Hello LT Support,
I want to use the LT3922 in Buck Mode as an LED Driver for a 12V/1,5A LED (one power LED). The supply voltage is 15V. Additionally I want to use the SSFM option with up to 2MHz switching frequency (base fsw=1,6Mhz*125%).
The goal is to achive very low noise EMI on PCB Level.
Will this application be possible without any issues with the mentioned configuration?
In the datasheet there is an example application (added as attachment). What exactly does the Q1 pnp do? A detailed answer in connection with the functional sequence would be great
Any advices, like which filter to add to the output/ input or which changes to apply to the example application are appreciated a lot.
Thank you!Max K.
Considering the chosen Buck switching frequency, 2MHz, and the output power, 18W, I think you may have heating issues, as your LT3922 overall losses are proportional to its fsw. Such a high switching frequency is interesting in terms of filtering, as it will reduce your filter surface.
The Q1 PNP is basically PWM-ing your Buck output to regulate your LEDs light intensity.
Concerning the filters, a quick calculus on the switching and dimming frequencies will help you out, make sure to have the level of your harmonics well reduce by a lowpass filter for example.
thanks for the reply.
Actually the Q1 is not switching, at least not in my simulation as you can see V(u_fb) green chart. The MOSFET 1(dark blue) is switching with 10kHz and the internal MOSFET(grey) is switching with 2Mhz. (External PWM dimming ist being used). The Voltage at FB is constantly at 0,75V. I kind of understand now the function. The Q1 pnp is inverting the voltage at the LED kathode for a successfull overvoltage lockout through the FB input.
Another question is:
Sorry I mistaken Q1 for M1.
I guess the explanation lies between the Boost and Buck modes of the controller, can't really give you an explanation.
But the resistor may give you a better compensation loop adjustment.
Actually it's a question of phase margin, just try and draw the bode plots of the Vc capacitor paralleled with the current regulation amplifier's output resistor, with and without Vc resistor. You will see that you back from -90° to 0° at high frequencies, which gives more phase margin to the overall loop, and more stability. It may be more important for a Boost converter.
Sopmodtheo thank you for trying to assist maxka. We appreciate your willingness to help other members however, your approach is not the one I would have chosen. The ADI approved approach is to first, use the LT3932 for buck LED driver with low EMI.
The demonstration circuit DC2286A already has EMI filters on it.
The demonstration circuit DC2247A has LT3922 EMI filters on it if you want to use them. If you want to use LT3922 as buck mode, you can. The transistor is used as a level-shift for over voltage protection. The idea is that a voltage is developed across the LED string which pulls a certain current through the upper resistor in the level-shifter.
That current goes through the collector and emitter of the transistor and creates a voltage from FB pin to GND. When the output voltage is too high (LED string is removed), the FB resistor goes to high voltage and the part recognized the open LED condition.
There are two configurations of the level shifter. One has a resistor divider at the base, and one does not. The resistor divider is used at the base of the transistor only when very low dropout in buck mode is needed. When that is needed, the base resistor divider allows the level shifter to still create a high (1.2V) voltage at the FB pin when the LEDs are open. However, if low dropout is not needed (low VIN to high open LED voltage), then the FB pin will not get squished down during an open event and the divider is not needed.
Still LT3932 is a better fit since buck LED driver with low EMI is more obvious solution here.