Twelve 0.5W SMPS LED drivers compared to one 6W LED driver?

[4pyros]

Attached is an example SMPS for a rear LED light...such a circuit will last for years and years.

How do you know that?

 

[Flyback]

from a short inspection of that circuit, you can see that that can be simply done without any electrical overstressing of any component of that circuit. Thus it will last for years and years...it has no 'lytic caps either.

 

[dr pepper]

Flyback is right on that one as far as I'm concerned.
Also allthough not a direct argument my opinion on the smps unreliability comment isnt shared.
Nearly every device these days esp domestic has a smps, which means theres bound to be some go wrong, its the most stressed part of many electronic devices, if we still had linear supplies in things then they would go wrong too, so its not so much that a smps is unreliable its more the fact that they are so common.
Smps supply failiures in more expensive commercial and industrial equipment is no doubt less common, as they had the budget to use better components and topologies.

 

[ChrisP58]

While it is certainly true that a linear regulator will be less efficient than a switcher, the LED array and the linear regulator will be far MORE efficient than the incandescent lamps that it is replacing. So, even though it isn't the most efficient topology, it still generates less heat than before. That, and the fact that a linear will be cheaper, is the comparison that the auto engineers are looking at.

 

[Flyback]

The attached is about the cheapest linear regulator that you can get for a rear/brake LED light.
Its for an executive car.
Those BJTs need to be SOT89's.
The PCB will either need to be bigger than an smps one so as to be able to pass out the heat, or the pcb will need to be mcpcb. So already we have extra expense due to it being a linear regulator.
Another point is that to save on components, only one actual string of leds gets its current regulated..the other strings are "regulated by proxy", and as such one assumes excellent thermal coupling between the BJTs and the LEDs, and also assumes that the BJTs all happen to be matched, -something that no manufacturer of BJTs guarantees for any given batch of BJTs.
Note that despite having a Car battery to power it, the most we can have is three in series, since in cold ambients, the red leds may have voltages near 3.3V, and with the supply cable volt drop, the supply at the lamp may well not manage that, and especially if a reverse polarity diode is used...So we end up burning off the excess power....If the driver is in a downhill traffic jam, then he/she may keep their foot on the brake pedal for a long time, and it'll be on full power.

Regarding incandescent rear/brake lights, they are a different story, incandescants don't suffer from heat the way that LEDs do. Incandescants chuck out a lot of their energy by means of infra-red, whereas LEDs almost totally rely on heat being convected away from them, or conducted into their footprint copper....Another point is that incandescent lights come as bulbs, which are easily popped out and replaced, the attached led linear regulator comes on a pcb/heatsink assembly, and needs the car to go into the garage for a good day at least for dismantlement, and replacement....all the more reason to use a cool running smps driver which will keep cool the internal ambient and make the leds last longer.

 

[ronsimpson]

Flyback, try this.
Vin max is only 20V. (probably OK for auto use) I think it is built for 100 to 200mA LEDs.
It has a little brother that is built for 20 to 40mA LEDs.

 

[Flyback]

Thanks, that TPS61165 looks very good. For a 36 LED rear light cluster, I would use three of them, each one as a sepic powering 12 of the LEDs.
The internal FET current limit can be as high as 1.44V, so to prevent overshoot etc on start up, (or when an indicator light is flshing on and off), I would short the COMP pin to ground with a BJT so as to effect a soft start every time the lamp flashed back on or turned back on.

TPS61165 is $0.84 on digikey @3000 pieces, and such a solution would be cheaper than a linear regulator solution.

The datasheet actually says its for a Rear car LED light usage, so I reckon manufacturers must already be using it. What I cant understand is, when such simple , cheap SMPS solutions are available, why people are considering linear regulators for rear car lights.?

 

[ronsimpson]

Data sheet Fig 18 shows a way of surviving much higher input voltages and reverse voltage.

I don't remember "over shoot". There is current limit during power up for a short time. Until the output cap gets charged up. (1uF)

SEPIC or Flyback. The flyback pulse needs to below 38V! The flyback mode reduces the p-p voltage on the switch by Vin. In flyback mode the first 12V is (more or less) free. The IC does not have to work as hard. If Vout is low like 12 to 18V then yes SEPIC is a must.

All these designs are thinking 3V LEDs. Are you using white or red LEDs?

 

[Flyback]

thanks, we are using red leds

 

[kinarfi]

Hello,
I need to drive 36 LEDs with 60mA in each LED. (my vin is 13.5V)
I wish to do this in 12 strings of 3_in_series. (so theres redundancy)


I wish to use twelve zxld1360 buck smps control IC's...(one for each led string)

Is this going to be more noisy (EMC wise) than using one 6W boost SMPS LED driver, and putting the 36 LEDs in series?

For get the power supplies and use 9 strings of 4 tied directly across the 13.5 v
https://www.dropbox.com/s/qa3lc3ukf36l7d6/20121212_111859.jpg
here's a photo of 6 strings of 4, 3 watt leds.
Are these being used on a vehicle?

 

[Flyback]

Yes it is for rear/brake light on car

 

[kinarfi]

Yes it is for rear/brake light on car

The third light that is in the center? What's the voltage rating on each LED, you already said they are 60ma.

 

[Flyback]

voltage nominal on each led is 2.5V
-but in cold ambient it could be as high as 3.3V.
-and of course, if our suppliers bump up the price of the leds, then we will want to change to a cheaper one, and that could be up to 3.5V max in cold ambient.

Its the leds that do the rear light, (with 25mA in the leds...) and then when the brakes are pressed, they get the full 60mA.

 

[ronsimpson]

I think there are two different circuits.
A 1Watt 'bulb' connected to one wire.
A 5Watt 'bulb' connected to the other wire.
And a ground wire.

 

[kinarfi]

I don't know the values, but the one I tore apart worked like this: the tail light voltage fed 4 out 7 lights via 2 resistors in series, when the brakes were applied, the brake voltage was applied to all 7 lights via 1 resistor each, thus upping the current of the 4 that were on as tail light from 25 ma to 60 ma and also feeding 60 ma to the other 3 that aren't powered up by the tail power, there were some diodes in there too, to prevent back feed from tail to brake or vice versa