Type 3 compensator for forward/buck converter

Question

I am trying to build a isolated forward converter with voltage mode feedback. For some reason, I am unable to achieve required crossover frequency. To simplify everything, I have change the circuit to a buck converter. To simulate it, I am using LTspice with FRA analysis and LT1247 as a controller.

I am using the part in voltage mode as described in an article from TI (for similar part):

Designed parameters:

• Vin = 8 V
• Vout = 3.3 V
• Ioutmax = 1 A
• fsw = 200 kHz
• Lout = 100 µH
• Cout = 100 µF

As I am using ceramic capacitors, I assume ESR of output cap to be zero. And therefore I need a Type 3 compensator.

As the filter has a peak at 1.6 kHz, I put a double zero at half - 800 Hz. Double pole at half of the switching frequency - 100 kHz. Crossover frequency to the 10% of the switching frequency - 10 kHz.

Then I use these equations to calculate all components:

The problem is that when I simulate the circuit, no matter what I set as the gain of the EA, I always get the crossover frequency similar to the frequency of the double zero.

For gain = 10, fz = 800 Hz, fco = 10 kHz, fp = 100 kHz:

For gain = 100, fz = 800 Hz, fco = 10 kHz, fp = 100 kHz:

For gain = 10, fz = 3000 Hz, fco = 10 kHz, fp = 100 kHz:

For gain = 100, fz = 3000 Hz, fco = 10 kHz, fp = 100 kHz:

The simulated circuit for gain = 10, fz = 800 Hz, fco = 10 kHz, fp = 100 kHz:

Link for download of the circuit: LTspice circuit

Can you please help me out where I made a mistake?

EDIT: I have changed the circuit to the one I have actually build. The results of the simulations are exactly the same.

I have calculated the gain of the power stage to be at low freq: Apwr = Vin / Vramp = 8/0.9 = 19 dB.

I also verified the gain by setting the gain of error amplifier to 1 and appling low frequency sin source to the amplifier and measured the output:

The simulated gain = 2.22/0.2 = 20 dB

I used the gain and LC filter parameters to simulate the circuit in Simplis: I had to lower the gain at crossover parameter a little bit to achieve desired frequency. The results:

Simplis circuit

Then I used the calculated parameters to the LTspice circuit. Output voltage of the LC filter:

The problem is that with the same parameters the circuit is not stable and not able to achive the stable output voltage.

I have measured the circuit in the lab, the output voltage is stable, but I measured the crossover frequency to be about 300 Hz.

So what am I doing wrong? Thank you

Answer 1

As pointed out by Hearth, the LT1578 is a peak-current-mode switcher so a simple type 2 compensator is needed. Before compensating the converter, you need to extract its control-to-output transfer function. For that purpose, you have to plot \$\frac{V_{out}}{V_c}\$ first and choose a crossover frequency \$f_c\$. You then extract the magnitude and phase at \$f_c\$ and, from there, you determine the components values of the type 2 compensator.

You can either take a look at my last book on loop control, or run the below automated template in LTspice that I will soon upload on my page. The calculation procedure is automated and the dc sources in the bottom of the schematic, display the computed values:

If you probe the op-amp output, you have a typical type 2 response:

In this example, I had a 10-dB gain excess at 1 kHz so the macro determines the components values to produce a 10-dB attenuation at 1 kHz with the necessary phase boost for meeting the phase margin requirement. A type 3 compensator will also be uploaded.