Hi Robert, i have a question regarding the output transfer function of a compensated power supply, ¿why does it has to be or look like a straight line from the 0Hz point which is DC level and starts high in gain and from there goes down with a certain slope in a straight line? why a straight line? instead of just having high gain at 0Hz and then not including any frequency?
Hi Carlos, that is a good question. The high gain is needed at DC. What the high gain does is that it lowers the error voltage at the output of your supply. So let say you have a Vout of 5V. If you gain is low then it would have maybe 5.123V. But with a high gain then it drive the 5.123 towards 5.000V. The high gain is set by the integrator capacitor. The integrator capacitor is the one connected in series with a resistor (integrator resistor). At low frequency, the Xc of the capacitor is dominate because it has a higher inpedance than the integrator resistor. So the gain will be Vgain=Xc/Rdivider. Then as the frequency is increased, the integrator capacitor will become a lower impedance and then the integrator resistor will dominate and that is where you have the midband gain.. I hope this make sense. RB
@@RobertBolanos ok but then why not use a linear transfer function with no capacitors, why should there be poles and zeros? Can't a linear amplifier give the necessary gain at any frequency?? Thanks for taking time to explain my questions Robert 🙏
@@carlosornelaslim9713 No you can not use a linear amp . If you use current mode, you have to use a type two compensation which a integrated cap in series with integrator resistor. In parallel to the integrator cap and resistor you need a rolloff cap. The output filter of the supply has a pole and at higher frequency, you have a zero. The type two pole and zeros are selected to cancel the pole and zero of the output filter. Please see my videos on loop compensation and it will make sense. Rb
Hi Robert, This is Victor Castrellon. Is it possible you provide me the file with this circuit in microcap format? I am trying to get familiar with microcap and I am having troubles in making the file to work. Thank you
The intent was to show the instability of the input filter if it is not damped properly. To properly design an input filter, you can watch the procedure that I posted on my channel. Keep in mind, that even if I had tried to lower the impedance of the input filter, if the filter is not damped, you still have a chance that the resonant may be higher than the input impedance of the supply. Damping the input filter should always be followed.
@@RobertBolanos both criteria have to be met to prevent oscillations is what I was trying to say. I have seen the other video as well, it's very well done.
@@hispaniola1684 Yes, that is correct. That is Middlebrook's criteria. I think I read somewhere that Middlebrook solved this problem on US submarines were having instability problems when they started connecting many supplies to the main submarine bus. Funny how things happen.
Thank you for your kind comments. Best regards, Robert
It is amazing that in 10 minutes you can explain this, I like the rule of thumb first and then maths later
well explained. thank you :)
Sir, what is voltage rating of capacitor (47uF) and power rating of resistor of series rc ?
Please suggest
Hi Robert, i have a question regarding the output transfer function of a compensated power supply, ¿why does it has to be or look like a straight line from the 0Hz point which is DC level and starts high in gain and from there goes down with a certain slope in a straight line? why a straight line? instead of just having high gain at 0Hz and then not including any frequency?
Hi Carlos, that is a good question. The high gain is needed at DC. What the high gain does is that it lowers the error voltage at the output of your supply. So let say you have a Vout of 5V. If you gain is low then it would have maybe 5.123V. But with a high gain then it drive the 5.123 towards 5.000V. The high gain is set by the integrator capacitor. The integrator capacitor is the one connected in series with a resistor (integrator resistor). At low frequency, the Xc of the capacitor is dominate because it has a higher inpedance than the integrator resistor. So the gain will be Vgain=Xc/Rdivider. Then as the frequency is increased, the integrator capacitor will become a lower impedance and then the integrator resistor will dominate and that is where you have the midband gain.. I hope this make sense. RB
@@RobertBolanos ok but then why not use a linear transfer function with no capacitors, why should there be poles and zeros? Can't a linear amplifier give the necessary gain at any frequency?? Thanks for taking time to explain my questions Robert 🙏
@@carlosornelaslim9713 No you can not use a linear amp . If you use current mode, you have to use a type two compensation which a integrated cap in series with integrator resistor. In parallel to the integrator cap and resistor you need a rolloff cap. The output filter of the supply has a pole and at higher frequency, you have a zero. The type two pole and zeros are selected to cancel the pole and zero of the output filter. Please see my videos on loop compensation and it will make sense. Rb
@@RobertBolanos Ohhh okay, so if i use a linear amp I'm running into a possible instability situation and my power supply can start oscilating??
@@carlosornelaslim9713 yes, that is why the type compensation is used to cancel the pole and zero and make the power supply stable.
Can you create video about PFC Circuit analyse and design ? Thanks for smps videos
Hi Erden, yes my goal is to make a video about PFC when I get more time to do it. Thank you for the feedback. Robert Bolanos
@@RobertBolanos thanks sır, for answer. You are very helpful. 👍
Hi Robert, This is Victor Castrellon. Is it possible you provide me the file with this circuit in microcap format? I am trying to get familiar with microcap and I am having troubles in making the file to work. Thank you
You didn't try to drop the impedance of the filter... Or at least determine the impedance of the power supply section.
Why?
The intent was to show the instability of the input filter if it is not damped properly. To properly design an input filter, you can watch the procedure that I posted on my channel. Keep in mind, that even if I had tried to lower the impedance of the input filter, if the filter is not damped, you still have a chance that the resonant may be higher than the input impedance of the supply. Damping the input filter should always be followed.
@@RobertBolanos both criteria have to be met to prevent oscillations is what I was trying to say. I have seen the other video as well, it's very well done.
@@hispaniola1684 Can you please explain what two criteria you are talking about? I not sure I understand.
@@RobertBolanos the output impedance of the filter should be lower than the input impedance of the converter, and the filter should be damped.
@@hispaniola1684 Yes, that is correct. That is Middlebrook's criteria. I think I read somewhere that Middlebrook solved this problem on US submarines were having instability problems when they started connecting many supplies to the main submarine bus. Funny how things happen.