LM3414 LED driver doesnt work for me

Thread Starter

arman74

Joined Jul 1, 2017
44
Hey guys.
Im designing a constant current LED driver with brightness control with pwm.
I want to use lm3414.In the datasheet it has been said that if i put a 7k resistor at the iadj pin the led max current will be limmited to about 400ma.But when i increase the pwm duty after the current reached 400 ma it wont get higher but when the pwm is rising the IC explodes and the output will connect to ground and the LEDs damage because of this. I have raised the inductor to 100 uh but the same problem appeared.
What is the problem?
I tried with 2 series 3 watt leds and with one 30 watt leds.
 

ronsimpson

Joined Oct 7, 2019
2,985
First: Good pictures.
Second: the LM3414 is not used like the data sheet.
R10 1k should be 0 ohms. Remove.
Where is the flyback diode?
1575055795149.png
Values missing.
 

Thread Starter

arman74

Joined Jul 1, 2017
44
Second: the LM3414 is not used like the data sheet.
R10 1k should be 0 ohms. Remove.
I placed this resistor between LX and Vin and the value is about 670k ohms.I saw it in one of the led drivers I had.

Where is the flyback diode?
I forgot to place that in schematics and i placed it manually on the board as below:
1575057991374.png
1575058014144.png


1575058256884.png
Above is the values .
PWM frequency is 10 kHz.
 

eetech00

Joined Jun 8, 2013
3,856
Hey guys.
Im designing a constant current LED driver with brightness control with pwm.
I want to use lm3414.In the datasheet it has been said that if i put a 7k resistor at the iadj pin the led max current will be limmited to about 400ma.But when i increase the pwm duty after the current reached 400 ma it wont get higher but when the pwm is rising the IC explodes and the output will connect to ground and the LEDs damage because of this. I have raised the inductor to 100 uh but the same problem appeared.
What is the problem?
I tried with 2 series 3 watt leds and with one 30 watt leds.
Which IC explodes? Which LED(s) is damaged? Which output connects to ground? Where do the external LED(s) connect?

eT
 

ronsimpson

Joined Oct 7, 2019
2,985
I do not understand R10 and D4. The IC can not work like that. The LED current must flow through R10. You limited the current to just uAs. The diode is in the wrong place. What is wrong with making the circuit in the data sheet?

10khz? Why so slow?
1575325505548.png
 

TeeKay6

Joined Apr 20, 2019
573
I do not understand R10 and D4. The IC can not work like that. The LED current must flow through R10. You limited the current to just uAs. The diode is in the wrong place. What is wrong with making the circuit in the data sheet?

10khz? Why so slow?
View attachment 193603
@ronsimpson
Except for the addition of "R10" (that at 670K does nothing), I believe that the schematic of post#5 does agree with the datasheet and the schematic you posted. I do agree that it is drawn in a difficult to comprehend fashion.
 

djsfantasi

Joined Apr 11, 2010
9,156
I placed this resistor between LX and Vin and the value is about 670k ohms.I saw it in one of the led drivers I had.


I forgot to place that in schematics and i placed it manually on the board as below:
View attachment 193417
View attachment 193418


View attachment 193419
Above is the values .
PWM frequency is 10 kHz.
The output sub-circuit doesn't match the circuit on the datasheet. You have the LX output from the junction of R10 and D4. That’s curious as the sample circuit doesn’t have a resistor there. You’re bypassing the flyback diode and allowing surges from the inductor where it should not go. You may have seen it, but unless you know ehy its thete, you should expect to have problems. Your 3414 may be toast.

On the schematic you shared, I didn’t see ‘led-‘ defined? There may be an error there as well.
 

ronsimpson

Joined Oct 7, 2019
2,985
I didn’t see ‘led-‘ defined?
I think "led-" = "LED".

After much redrawing the schematic, I think it is OK. It is just drawn in a way that is hard to see. Sorry it took me so long to see it.

What is the part number for the inductor? Maybe there is core saturation. That will stress the IC!

What is the part number for the diode? It must be fast or Schottky. It can not be 1N4007. That will stress the IC!
 

djsfantasi

Joined Apr 11, 2010
9,156
I think "led-" = "LED".

After much redrawing the schematic, I think it is OK. It is just drawn in a way that is hard to see. Sorry it took me so long to see it.

What is the part number for the inductor? Maybe there is core saturation. That will stress the IC!

What is the part number for the diode? It must be fast or Schottky. It can not be 1N4007. That will stress the IC!
Then, the circuit he provided is wrong. I can’t see how the circuit he provided matches the datasheet, no matter how I draw it. Or I could be wrong.

And no one has provided a justification for the resistor where it is drawn. It should be in series with the LED and not in parallel with any other components. Or I could be wrong.

Otherwise, the flyback diode is bypassed and back EMF from the inductor could damage the LEDs or the 3414.

It would be nice if the entire LED circuit could be drawn and shared.
 

Thread Starter

arman74

Joined Jul 1, 2017
44
1575800740550.png

The schematics is just as the datasheet and it has just a resistor in parallel with the flyback diode.The value is high and I get this resistor from some other documents. Ive tried without this resistor and the something happened.
I used a fast diode for flyback diode but i have not tried shotcky instead.

Sorry because of my late response , I was a little busy.
 

Attachments

Thread Starter

arman74

Joined Jul 1, 2017
44
Which IC explodes? Which LED(s) is damaged? Which output connects to ground? Where do the external LED(s) connect?

eT
Which IC explodes?

lm3414

Which LED(s) is damaged?

The led connecting to the connector i posted #14.


Which output connects to ground?

I did not understand what you mean.

Where do the external LED(s) connect?

I said thatin post #14
 

TeeKay6

Joined Apr 20, 2019
573
Which IC explodes?

lm3414

Which LED(s) is damaged?

The led connecting to the connector i posted #14.


Which output connects to ground?

I did not understand what you mean.

Where do the external LED(s) connect?

I said thatin post #14
@arman74
It appears to me that you have violated at least one limit given in the datasheet. Specifically, I first note that the 40.2K resistor you have connected to the FS pin of the LM3414 sets the internal oscillator frequency to a range from 420KHz to 580KHz--see section 6.5 (pg 5) of the current TI datasheet; for analysis we can assume the worst case of 580KHz. Next I call your attention to section 7.3.2 of the datasheet wherein is given a limit for the LED forward voltage as this inequality: (Vled/Vin) >= (400ns*Fsw). I asked in a previous post for the value of Vin, but you did not reply; let us assume 24V. Further, you did not state the forward voltage of the single "30W" LED you tested, so I will assume Vfwd=2.5V. Applying these values to the formula given in section 7.3.2 we have:
\[ Vled/Vin >0.4us*580KHz\]\[2.5/24 > 0.232\]\[0.104 > 0.232 \]
It is clear that the inequality is not satisfied. If Vin > 24V, then the error is worse. The most likely result is that the LM3414 therefore failed to control the LED current, and that overheated the LM3414 to the point of explosive failure and the LED overcurrent destroyed the LED.

Lesson: Read and obey the entire datasheets for all devices. I have not done so and make no claim that I have found all violations in this design. To be clear, there could be additional problems.
 
Top