Maker Pro
Maker Pro

Smallest Possible Linear LED Driver

modern_messiah

Dec 1, 2010
29
Joined
Dec 1, 2010
Messages
29
HI all,

I have been racking my brain the last two days trying to design a simple linear/buck constant current regulator for driving LEDs. I've looked at everything I can think of - discrete ICs from all the major players (TI, Diodes, Linear Tech etc) to building my own using MOSFETs and so on. I feel my knowledge is borderline hopeless, so I've come here to see if anyone can help out.

Now before I go on, yes I have read the sticky in this very forum on driving LEDs. I think I should list my requirements to show you why what's listed in the sticky won't work, or at least show me that I don't know how to make them work for me.

- Single Lithium Ion rechargeable power source (4.2V fresh, 3.7V nominal, 2.7V flat)
- Single HB LED (XPG, XPE, Nichia 219)
- Maxium current of 1000mA (over 800mA is desired)
- Maximum Vf or 3.3V will be needed
- PWM control. I will be pairing the driver with an uC to control brightness.

So as you can see all I really need is a linear LED driver. I stumbled across a circuit that meets all my requirements except for the current output (this circuit only manages 400mA). All the circuit contains is a 400mA low drop out linear regulator (PJFQ), an output capacitor and a PIC (Microchip 12F629). Photo of the circuit is below:



How does this circuit function? The confusing part for me is that 99% of the LDOs I find only have an enable pin (which is not appropriate for PWM control) and secondly LDOs have a fixed output voltage.

It looks like the LDO output pin is connected to the +ve terminal of the LED, but then the LED -ve terminal is connected to ground...LDOs output a constant voltage so how can this be working? I can't see how the uC is connected up as the signal traces are too tight to differentiate :(

Any help would be highly appreciated as I don't think I can look at anymore datasheets!

Thanks,

- Matt
 
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The ones using a switchmode regulator are not much larger and will result in significantly longer battery life (and far less heat)
 

modern_messiah

Dec 1, 2010
29
Joined
Dec 1, 2010
Messages
29
OK so after consulting with a friend I think I know how it works. I also understand why you are suggesting a switch mode regulator...if you have one in mind or can point me in the right direction I'd be very grateful - I've seriously spent the last 2 days solid looking and just can't find one!!!!
 

rob_croxford

Aug 3, 2010
262
Joined
Aug 3, 2010
Messages
262
http://uk.farnell.com/diodes-inc/zxld13625ta/ic-led-driver-buck-1a-tsot23/dp/1843876

Here is one i am currently using. Its a switching high side current regulator. The current level is sensed on the high side and the output current is either set by the sense resistor or can be directly controlled using either PWM or an external DC voltage. Read the data sheet for full calculations.

I would also suggest having a secondary control loop for voltage as this must be fed with a constant voltage.

Rob
 

modern_messiah

Dec 1, 2010
29
Joined
Dec 1, 2010
Messages
29
Thanks Rob. The input voltage won't work for me but looking at Farnell I seem to be finding a few that might be appropriate so I'll check them out. The only issue I see is that the switching regulators require a lot of external components which doesn't help me too much. I also have a good understanding of how the driver in the image in my original post works so if need be I can probably fall back on that.
 

rob_croxford

Aug 3, 2010
262
Joined
Aug 3, 2010
Messages
262
yes that was the 60V version i believe they do an entire range. To be honest most semiconductor manufactures do some kind of range of current regulators. LT and Zetex being two that i have used.

Using a chip similar to this and a Uc you should be able to develop a fairly tight control loop that doesnt requre too many external components. The best way i have found to do it is to set the current using a variable DC output from a Micro with a built in DAC. This removes the need for PWM.

As you are running from a battery you should not need to much line filtering and there will be no need for a seperate switcher to convert voltage levels.

Also from my calculations runnign a pic on 3.3V (DAC output 0-3.3V you will need an 8 bit resolution for decent dimming) and setting the current sense resistor to 0.2 ohms you should be able to vary the output from 350mA up to 1A.

Unfortunatly the external inductor and didoe are essential to operation however only a very small inductor will be neeed. And you could use a planar inductor making the design even smaller.

Rob
 

rob_croxford

Aug 3, 2010
262
Joined
Aug 3, 2010
Messages
262
Also if you only want a single output current you can set this using either a voltage regulator or just a divider strait from the battery. That way you can erradicate the uC from the design.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
One of the issues is that the battery voltage can vary both above and below the LED Vf. This necessitates a rather more complex design based on buck/boost architecture.

There are some remarkably small boards available on eBay, but I'm not sure if their input voltage range goes low enough to suit your purpose.

Life gets easier if you can have 2 batteries (smaller ones may be OK)

Here is a possible solution.

Here is someone selling something they claim is good for this, but it appears to be a simple linear current regulator (so the LED will get dim as the battery voltage falls below 3.3V (ish)

Here is another option which *seems* to fit your needs.

Beware that some of these solutions will discharge your battery way past the point of no return if you let them.
 

modern_messiah

Dec 1, 2010
29
Joined
Dec 1, 2010
Messages
29
Thanks Steve(?),

That circuit looks good except for the component count being so high. One thing I don't get is it states it's a boost converter but there are times when the forward voltage of the LED (3.5V) is actually lower than the battery voltage (typically 4.2V fresh off the charger). The driver makes no mention of buck, but there clearly needs to be a step-down stage incorporated otherwise it just won't work with fresh lithium ion cells...

As for the drivers you have linked I am MORE than aware of the first set - these are all I;ve been using for the past 2 years :eek: The regulators on these boards are called AMC7135 linear constant current regulators. They can be paralleled to add their output. each chip is good for between 350 and 380mA. Paralleling 8 yields the 2.8A output these boards provide. The brightness can be controlled by connecting a PWM input to the Vdd pin of each regulator.

I've since learnt the circuit I initially asked about works in effectively the same way. A friend figured it out (he knows a bit more than I do that's for sure):

The TPS73601 is being used as a current regulator. Since there is nothing connected to the FB pin (or it’s tied to ground), the regulator doesn’t pay any attention or make any attempt to regulate voltage directly. It will output voltage as necessary. It has an internal 650mA current limit built in though – this means that it will output only as much voltage as is necessary to drive 650mA; No more.

So there’s the regulation concept. Quite nice actually.

I’m sure PWM is then being used from the microcontroller to the EN pin on the regulator for brightness control. Just because the Microcontroller doesn’t have a hardware PWM module doesn’t mean that the software designer couldn’t simulate it by turning a pin on and off very quickly in code..

while (medbright) { LATA |= (1<<5); Delay_us(500); LATA &= ~(1<<5); Delay_us(500);
}

That would give you a 1Khz PWM at 50% duty cycle on Pin RA5, for instance. There’s no problem with doing it that way when the microcontroller doesn’t have anything else to do anyway.. Hardware PWM is nice because your micro can continue to do other things while generating a PWM signal, but for a flashlight driver, it’s hardly necessary.

They basically function in exactly the same manner. The only difference is the regulator is a bit more powerful than the AMC7135 chips. I'm working on a simple circuit now that works the same way, but with a different regulator (good for 1A) and an ATMEL uC.

Linear regulators suck, but the space this driver requires is minimal (looks like I can fit it on an 8mm PCB) so I'll stick with it for the time being :)

- Matt
 
Top