driver circuit designs?

[sysadmn]

You can get started with AVR programming for $12-20 or so. The bad news is that the low power versions are 1.8v; no go for a single AA.

 

[Curious_character]

I see two immediate problems with that circuit:

1. It's a voltage regulator. You need a current regulator to drive an LED. It can, of course, be modified, but can't be used as-is -- unless you want to use a dropping resistor in series with the LED which would result in poor current regulation and worse efficiency.
2. The diode in series with the output will drop around 0.5 volt, which would dissipate about 0.5/(3.6 + .5) = 12% of the power to a single LED. Even with 5 volt output, the regulator is about 80% efficient according to the app note. At 3.7 volts out it would likely be less. This is o.k., but not the best for a switching regulator. The more efficient designs avoid a series diode.

c_c

 

[hank]

Here are a few more, just "found, not understood" -- mostly for the 3xAAA "cheap Chinese multiple LED" lights that are around everywhere, some stealing one of the battery slots for a driver.

Lots of different ways, posting it for reference if anyone's got comments
http://www.thompdale.com/flashlight/flashlight.htm

 

[hank]

A tidbit from here:

http://www.edn.com/blog/1700000170/post/1960015996.html

"... I was doing the Analog Seminar with National Semiconductor. We were in England. A guy came up to me at lunch and said: "If I see another white LED driver circuit in your ads I am going to firebomb your headquarters." I chucked knowingly. Myself and many of my friends have also become saturated with LED driver circuits. That was several years ago. Now there is a whole LED driver industry. Now far be for me to risk immolation from you, the angry reader, but I saw a good LED driver application article in Auto Electronics magazine
http://autoelectronics.com/body_electronics/lighting/709AETF1.pdf
(pdf). What caught my eye was Figure 6. It was captioned 'A low-noise application circuit that can meet CISPR 25 level 5 specs.'

Aha, I thought, so all that switching in these LED drivers is causing conducted noise to backflow into the vehicle's power wires and radiating out, causing regulatory failures. Just a good thing to know if you are designing some LED driver circuits...."

 

[OddOne]

1. It's a voltage regulator. You need a current regulator to drive an LED. It can, of course, be modified, but can't be used as-is -- unless you want to use a dropping resistor in series with the LED which would result in poor current regulation and worse efficiency.

You can "hotwire" a voltage regulator to act as a constant-current regulator if you use its sensing mechanism to detect current instead of voltage.

To explain...

Switchers use some sort of feedback loop to operate, and what they generally do is look for a specific voltage at one of their pins. They often use a single resistor or a resistive voltage divider to give the feedback pin the correct voltage when the output voltage is what you want it to be.

If you use a current-sensing resistor in series with your load, pick its value to reach that specific desired voltage when your desired current flows through it, and have the switcher look at that, the switcher will suddenly be regulating current instead of voltage.

This means of application does require an efficiency-robbing sense resistor in series with the load, so it's crucial to select a switcher than needs the lowest feasible feedback voltage. Fortunately quite a few work with 100mV or less.


This is the trick at work in using an LM317 as a linear current regulator - the resistor you pick drops the right voltage to trigger its regulation when the desired current flows through it.

oO