice guys. im a major beginner in electronics, despite doing it for 1.25 years in school. we dont actually learn theory or how to design our own circuits, we just make pre-designed ones.

mrmonteith · 3rd March 2008, 02:05 PM

I'm not sure why the reason for the PWM. If you want to make an LED light source efficient you can use just a simple timer circuit like a 555 timer and set it so the ON time of the LED is a small percentage of the whole cycle. Say the LED is only on 1/10th of a cycle and run the cycle fast enough your eyes can't see the different. The other part is the current-limiting. I guess there are several ways you can go about that as you've seen. It depends on how simple you want to keep the circuit vs how much you're willing to give up a little power loss to the circuitry.

Michael

Hero999 · 3rd March 2008, 02:44 PM

But that doesn't increase the efficiency, the power dissipated in the resistor relative to the LED is the same regardless of the duty cycle.

mrmonteith · 3rd March 2008, 02:49 PM

Someone addressed the current usage issue. I'm mainly tackling it from the perspective of power used over time. Our eyes can still perceive the light as being on the whole time but yet power is only applied to the LED 1/10th of the time versus just a steady flow. The two combined I think would be the optimal solution.

ronsimpson · 3rd March 2008, 02:55 PM

PWM an LED is a very good way to efficiently drive LEDs. Fairchild has a series of PWM ICs that drive LEDs. They limit the LED current by measuring the current. The LED brightness remains constant over changing battery voltage. In the case of the FAN5333 there is only 0.1 volt loss across the current sense resistor.

http://www.fairchildsemi.com/ds/FA/FAN5330.pdf

Fairchild has a number of similar parts. There must be 50+ parts built to drive LEDs. Start out by reading the FAN5330 data sheet, application notes and demo board information.

Zetex has a very good book (application note) on PWMs for LEDs. See DN81 Lighting handbook!

Hero999 · 3rd March 2008, 03:15 PM

The FAN5330 isn't just a simple PWM controler, it's a propper little switching regulator which is much more efficient than a simple series resistor with or without PWM.

All I'm saying is that with a simple series resistor, it doesn't make any difference whether you use PWM or not, the efficiency will be the same.

Diver300 · 3rd March 2008, 06:51 PM

The data sheet for the FAN5330 has this to say about PWM dimming:-

"The maximum frequency should not exceed 1kHz to ensure a
linear dependence of the LED’s average current."

1 kHz is too low if you want a torch that doesn't seem to flicker. You won't be able to see it if everything is stationary, but if you move the torch quickly in the dark, it will be most distracting.

You would be better to use the circuit on page 8 of the FAN5330 data sheet that gives dimming with no fast changing of the LED current.

audioguru · 3rd March 2008, 10:23 PM

Quote:

Originally Posted by mrmonteith

Someone addressed the current usage issue. I'm mainly tackling it from the perspective of power used over time. Our eyes can still perceive the light as being on the whole time but yet power is only applied to the LED 1/10th of the time versus just a steady flow. The two combined I think would be the optimal solution.

That is not true. Your vision is slow. You see the average current of a pulsed LED which seems dim. That is why a multiplexed display needs to drive the LEDs with very high current just to be seen.
DC looks bright. Very narrow pulses seem like the LED is not turned on.

ronsimpson · 3rd March 2008, 10:42 PM

It is best to use PWM to dim a LED because many types of LEDs (white) act different at low current. A 20mA LED operating at 1mA may appear to have a different color, while a 20mA LED operating at 20mA for 5% of the time has the same color. I usually try to keep the frequency above 100hz but defiantly below 1khz.

Maybe I should explain PWM. Using a FAN5330 there is two types of PWM happening. The first is happening at 500khz and takes the battery voltage and steps it up to the LED’s voltage to maintain 20mA. The optional second level is to turn on/off the current flow to make the LED appear dimmer. This happens at perhaps 200hz.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Latest
Junebug Assembly Tips	Krumlink	Micro Controllers	187	18th May 2008 03:19 AM
555 timer flashing LED circuit	matt_hau	General Electronics Chat	15	11th February 2008 02:08 AM
LED Matrix Project Debug, Please Help :[	mekohler	Micro Controllers	2	7th December 2007 06:56 AM
solar blinking LED HELP	team_nightstalker	Electronic Projects Design/Ideas/Reviews	33	26th October 2007 05:04 PM
P channel FETs vs N channel FETs for LED Matrix applications ??	tdg8934	General Electronics Chat	3	9th September 2007 05:16 PM

14th February 2008, 12:18 PM	(permalink)
Da_vinci_cod	LED torch/spotlight hello all, I'm planning to make an LED torch. however, wanting to make it efficient, I did some "research" and unearthed these two articles: http://electronicdesign.com/Articles....html?feed=rss and http://superpositioned.com/articles/...light-with-pvc In these articles (which are on the same circuit design by the way) they say that they use Pulse Width Modulation in the circuits. They also use 7 LEDs in series. they both say that the design is effiecient. I was wondering if it would be better if i connected the LEDs in parallel rather than in series, how to change this circuit so that the LEDS are in parallel, and if PWM is good to use in this case or not. can someone help me here?

14th February 2008, 12:27 PM	(permalink)
Nigel Goodwin	The best way is in series, as you only have a single current limiter - in parallel you need one for every LED. __________________ PIC programmer software, and PIC Tutorials at: http://www.winpicprog.co.uk

15th February 2008, 04:18 PM	(permalink)
Hero999	The beat way is to use a switching regulator to limit the current rather than a simple resistor. PWM doesn't save you any power because the power dissipation in the resistor relative to the LED is always the same. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

3rd March 2008, 11:07 AM	(permalink)
Da_vinci_cod	thanks for the advice guys. im a major beginner in electronics, despite doing it for 1.25 years in school. we dont actually learn theory or how to design our own circuits, we just make pre-designed ones.

3rd March 2008, 02:05 PM	(permalink)
mrmonteith	I'm not sure why the reason for the PWM. If you want to make an LED light source efficient you can use just a simple timer circuit like a 555 timer and set it so the ON time of the LED is a small percentage of the whole cycle. Say the LED is only on 1/10th of a cycle and run the cycle fast enough your eyes can't see the different. The other part is the current-limiting. I guess there are several ways you can go about that as you've seen. It depends on how simple you want to keep the circuit vs how much you're willing to give up a little power loss to the circuitry. Michael

3rd March 2008, 02:44 PM	(permalink)
Hero999	But that doesn't increase the efficiency, the power dissipated in the resistor relative to the LED is the same regardless of the duty cycle. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

3rd March 2008, 02:49 PM	(permalink)
mrmonteith	Someone addressed the current usage issue. I'm mainly tackling it from the perspective of power used over time. Our eyes can still perceive the light as being on the whole time but yet power is only applied to the LED 1/10th of the time versus just a steady flow. The two combined I think would be the optimal solution.

3rd March 2008, 02:55 PM	(permalink)
ronsimpson	PWM an LED is a very good way to efficiently drive LEDs. Fairchild has a series of PWM ICs that drive LEDs. They limit the LED current by measuring the current. The LED brightness remains constant over changing battery voltage. In the case of the FAN5333 there is only 0.1 volt loss across the current sense resistor. http://www.fairchildsemi.com/ds/FA/FAN5330.pdf Fairchild has a number of similar parts. There must be 50+ parts built to drive LEDs. Start out by reading the FAN5330 data sheet, application notes and demo board information. Zetex has a very good book (application note) on PWMs for LEDs. See DN81 Lighting handbook!

3rd March 2008, 03:15 PM	(permalink)
Hero999	The FAN5330 isn't just a simple PWM controler, it's a propper little switching regulator which is much more efficient than a simple series resistor with or without PWM. All I'm saying is that with a simple series resistor, it doesn't make any difference whether you use PWM or not, the efficiency will be the same. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

3rd March 2008, 06:51 PM	(permalink)
Diver300	The data sheet for the FAN5330 has this to say about PWM dimming:- "The maximum frequency should not exceed 1kHz to ensure a linear dependence of the LED’s average current." 1 kHz is too low if you want a torch that doesn't seem to flicker. You won't be able to see it if everything is stationary, but if you move the torch quickly in the dark, it will be most distracting. You would be better to use the circuit on page 8 of the FAN5330 data sheet that gives dimming with no fast changing of the LED current.

3rd March 2008, 10:42 PM	(permalink)
ronsimpson	It is best to use PWM to dim a LED because many types of LEDs (white) act different at low current. A 20mA LED operating at 1mA may appear to have a different color, while a 20mA LED operating at 20mA for 5% of the time has the same color. I usually try to keep the frequency above 100hz but defiantly below 1khz. Maybe I should explain PWM. Using a FAN5330 there is two types of PWM happening. The first is happening at 500khz and takes the battery voltage and steps it up to the LED’s voltage to maintain 20mA. The optional second level is to turn on/off the current flow to make the LED appear dimmer. This happens at perhaps 200hz.