• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

LEDs in parallel in tube light will share current well?

Status
Not open for further replies.

Flyback

Well-Known Member
Hello,

We have had a LED tube light designed for us, vin=20v-140v, Output = ten paralleled banks of four_LEDs_in_series. It’s a Buck converter with output current regulated to 1.67A
All LEDs are on a long piece of MCPCB. The LEDs are 2.5cm apart.
No dropping resistors at all.

The designer told us that we will not get thermal runaway in any of the paralleled LED strings, because all the LEDs are in this totally unventilated tube, and so they will all be “frying” in each others heat….(so none can ’run away’ thermally). Do you agree with this?

Also, the LEDs are 700mA Luxeon Rebel Plus LEDs, but are only run at 166mA each, and LEDs operated at the lower end of their V/I curve share current better when paralleled. Do you agree?
 
Last edited:

alec_t

Well-Known Member
Most Helpful Member
I'm not convinced about the thermal runaway. IF all LEDs are well-matched and IF they all remain at the same temperature there's obviously no problem, but if (for whatever reason) there's a temperature gradient in the tube then the hotter LEDs will have a lower Vfwd so will tend to hog the current and get hotter still. Even a small temperature difference could have a big effect on LED current sharing.
TempGrad.PNG
 

Flyback

Well-Known Member
thanks, the designers have told us that there could not be a temprature gradient in the tube, because its a small tubular volume, and all the leds sit on the same piece of MCPCB, and even though they are 2.5cm apart, its inconceivable that there could be any significant temperature gradient between any one LED and any other.
 

audioguru

Well-Known Member
Most Helpful Member
LEDs in parallel must have their forward voltages matched. Having three LEDs in series for each string helps make their forward voltages more even but what if one string of 3 LEDs has a lower forward voltage and another string has a high forward voltage? Then the lower voltage string will burn out soon and the higher voltage string might not even light.
 

Flyback

Well-Known Member
Thanks, i see, ..so, if we put a 1 ohm 1206 in series with each string of leds, then that drops 100mV for every 100mA of extra current through it.....and looking at the page 9 datasheet Vf vs I curve, this will mean quick equalisation of the currents in the LEDs.

Luxeon rebel plus led
http://www.lumileds.com/uploads/380/DS107-pdf

...i am sure anyone reading this will realise that where led tube lights are concerned........there is only one topology for driving the leds...a simple buck converter....anything else is over-engineering. Do you agree?
 

alec_t

Well-Known Member
Most Helpful Member

Flyback

Well-Known Member
thanks, agreed, i should have said i was talking about cases where the vin youve got to play with is at least 20v or so.....because then you can just shovel in as many 3 or 4 led strings as it takes.

In truth, with a 20-140v input, i woudl have personally done a bukboost and put the leds in a single series string to the tune of 20w.......but everyone seems to be using bucks these days and just using shovel in enough paralleled strings to be able to do that.
 
Last edited:

ChrisP58

Well-Known Member
Most Helpful Member
The LEDs on the edge will be at a different temperature than the ones in the center of the board.

Instead of paralleling a bunch of series strings, I would have series'd a bunch of parallel clusters. And made those parallel clusters thermally monolithic.
 

Flyback

Well-Known Member
And made those parallel clusters thermally monolithic.
...thanks, but this is beyond our cost limitation
The LEDs on the edge will be at a different temperature than the ones in the center of the board.
Thanks, but in a totally enclosed tube, with all the leds sitting on the same bit of aluminium, surely they will all be pretty much the same temperature...i mean, i cant see the ambient air temperature being more than a degree different at the ends or the centre?
 

ci139

Active Member
LEDs operated at the lower end of their V/I curve share current better when paralleled. Do you agree?
the V-I curve /¯¯ ← is like there so at low power dU is relatively greater than dI (a current diode) and vice versa for higher power (a zener) - but that does not state much about the current sharing -- what i know is that at near max. power the voltage drop may vary some 500mV or even more for the same type 3V LED-s (and not the cheapest ones) as in fact it varies a lot less for low brightness cheap 2V LED-s
 

Flyback

Well-Known Member
Thanks, page 20 of the following LED datasheet shows just how low the dynamic impednace gets as current rises in the LED.
If paralleling leds, then we should work in the high z region, do you agree?
After all, in that region, any LED that hogs I will see its vf go more rapidly up, which in turn will bring other leds into conduction, and stop the hogging. Do you agree?

http://www.lumileds.com/uploads/571/DS144-pdf
 

alec_t

Well-Known Member
Most Helpful Member
any LED that hogs I will see its vf go more rapidly up, which in turn will bring other leds into conduction, and stop the hogging. Do you agree?
No. It won't stop the hogging, but it will limit it at some equilibrium point.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top