clever led driver / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

clever led driver

1 2 3 next

Move Thread

LAN_403

polop

Wed Jun 17 2009, 05:03PM

Registered Member #1537 Joined: Thu Jun 12 2008, 06:44PM
Location:
Posts: 51

HI there i have a p7 led (google it, there awsome) that requires a 2800mA constant current supply (Vf around 3.6v-4v). i know a resistor can work but drawing this kind of power, it would need to be huge!! i also know about the lm317 and its derivitive, but this is almost as inefficient as a resistor.

How could i make simply and cheaply some kind of switching supply that takes voltage from the 5-9v range and puts out 2-4.5v (Vout shifts in order to keep current regular)

You can buy circuit boards that do this that are specificaly designed for this led, that are also very cheap- but they only push 2A and seems a waste to underdrive such a brutal LED.

any ideas , cheap efficient small driver at 2.8A

tobias

Wed Jun 17 2009, 07:33PM

Registered Member #1956 Joined: Wed Feb 04 2009, 01:22PM
Location: Jersey City
Posts: 172

Yeap..

You can try a ncp3063.. It's a DC/DC chip that has current limiting. But it can supply only 1.5 A without a driver.

Maybe you can find a way to use a FET after it to give you the power you need.
The IC is very cheap.. less then a buck.

Or a LM3401 (less then a buck too).

Conundrum

Wed Jun 17 2009, 07:51PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4061

MC34063 in constant current mode? set the output voltage and cutoff resistor and test with a dummy load with similar Vdrop (several silicon diodes in series work well)

i'd also use a zener on the output just to be safe, set around 1V higher than the rated Vmax..

regards, -A PM me if you need any help with this

mikeselectricstuff

Wed Jun 17 2009, 11:31PM

Registered Member #311 Joined: Sun Mar 12 2006, 08:28PM
Location:
Posts: 253

Altjhough there are many switchmode buck reg chips that can be used for LEDs, by connecting their voltage feedback input across a current-sense resistor, however the typical 1.2v feedback voltage limits efficiency - there are a number of buck reg chips designed for LED apps, which have a much lower feedback voltage threshold - typically 0.2V. Most of the main semi mfrs have something in their range - I know Maxim have an external-FET part which should be easily scalable to larger O/P currents.

wastehl

Thu Jun 18 2009, 03:10AM

Registered Member #1188 Joined: Sun Dec 23 2007, 12:18AM
Location: Ventura CA USA
Posts: 2

There are IGBT gate driver chips that will handle this and give you the added capability of putting a simple pulser circuit on the enable line, useful for signaling, etc.

jonny5

Thu Jun 18 2009, 06:00AM

Registered Member #1807 Joined: Tue Nov 11 2008, 07:36AM
Location: San Luis Obispo, California
Posts: 19

A Buck converter controlled by a 555 timer and a comparator?

The circuit uses a one-shot configured 555 timer and a comparator to implement a wacky controlled off-time buck converter.

6 VDC input, the LTspice simulation of the LED current:

12 VDC input, simulated LED current:

The simulation suggested that the 555 timer provided adequate gate drive at relatively low frequency (20KHz). Higher frequency operation could be obtained by decreasing the one-shot interval. This would permit the use of a smaller buck inductor and/or less current ripple. However, the 555 output might not be sufficient for high frequency switching. Perhaps a NPN/PNP voltage follower between the PMOS gate and the 555?

The voltage reference (150mV...really whatever you want) might be obtained via a voltage divider off of a linear regulator/zener/etc. Unfortunately, the comparator needs a (close to) rail-to-rail output and the ability to handle small input voltages (ie. voltage reference).

At high input voltages (15VDC and higher), the free-wheeling schottky diode dissipates a fair amount of power (over a watt)! Lower input voltages result in less average schottky diode current.

At low voltages (5VDC in the simulation) the switching frequency drops dramatically as the circuit attempts to maintain the desired LED current.

Overall, the circuit looks like it might drive a high-powered LED relatively efficiently. Of course, the only way to tell is to build the circuit and see. One more project to throw on the pile!

Dr. Slack

Thu Jun 18 2009, 08:09AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

There are two alrternatives. Put something together yourself from an inductor, comparator and FET and learn about switching supplies. Or, obtain something ready made.

For ready made, I have always headed stright for the maxim-ic site, for two reasons. The first is they have a huge raft of parts. The second is that they have a policy of sampling, even to (especially to) amateurs, so there are no eithical problems about getting their standard two-off sample quantity for nothing.

I say "have always", because "progress" has meant that fewer and fewer new whizzy clever parts these days are available in a package that you can solder wires onto. Even if you can get fine wires onto the QTFN pads under a microscope, you still may have to get 1 watt out of the power pad in the middle, which is tough without a board. So increasingly my searches there start with DIP or SO package availability as the first criterion.

You can get a MAX16834 3A buck LED driver in an SO package. It needs an external switching FET (and of course an external inductor and a few Rs and Cs), but otherwise controls the whole driving job in one.

A MAX15037 is an internal switch buck 3A PSU, which can be configured for constant current output by returning the load current through a sense resistor to ground across its 0.6v Vsense feedback input, but it's in a TQFN, so nearly impossible to connect without a board.

Linear, TI, NatSemi (I think) and a few others all do LED drivers and PSUs as well. You may find better combinations of features with them. None are going to be more than $1 or $2, so if you have to buy from digikey, it's not going to break the bank.

polop

Thu Jun 18 2009, 05:22PM

Registered Member #1537 Joined: Thu Jun 12 2008, 06:44PM
Location:
Posts: 51

is the schottky diode need to be of any particular rating, for the 555 design ?

GeordieBoy

Thu Jun 18 2009, 05:26PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Schottky diode must block the full maximum supply voltage in reverse-bias, and be able to support the required output load current in forward conduction.

polop

Thu Jun 18 2009, 06:05PM

Registered Member #1537 Joined: Thu Jun 12 2008, 06:44PM
Location:
Posts: 51

are these inductors alright?

they should i guess take the full current to the led. again 555 circuit
edit, i guess they are too low powered, winding my own i guess iis the best way i guess i have some torriodal cores from a psu, about 1cm diameter with perhaps a 4mm hole,

how critical is the exact inductance

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob