Topics:

**Output filter capacitor requirement in flyback converter** on

Automation Technologies
Start by

Malik

08-21-2013 11:12 PM

# Output filter capacitor requirement in flyback converter

Requirement of output capacitor value in flyback converter reduced if I use CCM (current control mode) as compare to VCM (voltage control mode) controller and experimentally proved in one of my flyback converter stability analysis....

Want to know more about this two mode. It will be great if anyone can post some comments or link which describes mathematics behind this...It will greatly help me to cross verify my understanding and analysis.

08-21-2013 11:12 PM

Top #2

The output cap will be about the same for both, usually. It is determined partly by the output ripple and noise requirements, which is independent of the control mode.

It is also determined by the output impedance. For both control modes, the crossover frequency will be approximately the same, although there are cases where current mode is definitely better if the RHP zero becomes low.

If you are in DCM all the time, the cap will be the same for either current-mode or voltage-mode.

If power supply rejection ratio is stringent, current-mode will be far better, unless the voltage-mode control uses feedforward, a topic being discussed in another thread on this site.

You can find voltage-mode and current-mode equations and software models in the Design Center. Please let me know if you are not sure which ones to access.

08-21-2013 11:14 PM

Top #3

Jurish

08-21-2013 11:14 PM

In the Fly back converter the output capacitor sizing depends on the load current, since the load current is supplied by the output capacitor when the power switch is ON and primary winding of the transformer is storing the energy.

Count = (Iout x Duty cycle) / (ripple voltage x switching frequency)

The control loop is independent of the output capacitor sizing.

08-21-2013 11:14 PM

Top #4

Well, that is the ripple requirement, and you are correct, it is independent of control scheme. Whether you are using voltage-mode, current-mode, analog, digital, none of them affect the ripple on a single cycle.

When talking about step load requirements, though, or output impedance, the control loop performance is paramount. The output impedance is equal to the impedance of the capacitor divided by 1+T. So the harder you push in the loop, the more the output capacitance can be reduced if step load response is your limiting constraint.

On ripple, the picture gets a bit more obscure if you implement a second-stage filter. This will reduce the capacitor requirements on the output.

08-21-2013 11:15 PM

Top #5

Hi..I am talking about step load performance only. I tried to find it from analytical analysis, but did not get enough proof what I found on lab. I do not know, Just a guess could be as it is bcz of DCM flyback converter and with very stable source I have applied. ....(confused)

08-21-2013 11:16 PM

Top #6

I would be cautious of reaching a conclusion based on the first paragraph of proving it on one sample of your flyback converter. I have seen many failures based on a sample of one where you reach a conclusion. Many of the posts have offered very experienced advice for you to review. I wish you success.

Double check the finding based on the one unit. Also expect the performance to change with temperature which is often overlooked.

08-21-2013 11:17 PM

Top #7

Thanks for ur reply. I
agree with you. This is may not be a final conclusion. Just observed on
lab and though to double verify from the this group where best power
supply design engineers participates....I have to go in a long
way.......

Its really need a systematic process to do all the test and record all
the test results.Can not do anything in hurry and draw any conclusion.
So every time any doubt comes in my mind, I tried to get answer from Dr.
Rays design centre or from this group.

08-21-2013 11:18 PM

Top #8

Usually we consider that the capacitor is perfect to determine its value versus the converter parameters (Cout = (Iout x Duty cycle) / (ripple voltage x switching frequency) . this is not true in reality because at the switching frequency, usually the capacitor impedance is almost equal to ESR, so ESR is very important to determine the output voltage ripple waveform and Delta Vp-p. What do Dr Rays thinks about that, please correct me if I m wrong, Thanks.

08-21-2013 11:19 PM

Top #9

For electrolytic capacitors, you are quite correct. The ESR corner is below the switching frequency, so the ESR determines the ripple, not the capacitor value.

If you use MLC capacitors, then the ESR is very low and it is the value of capacitor that counts.

In either case, it is often the step load response that determines the choice of output capacitor, not the cycle-by-cycle ripple. This is especially true if you use a 2-stage output filter to attenuate the switching ripple.

08-21-2013 11:19 PM

Top #10

MLC capacitors are used just about exclusively in point of load buck converters. They are high cost, but high performance, and the main requirement is lots of capacitance in a very small space.

At the other extreme, electrolytic capacitors are cheap, hence most often used in flybacks where cost is the number one objective.

It all depends on what your system requirements are.

08-21-2013 11:21 PM

Top #11

All the discussion points are very valuable for me to go in more depth of the analysis. Thanks all for participating and finding ur valuable times.

What I feel is, the second pole given by the inductor is virtually eliminated in case of current control DCM . Therefore , simple dynamic can be approximated here. Also current control have faster response to load changes rather than voltage controlled. So comparative requirement of output capacitance value is lesser here. What do you think?