LTspice model for COB LED

[Flyback]

Hi,
Do you know how I can get a LTspice model for the LC019D and LC033D LEDs from Samsung?
Ie something that gives the same I,V curve when a DC sweep is done?
(pages 26,27 of datasheet)
**broken link removed**

I did the attached but its not accurate

 

[eTech]

Hi,
Do you know how I can get a LTspice model for the LC019D and LC033D LEDs from Samsung?
Ie something that gives the same I,V curve when a DC sweep is done?
(pages 26,27 of datasheet)
**broken link removed**

I did the attached but its not accurate

Hi

Check back here in 24 hrs and will post a model(s).

eT

 

[alec_t]

How about this?

 

[ronsimpson]

The how is interesting to me.
It looks like you need 40 "diodes" to get a part that starts to conduct at 30 to 32 volts. But the slope of the curve is too flat at 1V/amp. Adding 2 ohms will increase the curve by 2V/A.

 

[eTech]

Hello

The Leds your interested in can be built using the LC003D as a base model.
I've determined the IV curve for the LC003D. All you need to do is use the "M" parameter to connect the LC003D in parallel to produce the other models. See example below.

eT

 

[eTech]

Hi,
Do you know how I can get a LTspice model for the LC019D and LC033D LEDs from Samsung?
Ie something that gives the same I,V curve when a DC sweep is done?
(pages 26,27 of datasheet)
**broken link removed**

I did the attached but its not accurate

See post #5

 

[Flyback]

Thanks alec_t.....your way of doing it is the onlyu one i seem to be able to use.
I did another one for the LC019D as in page 26 of the top post datasheet. Its attached.

I suppose it uses the standard diode model of LTspice, and then the resistor is added to increase the gradient as required.

eTech i like your approach with the model statement, i must admit i havent been able to make it work for me....certainly due to my lack of skill only

 

[alec_t]

I suppose it uses the standard diode model of LTspice, and then the resistor is added to increase the gradient as required.

Yup. No bells and whistles.

 

[eTech]

Thanks alec_t.....your way of doing it is the onlyu one i seem to be able to use.
I did another one for the LC019D as in page 26 of the top post datasheet. Its attached.

I suppose it uses the standard diode model of LTspice, and then the resistor is added to increase the gradient as required.

Yes...but it isn't very accurate since you asked to produce the diode IV curve.

eTech i like your approach with the model statement, i must admit i havent been able to make it work for me....certainly due to my lack of skill only

Whatever works for you

The way I've done it is really very simple.

The LEDs with the multiple strings of LEDs in parallel are all made from the LC003D (this is deduced from the datasheet).

Once the IV curve for the LC003D (12 LEDs in series) is determined, just connect multiple LC003D in parallel (using the "M" parameter) as needed to produce the remaining models.

In the Post#5 example for the LC019D, this LED has 6 series strings. So using the diode model for the LC003D, and simply added M=6 at the end of the model name specified in the symbol.​

Anyway....glad you got what you needed...

eT

 

[Flyback]

Thanks eTech, i must admit though, the M=10 version for the LC033D COB doesnt seem quite accurate, as attached.

 

[ronsimpson]

doesnt seem quite accurate

The voltage for a LED is not very accurate. There are min/typical/max numbers. When doing a production design you need to test all three conditions.