Is Buckboost LED driver going to be unstable?

[Flyback]

Hello,
I have designed a buckboost LED driver and have only so far simulated it. (in LTspice)
The schematic is attached.
Also attached is a LTspice waveform from the simulation. It shows the COMP pin and LED current at startup.
(COMP pin is output of error amplifier)
When I calculated the feedback loop with a certain method, it gave a crossover frequency of >100KHz and no phase margin, and yet the simulation runs as smooth as can be.
spec of buckboost led driver is:-
Constant off time,
fsw = 119KHz at 10vin
vOUT = ~35v (9 series LEDs)
ILEDS= 500mA
CCM
LTspice simulation also attached

 

[crutschow]

100kHz is outside the Nyquist frequency (59.5kHz) of your switching waveform and anything that happens above that frequency is not likely to have any effect on the loop stability. Since your output has no filter capacitor, there is not the normal low resonant output frequency due to the output inductor and capacitor tank circuit that you need to compensate. That is why the loop appears stable in the simulation. I would think the actual circuit would behave likewise.

 

[Flyback]

thanks, there is an output capacitor,its 10uF, - if you look at pages 16 and 17 of this datasheet, you can see the equations for the RHPZ and power stage pole of a buckboost LED driver..........(equations 17 and 18)
**broken link removed**
If you calculate it , you see that the RHPZ for this buckboost is 1340Hz, and the power stage pole frequency is 4440Hz.
..As you can see, that is a disaster for doing any kind of loop compensation.

The output capacitor cannot be any bigger due to cost, space and the internal ambient being too hot for bigger electrolytics.

You can see that the RHPZ is a disaster, its too low, due to the low dynamic impedance of the leds.....but...
...I actually think that putting the LEDs dynamic resistance into those feedback loop equations is the wrong thing to do, I think that Resistors R13 and R18 act as an "upper divider resistor"...and for the circuit of the top post, I believe that the LEDs dynamic resistance should not be a factor in the feedback loop equations....do you agree?

 

[MrAl]

Hi,

It looks like on page 17 and above they tell you how to compensate. A capacitor to set a pole at a lowish frequency, and possibly a current sense low pass filter set at a highish frequency. The capacitor of course limits the speed of response to a great degree, but this isnt a bench power supply just an application specific power supply for an LED which although should not be too slow does not have to be super fast either.