Type two compensator in flyback converter for LED driver in DCM mode.

[Patrick66]

Hello everyone, I would like to ask why is the type two compensator different from the error amplifier in the video (Figure is shown below) compared to Figure 1. Based on Figure 1, the Vo is connected to form a loop with Ve whereas the video in PE# 56 Vref is connected with Ve to form a loop. Is it the same? Thank you in advance.

Video Link:

[Terry Bites]

The error amp provides "proportional control" of the converter.
But the converter has a complex phase and frequency response.
To compensate for this you need to compensate the phase and frequency repsonse of the feedback loop.

If you want to get into the nitty gritty of control loop compensation in LED supply read the attched. Yawn...

[psydaddy]

The comparator itself lags in response to adjust the output because the feedback loop has a limited response speed, so it cannot adjust the output instantaneously. The compensation network is added to shape the loop response and maintain stability.
Without proper compensation, the PSU can overshoot and undershoot current to the output during very short intervals. This can cause oscillations or a triangular/ripple-like waveform that may look like ringing. The compensation network reduces these oscillations and gives the loop a more stable and predictable response.

[mtwieg]

I don't think the OP is asking about the purpose of compensation in general, they're asking about difference in implementations of the error amplifier (inverting vs non-inverting).

Error amplifiers can be made using inverting and non-inverting opamp circuits. The main difference is, obviously, the 180 degree phase shift. In that flyback circuit, the optocoupler circuit acts as an inverting stage (when Ve rises, Vc falls). If the error amplifier were an inverting opamp, the overall feedback would not be negative, and the flyback would not work.

But there's also a more subtle difference between inverting and non-inverting amplifiers. For the same component values, their transfer functions (pole and zero locations) are not the same. Inverting is H = Z2/Z1, non-inverting is H = 1+ Z2/Z1. This means non-inverting error amplifiers have a limitation that their gain is always has magnitude of 1 or higher, which can be problematic in some cases (sometimes you want gain to go towards zero as frequency goes towards infinity). Another way of looking at it is that the non-inverting circuit will always have equal number of zeros and poles.