Linear LED driver works due to emitter degeneration?

[alec_t]

What are Q3 and Q4 in LINREG 1_TL431REG designed to do? As shown it seems that if TL431 fails then Q4 turns on, Q3 turns off and the current through D2,D7 rises to ~470mA .

 

[Flyback]

What are Q3 and Q4 in LINREG 1_TL431REG designed to do? As shown it seems that if TL431 fails then Q4 turns on, Q3 turns off and the current through D2,D7 rises to ~470mA

ok, I certainly don't get that current when I break the connection to the TL431 in either schem........if I short out the TL431, I get no led current, and if I open the TL431, I get about 70mA led current.

if one of the leds from the 2_string fails open, then Q4 turns off, and Q3 turns on and turns off all the leds........that's what its supposed to do.....admittedly its a bit of a strange type of protection circuitry, but hay-ho, it didn't cost as much as a proper driver with fault reporting etc

 

[alec_t]

I certainly don't get that current when I break the connection to the TL431

Depends where you break the connection. Try disconnecting R7.

 

[Flyback]

disconnecting R7

I did that, and I get 64mA in one string, and 54mA in the other

 

[alec_t]

I agree, for the LINREG 1_TL431REF; but I was referring to the LINREG 1_TL431REG sim.

if one of the leds from the 2_string fails open, then Q4 turns off, and Q3 turns on and turns off all the leds........that's what its supposed to do.....admittedly its a bit of a strange type of protection circuitry

Indeed. Why turn off all the LEDs just because one has failed? I don't see the logic in that.

 

[Flyback]

I am also referring to that...I just simulated it... here are the the.sub and .asy files of the TL431

 

[alec_t]

Interesting. I was using an alternative model of the TL431A. I'm surprised the two models give such different results. I think I prefer the model you posted.

Edit: as to your question ....

Please could you advise on why the designer did not use an TL431 to provide closed loop negative feedback

... if you run both sims for different temperatures (I tried 10C to 50C) you'll find (surprisingly?) that the LED current varies less with temperature when the 431 is used without the closed loop feedback.

 

[MrAl]

Hi,

Yes the TL431 can oscillate in some circuits so it takes a bit of doing to get it compensated sometimes. When used as just a regulator is seems easier to work with.