delabs Circuits

Monday, January 19, 2015

Mystery of the Switching Paradigm

I was not able to solve a issue with an Enthusiast who tried building a Power Circuit Design of mine. It was a design that was not complete in documentation. I had just put it online for people to study. Marin built it and went thru a lot of trouble.- This was in 2005-6

Story Starts Here -

I built switching power supply with SG3524 (project del20020) from your site

It is working, but not good :-( Voltage and current regulations are very bad. I connect  stabilized 13V to input, and from another stabilized power supply I connect 12V for SG3524 and LF358. For testing I use mosfet BUZ11, and various ferrite inductors, but results are bad.

On output I cant get more than 6V, and output 470uF elco sounds like "bzzzzzzzzzzzzzzz"   I checked, and rechecked all connections several times, but everything is like on your schematic diagram.   Could you tell me if your schematic diagram has any bug?

Did you built any working switcher with this schematic?   Maybe is problem in inductor. Is it critical? I tryed with few toroid and ETD44 EI core with 4 to 40 turns of wire, with no luck :-( Which corre and numbers of turns do you suggest?

Power Supplies Section

Mail from Ing. Marin Antic

Mystery of the Switching Paradigm


delabs Reply

This i  a 15 year old circuit and working well, there may have documentation errors. i will look at it when i have some time. meanwhile troubleshooting is better.

The circuit itself can have errors, and then while you breadboraded there could be mistakes. the inductor is required for proper working around 50uH you will need a LCR meter or make a test jig.

Where is the bzzzz sound coming from is it a lose core of inductor vibrating. you said voltage is 6V does it give good curent, is load regulation good.   the input power needs to be 3-4 volts more than what you need a t output try 15V and also does input power suppy have good current rating like 5 is the input power supply able to able to drive the load you want. probably the input power supply may be loaded.   see for any heating parts. the problem is in the connections or the feedback path. look around the LM358, you can chat with me online. i will certainly try my best to help you.



I think that your schematic can't work good. Do you have continuous voltage and current regulation from 0-12V and 0-5A on your swicher? I think it is impossible with this schematic...

When I modificate sg3524 output stage to the mosfet, I got full voltage on the output not only half like before. But voltage regulation is not continious. Now I can regulate voltage only 6-12V. 0-6V not working. On 6V is step, and then continious regulation.

 I tryed many modifications on connecting LF358 to the output, an now is working. Not very good, but good :-)   Current control is not working ok. What do you think to move current sensing on positive output, not on negative?

mail from IMA


delabs Reply

Just check these points.

vary inductor air gap, make it also more like 100uH. Keep current limit to minimum and try. Or to locate problem disconnect pins  4-5 and then connect 4-5 to ground. Now current control has been disabled, see if vtg control and load-line regulation ok.   C2 and R10 fix frequency of SMPS 2.2k and 10nF plastic, verify.   Mosfet drive part may have to be modified bit as per circuit attached.   try all in the order show or as your understanding..

(see a application note below, not the circuit being discussed but the chip)

sg3524 smps