Output Voltage Regulation

Hi,

Your schematic looks very good and I do not see any obvious reason for the load regulation to be as poor as it is. I have only two thoughts right now.

1. How and where are you measuring the output voltage? I would assume you are measuring directly at the output terminals of the power supply. This must be down to make sure you are not getting any kind of I*R drop of the cables that are carrying the actual load current. Since you have different values of different line voltages at the SAME load conditions I am quite sure this is not happening but please check.

2. You have some good layout notes on your schematic which is great to see! However you have used a "GND" symbol at the return of the output and a "GND" symbol at the feedback. Be sure that in the layout these two points are connected via low impedance path. It is possible that load current flowing through the return is causing a potential difference between the feedback circuitry and the output terminals.

If you have a layout you can post it or private message me to get the file to me and I'll take a look. Hope this helps!

-Wesley

Hi Wesley,

Thanks for your reply. I mentioned in the first email that capacitor Cb was removed... well I thought I had. I found that it was still there, and after removing it, the load regulation is much better. What does capacitor Cb do in the circuit?
Sorry I had to remove the schematic from this forum. I'll send you the schematic and layout through email later. Thanks!

Hi Wesley,

Thanks for your reply. I managed to improve the regulation by removing Cb (which is in parallel to the current limitting resistor of the Optocoupler's led on the output side).

When the resistive dummy load was used, the power supply worked fine. However, the moment when it's connected to the actual load, which is a capacitive load, the power supply went to a momentary shut down (or a restart) and came back on again in about a second or so and stayed on.

I was using Cb of 1uF, should I try 100nF? Can you please explain or recommend an article about how the feedback compensation loop works? Are there any formula to calculate what value to use?

Thanks!

Hi again,

I forgot to mention that I also tried having an NTC in series with the load to reduce the initial current drawn, with this it works fine. However, this isn't an ideal solution as the NTC would have been too hot for the next start up.
Thanks.

A rigorous explanation of the feedback circuit is unfortunately out of the scope of this forum. However the Value of Cb, which usually is placed in series with another resistor, forms a phase boost network. This helps to ensure loop stability by increasing the phase margin.

My best recommendation is to use PI Expert to enter your design specifications and it will calculate the necessary values and show a schematic of the feedback circuit recommended.

I ran a design similar to yours and have attached it for your reference. PI Expert does not include a phase boost network for this design so you may want to try removing it from your circuit and seeing how it's performance changes.

I'll get an answer for you about the capacitive start up problem.

-Weseley