LED driver to run off 4.8V

Thread Starter

ian field

Joined Oct 27, 2012
6,536
Does anyone have a favourite circuit for a LED driver from 4.8V they're willing to share?

Preferably a very simple buck type. There isn't much room in the bicycle lamp housing - I was thinking of a "dead-bug" build that can squeeze in the dead space around the reflector.
 

Thread Starter

ian field

Joined Oct 27, 2012
6,536
LED voltage and current? Is it battery operated or from a generator?
I was sort of thinking I might plump for a white LED just to be boringly conformist, last time I looked the Vf was around the 3.4V mark.

But since LEDs need current regulation, I didn't think of voltage as the foremost specification.

4.8V is pretty much the nominal voltage of 4x nickel chemistry cells.
 

dl324

Joined Mar 30, 2015
16,846
Current regulation isn't critical for many applications. Personally, I'd just go with a series resistor. If you want a current source (sink), you can do it with 2 transistors and 2 resistors. There was a thread last month about help understanding a current source. Don't pick the one started this month. That thread didn't have anything to do with a current source...
 

Thread Starter

ian field

Joined Oct 27, 2012
6,536
Current regulation isn't critical for many applications. Personally, I'd just go with a series resistor. If you want a current source (sink), you can do it with 2 transistors and 2 resistors. There was a thread last month about help understanding a current source. Don't pick the one started this month. That thread didn't have anything to do with a current source...
A series resistor isn't quite a buck type regulator.

The original bicycle lamp already has one of those - I'd like to improve on that.
 

RichardO

Joined May 4, 2013
2,270
The original bicycle lamp already has one of those - I'd like to improve on that.
How much current does the LED draw? The bike headlight I made is barely bright enough at 3 watts. I need more like 12 watts to ride at full speed. You will probably get better results because you are starting with better optics than I have.
 

Thread Starter

ian field

Joined Oct 27, 2012
6,536
What would you gain by having a regulator that consumes power without providing illumination?
Probably the same as most people can gain by replacing a power dissipating dropper resistor with an efficient step down SMPSU type design.
 

Thread Starter

ian field

Joined Oct 27, 2012
6,536
How much current does the LED draw? The bike headlight I made is barely bright enough at 3 watts. I need more like 12 watts to ride at full speed. You will probably get better results because you are starting with better optics than I have.
ATM: I don't know all the details, the LED package type could be 1W or 3W, but the specification for the bicycle lamp only claims to be 0.6W.

The LED package has a heatsinking solder base as well as the solder lugs. In the lamp, that base isn't soldered, but was installed with thermal paste to an array of through plated holes for dissipation - I assume that's why it was de-rated to less than the probable 1W of the LED.

A basic circuit idea would be a step in the right direction, I can tweak component values to get it just right.

There's an SOIC-8 chip on the board that I can't identify any number even with a jeweller's loupe - that could possibly be the chip for the push button ON/OFF. If it wasn't for that - I'd just replace the 4x AA holder with a single D cell and use a blocking oscillator Joule thief circuit.
 

dl324

Joined Mar 30, 2015
16,846
Probably the same as most people can gain by replacing a power dissipating dropper resistor with an efficient step down SMPSU type design.
Here you go:
upload_2015-10-31_12-16-47.png
This is the current source that was discussed in the thread I referenced. Instead of dissipating the power in a resistor, you can dissipate it in 2 transistors and 2 resistors. The only advantage is that you can set a constant current which a resistor or buck regulator wouldn't have done.

I guess another advantage would be that this is more efficient than any buck regulator.
 

Thread Starter

ian field

Joined Oct 27, 2012
6,536
Here you go:
View attachment 93851
This is the current source that was discussed in the thread I referenced. Instead of dissipating the power in a resistor, you can dissipate it in 2 transistors and 2 resistors. The only advantage is that you can set a constant current which a resistor or buck regulator wouldn't have done.

I guess another advantage would be that this is more efficient than any buck regulator.
You need to go back to the drawing board!

A linear regulator dissipates (wastes) energy just the same as a resistor.

If you can't do it, stay out of the way and make room for someone who can.
 

Thread Starter

ian field

Joined Oct 27, 2012
6,536
Believe it or not, I was actually trying to help you.


If I respond to another of your posts, please remind me again not to waste my time.
It is actually hard to believe, since most of what you suggested was contrary to what I asked - and some of it just plain wrong.

Here's hoping you remember not to waste my time in future....................
 

dl324

Joined Mar 30, 2015
16,846
It is actually hard to believe, since most of what you suggested was contrary to what I asked - and some of it just plain wrong.

Here's hoping you remember not to waste my time in future....................
Again, sorry for trying to help and for wasting your time.
 

NCSailor

Joined Jun 15, 2013
33
It is actually hard to believe, since most of what you suggested was contrary to what I asked - and some of it just plain wrong.

Here's hoping you remember not to waste my time in future....................
I think Dennis was way too polite. You came here and asked if anyone had favorite circuit to solve your problem. You failed to provide many of the requested details. It appears to me that you already had a solution and were just here hoping that people would see how "smart" you are.

Thanks for wasting all our time.
 

ronv

Joined Nov 12, 2008
3,770
I was sort of thinking I might plump for a white LED just to be boringly conformist, last time I looked the Vf was around the 3.4V mark.

But since LEDs need current regulation, I didn't think of voltage as the foremost specification.

4.8V is pretty much the nominal voltage of 4x nickel chemistry cells.
You may be wasting your time with a buck converter. If the LED is 3.4 volts and the diode in the converter is .6, the current sense voltage .3 and maybe .5 for a FET and inductor it's a wash.
 
Top