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

[Flyback]

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?

 

[alec_t]

I'm guessing
1) the total noise energy would be similar for the two cases,
2) the twelve smps would not be switching at exactly the same frequency, hence their energy would be spread more widely over the RF spectrum, albeit with lower energy per smps.

 

[Flyback]

..So basically can we conclude that the conducted and radiated emissions would be no worse with twelve half Watt SMPS's , rather than one 6W SMPS?

 

[4pyros]

I think its mostly going to depend on the quality of the SMPS.

 

[dr pepper]

You could say 12 cheap psu's would create an electrical racket whereas 1 reasonable suply would be quieter, later offline control chips use frequency jitter which reduces radiated noise a lot.

 

[Flyback]

By "electrical racket" I presume you mean the 12 0.5w PSU's will beat with each other?
The beat frequencies will all be less than the individual switching frequency, and so should not pose a radiated emissions threat, and with each only being 0.5w, I doubt there will be much of a conducted emissions threat?

 

[dr pepper]

Sort of.
Your points are valid.
My main point was the school of thought that it is better to have one well designed supply instead of several cheap design ones.

 

[ronsimpson]

I need to drive 36 LEDs with 60mA in each LED. (my vin is 13.5V)

So this is a auto (car) project? 13.5V
What color of LEDs? White?
Notice that the IC is limited in what Vin and what Vout it likes. If this is 'auto' then Vin will vary over a large range. Vout will change with LED batch. See LED data sheet.

I wish to do this in 12 strings of 3_in_series. (so theres redundancy)

Sounds like a tail light or head light. "redundancy"??
Is redundancy more important than cost? Twelve strings seem like over kill. Would 2 or 3 or 4 "X" be ok?

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

Nice little part. Know it well.

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

IC-pin1 has sharp edges to the voltage. Keep that area small. also There are current flow loops keep them small. I would add inductors to power and ground to isolate this project from the rest of the world. (make LC filters) like on a power line switching power supply.[/quote]

 

[simonbramble]

IC-pin1 has sharp edges to the voltage. Keep that area small. also There are current flow loops keep them small. I would add inductors to power and ground to isolate this project from the rest of the world.

Pin 1 will have sharp edges - agreed, so you need to put a ground plane underneath this node and keep it physically small. However, this is a boost architecture (or at least the schematic on the front page of the datasheet is), so the inrush current will be smooth (you have a 47uH inductor limiting the inrush current). Likewise with the current flowing in the LEDs on the discharge cycle of the LEDs. This will be smooth too. You will get a high di/dt through the rectifier Schottky diode though, so keep this line short and don't run any tracks parallel with it. The LTC3755 or the LT3518 also will do a neat job of this too
https://www.linear.com/product/LT3755
https://www.linear.com/product/LT3518

Simon

 

[Flyback]

It is for auto motive brake/rear light.
I am not sure exactly how many independent led drivers the regulatory bodys insist on for a rear/brake LED light. If I could get away with one smps LED driver then I would. The thing is, most, if not all rear/brake led lights use about 12 strings of three_leds_in_series and a linear current regulator in each of the twelve strings (as attached). To my mind this is madness, but everybody I speak to says its done like this because the regulatory bodies like to see 12 independent current regulators, so if one fails, the others may still work.
-It doesn't make sense to me, but just about every car on earth that has LED lights for rear/brakes/indicators, uses linear led drivers. There must be some reason for this, but I cannot work out what?


The ZXLD 1360 is buck only...it has a low side fet.

 

[Nigel Goodwin]

There must be some reason for this, but I cannot work out what?

What's to work out? - cost, simplicity and reliability.

 

[ronsimpson]

I have seen stop lights with 1/4 to 1/6 of the LEDs out. Redundancy is fine with me. I would think 4x or even 2x is find.

Auto is hard to do. I made police car computers for a Texas company years ago. They put a prototype in a black car, faced it south, in the summer, ran the heater full on, and left it on all day. The plastic case milted down onto the electronics but the computer never failed. We went into production. A year later I found that it would not start up at -60F. (I tested it at -40F) It would run at -60 and start at -40. I changed a resistor to drive the base of the switching power supply transistor harder and all is fine. From Texas to Quebec is a long road.

ALSO

13.5v is not the supply voltage. It can be almost anything. We put recorders on cars, buses and trucks to monitor the voltage. When it is cold the batteries are not good. I have seen people "jump start" with the cables on backwards. Some truckers and bus drivers will "24V" jump start. It is too common to use a electric welder (DC) to jump start or charge. When a battery gets old and goes high resistance the alternator will output a high voltage. I am not saying your light should work at 24V, I am saying it should survive 24V for a fraction of a second.

 

[Flyback]

I would have thought the smps version is cheaper and more reliable. -Because there is less heat, and also, less need for heatsinking, ie lower weight of copper can be used on the pcb, etc etc.
Indicators /brakes and stop lights are all in the same enclosure, and if all done with linear regulators, then that's too much heat in a confined, non ventilated space.

 

[dr pepper]

I put together some lights for my pushbike, the drop in volts was so low I didnt bother with a switcher as there wasnt a lot to be gained.
The same could be true for your project esp if you have the option of wiring the leds in series/parallel to get close to the battery voltage.
Someone said that commercial vehicle headlights that use leds are driven with pwm, and over time the pwm is lengthened to make up for the reduction in brightness of the leds over their lifetime.
I didnt think leds 'wore' out, that might be a load of cobblers.

 

[ronsimpson]

I didnt think leds 'wore' out, that might be a load of cobblers.

Some LEDs use phosphor to get the right color of light. I cant prove it now but I thought the phosphor in white LEDs has a ware out function.

 

[alec_t]

I didnt think leds 'wore' out, that might be a load of cobblers.

They do indeed wear out (fade with time). On a previous car of mine the security active LED on the dash had become quite dim after ~8 years (~70k hrs) continuously on, so I replaced it.

 

[dr pepper]

Explains why th eled on my psu I built 20 years ago isnt as bright as I thought it used to be.
The led on the security system on my wifes old banger >10 years is still nice and bright though.

 

[Flyback]

Yes LEDs do "wear out", getting about down to 70% of initial brightness. Though it does depend on the heat to which they are exposed throughout their lifetime...hotter = wear out quicker

 

[Nigel Goodwin]

I would have thought the smps version is cheaper and more reliable. -Because there is less heat, and also, less need for heatsinking, ie lower weight of copper can be used on the pcb, etc etc.
Indicators /brakes and stop lights are all in the same enclosure, and if all done with linear regulators, then that's too much heat in a confined, non ventilated space.

You would think wrong - as you're presumably a university graduate with no practical experience?.

Switch mode circuits are more complicated and less reliable, and have been since the first ones (first domestic use was in the Thorn 3000 colour TV).

Switch-mode supplies are used because they have a number of advantages - not all apply to all types though.

1) Smaller.

2) More efficient.

3) Lighter.

4) Cheaper - assuming you're replacing large heavy components with small light ones (such as mains transformers).

Anyone with practical experience knows how unreliable SMPSU's are - with the majority of electronic faults occurring in the SMPSU section of faulty devices.

 

[Flyback]

Anyone with practical experience knows how unreliable SMPSU's are - with the majority of electronic faults occurring in the SMPSU section of faulty devices.

For a low power SMPS like for brake/rear/indicator LED lights, such SMPS's can be very easily designed and made without electrolytic capacitors, and without overstressing any of the SMPS components in any way...such SMPS's will last far longer than linear regulators used for the same purpose.
Heat is the worst enemy of electronics components that are not electrically overstressed.
Linear regulators will result in hotter running electronics than SMPS's.
If we are saying SMPS's really are that unreliable, then howcome they are starting to get used for front LED headlights?....the power here is 20 to 40W, and at this level as im sure you agree, linear regulators are totally out of the question.
Even cheap switch mode CFL drivers are lasting for years and years....And SMPS's stuffed into the base of LED lightbulbs are lasting again for many years.

To do linear regulated LED brake/rear lights, one really needs regulation in every string of three leds, and that ends up being a fair few components.

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

As to the question about experience..i have worked in 18 different electronics co's...and have a few designs out there in the market.