2mm x 2mm boost circuit

[PhotonBoy]

I'm not an electronics expert, but this looks interesting:

http://biz.yahoo.com/prnews/031014/sftu148_1.html

Summary of Features & Applications:
MIC2287 / MIC2289 White LED Driver ICs

-- 2.5V to 10V input voltage range
-- Output voltage up to 34V
-- Up to 500mA input switch current
-- Internal Schottky diode (MIC2289)
-- 95mV feedback voltage
-- 1.2 MHz PWM operation
-- Over Voltage Protection options of 15V, 24V, 34V
-- <1% line and load regulation maintains constant LED brightness
-- 1uA shutdown current
-- Over temperature protection
-- UVLO
-- 2x2mm MLF-8L package options
-- Thin SOT-23-5L package option (MIC2287 only)
-- -40degC to +125degC junction temperature range

Applications

-- Constant-Current White LED Driver for LCD Backlighting
- Cell phones
- PDAs
- Digital cameras
- Camcorders
- GPS systems
- IP phones
-- Camera-flash LED drivers
-- LED flashlights

 

[shankus]

Yes, that is very interesting...
2mm by 2mm!?

 

[Negeltu]

2mm x 2mm???? Maybe it's a typo on their site... hmmm

 

[Steelwolf]

Can't be a typo. They are referring to a standard SMD size, right? MLF-8L. How big would that be? The 4 external components would not be included in that measurement. So 2mm x 2mm is a possibility.

 

[php_44]

There is a spec available on the Micrel web site. The 2mm X 2mm package has 8 pads on the underside of the package(!!!) It requires a PCB layout with 0.25mm pads on 0.5mm centers - that's 10mil pads with 10mil gaps. Whoa - try making a PCboard for that at home, then try to get that thing soldered without shorts or opens. You'd have to use solder paste and a toaster oven or similar method. It would be a challenge to say the least. I wonder if you could get a pre-made breakout adapter board (though that would negate the space savings of the tiny chip)?

There is a version of the chip that's available in a low profile SOT23-5L package (5 normal but small pins). A hobbiest with patience should be able to make a PCB and get it soldered down without too much trouble. One can even get pre-made SOT23 generic PCBs for that package.

The max switch current is a nice 500-700mA, but their graphs indicate running series strings of 3-8 LEDs at a 25mA max current. It's sort of optimized for 3-12V LiION or multi cell applications driving a string of smaller LEDs. It's a pity because the small size would make it nice for a single cell single luxeon driver.

 

[danielo_d]

php 44,
you're right. At their site and in their pdf spec sheet, they reference data involving driving a series of *edit* smaller *edit* LEDs. No mention of driving Luxeon capacities. /ubbthreads/images/graemlins/icon23.gif [not that I'm at that level, yet]

Nice possibilities, though... /ubbthreads/images/graemlins/smile.gif [that is you can handle the form factor.] Maybe something lambda may be able to do.

Danno

 

[UK Owl]

Can't help but think that it is about time someone made a complete dc/dc convertor chip that only needs 1 external component (i.e. a resistor to set the current) that will automatically step up to the voltage required to provide the current requested (e.g. between 3.4 and 13.6 volts for 1 to 4 White LED's)

Given the rapid uptake of white LED's, this would probably become very popular especially if it could be made to run from a supply down to 2 volts or less.

Is they offered all that on a 2mm x 2mm chip, I have no doubt that paypal would have already be sent by many CPF members !

 

[koala]

Unfortunately, you will need an inductor which is always the main component that contributes to the efficiency factor.

Vince.

 

[vicbin]

IMHO Linear LT 1932 is better (min Vin = 1 Volt) instead of 2.5 V for single cell application, still not for LS though.

Linear LT1932

Datasheet : LT1932 Datasheet (PDF)

 

[Steelwolf]

What is the minimum number of external components that a DC/DC circuit could possibly need? Definitely an inductor, no choice about that. No need for a sense resistor if it is a dumb convertor. Capacitor to smooth the ripples is optional. Anything else?

 

[UK Owl]

The point I was trying to make is why not produce a chip with the inductor and capacitor built in. Ideally this would boost up to the output voltage needed to provide the current determined by a single external resistor.

 

[CM]

Inductors are very difficult if not impossible to implement in a chip in the storage capacity that's required to drive LED's. You also need fairly large capacitance for input and outputs on the order of micro farads. Not going to happen soon I'm afraid.

CM

 

[alternety]

May I suggest that as white LEDs enter mainstream lighting there is a real need for a chip that drives variable length strings of diodes but from a higher supply voltage. Undercounter lighting for instance normally usees 12 or 24 VAC. These easily available transformers can be used as a supply and the low voltage wiring simplifies things.

A constant current driver essentially deals with variable numbers of diodes if the output voltage range is sufficient.

An transformerless off-line switcher that drives LED strings could be real useful as well. Low weight and small size for lighting fixtures.

 

[Jarhead]

There is a circuit that bucks down the voltage:

http://www.supertex.com/pdf/misc/HV9906DB6L.pdf

80 to 135VAC, 47-63Hz or 100 to 200VDC input
(fyi, it can be modified for much lower voltage
input, I got one running off of 12V)

Specifically for Luxeon devices.


See also:

http://www.supertex.com/pdf/datasheets/HV9904.pdf

10-400 VDC or 10-265 VAC input.


for regular devices.

 

[Steelwolf]

UK OWl: Inductors and capacitors, unfortunately, have physical size limitations. Even transistors too, actually. There is a certain minimum size for the power levels they are expected to handle. You could make them smaller, but they will also be able to handle less power. In some of the chips, the onboard transistor is used to switch a larger power transistor to allow higher currents to be handled. And what is the tiniest SMD inductor or capacitor you have ever seen? How much current could they take? How much inductance or capacitance could they provide?

That is why it isn't possible to get all the components on to a IC chip. But it can still be very small. Look at Wayne's sandwiches and LEDsAsSmallAsPossible's work. The downside is that you need to be able to work with SMD components.

 

[UK Owl]

No offense guys, it was just a suggestion of mine. I know that by using SMD you can reduce the circuit size such that many members have built their own DC/DC convertor small enough to fit onto a board, brive 4 LED's and still fit in the head of a minim*g. I would love to be able to do this but working with discretes, due to my limited soldering ability, I am unable to.

I have build DC convertor boards quite small but not small enough to fit into a minim*g. I understand that a 2x2 footprint is un-realistic, but in a standard discrete pachage e.g. 8 pin DIL, there is a lot of spare space that is little but resin filler.

Would it be nice if this could be used to house the rest of a DC convertor circuit ?

Note : I edited this because I left the 'un-' of 'un-realistic.

 

[Steelwolf]

UK OWl: No need to apologise. I am still working with discretes too. I try to freebuild, ie, solder all the parts together without a circuit board, in order to reduce the overall size.

WRT the minim*g, is it a 2AA or 2AAA size? With those, I typically use the space of 1 cell for the booster so that the light runs on 1 cell. To get a booster circuit in to the head, I think the smallest "discretes only" circuit that would fit would be the SatCure circuit.

 

[UK Owl]

The Minimag I was refering to is a 2 AA version, but to utilise the existing 'switch' facility they have, and to allow the greater efficiencies, I wish to use two cells over just the one.

I have made invertors before using an adaption of the satcure circuit, but have experienced difficulty getting the power transfer or efficiency I desired (note: transformer was wound on a 3mm dia x 6mm long ferrite tube.

I have also had some success using the 'satcure' mod to power a fluorescent tube.

i.e. wind on a small audio transformer (current winding 10 turns, base bias winding 15 turns) but add a third winding of about 200 turns. Connect one end of this trird winding to the 0v rail, and connect both ends accross a 6 inch fluorescent tube. attach a piece of insulated wire to one end of the tube, and run this alongside the tube, attaching with tape if desired (this is to allow tube excitation), you now have a circuit that will power the tube from a PP3 battery.