Problem with my design

I agree with you. It is very possible that the auto restart is caused by overload, would youlike to double check your transformer and spreasheet?

If everything is ok,then you can try a bigger part.

Regards

Thank you for you reply.

I have checked my tranformer :  Lp = 1170uH but I am not sure about leakage inductance. The value that I measured is 25uH (with all other ouput shorted) but I don't believe that is correct, so my snubber may be wrong.

By the way, I would like to ask about  snubber circuit :

- I design snubber circuit with reference to  AN-4147

- The value of Vsn I measured across R1 is only 170VDC with the load 1.5A . May be it will increases when the load is > 1.5 A or not ?

- With this value of Vsn, I re-calculate the L leakage

           Llk = 2 * Vsn * (Vsn-VOR) / (R1 * fS * IP * IP)

it is only about 3 uH ???

- So how to measure the leakage inductance  with high accuracy ??

That is to say that  I have to re-design the snubber after having the exact Llk.

Waiting for your reply. Once again, Thanks.

Hi,

This is the waveform of V drain with the load approximately  1 A. With the load > 1.5 A, the output flash so I cannot draw the wavform.

The question is : while I have designed the converter in CCM mode but it seems to be working in DCM mode ???

Why and How to modify this converterto work at maximum output = 5 A ??

Thanks.

PLS check the drain current to verify the current limit.  If the current limti is lower than your expectation, then you know the root cause and solve it.

Thanks, 

But I dont know how to draw the drain current because I dont have current prob.

Should I add a small resistor (0.47 ohm) between pri inductance and Drain pin then draw the voltage across this resistor ? 

I guess we use the same way  to measure the leakage inducatnance. We short the secondary winding,and measure the leakage at 100kHz.

The equation could be wrong, because the magnetizing inductance also affect the voltage on the snubber cap, and thus i do not trust the equation. The 170V across snubber cap looks fine for me, since your Vor is around 130V.

Your problem is caused by the load capability. So source the load with a normal electronic load, and verify  if the current limit can meet your expecation. If the current limit and switching frequency are good, then you have to double check the magnetizing inductance and also the polarity of your output cap. As you know, 5A and 1.5A has huge difference.

I observed and saw that the Vsn increases as the load current increases. At load=0 Amp, Vsn = Vor, sure. So I think that I will not care about snubber circuit any more. It may not be the main cause to this problem.

My actual load is a DVR (Digital Video Recorder) but the LED also flashs when sourcing with this load.

I have checked all the things you suggest :

     - magnetizing inductance : Lp around 1150uH (the designed value is 1122 uH)

     - Switching freq : 98.8 kHz (I measured when draw the waveform)

     - Output cap polarity is correct.

Only the drain current I dont know How to measure. Should I add a small resistor (0.47 Ohm) between Lp and the Drain pin then draw the voltage across this resistor ? Or you could advice me another way ?

Thanks.

Hi,

To make it possible to measure the drain current, you can connect a 1 ohm resistor between the source pin to the DC minus, then you can hook you probe between this resistor. The voltage cross the resistor will represent the drain current accordingly.

Oh my God, I have bought a faulty Topswitch. You know, after changing everything as : transformer, output rectifier, output capacitor, ...; the board still flash when loading > 1.5A.

Finally, I decide to buy another Top224Y.  It was unexpected, the circuit is now running very good.

In order to make sure this Top224Y is faulty, I bought some more.  Indeed, one of which has similarly fault.

Now I knew that the topswitch that I bought was not a new one but a second-hand  wich was collected from the old power supply boards.

Has someone face the problem like this ? It is a valuable lesson to learn !!!!

Thanks for all.

Hi,

I did not hear some cases like this before. I am glad that your powre supply is working now.

Thank you,

Just a small problem with the load regulation.

I change the feedback circuit from Zener diode to TL431 as attached picture.

The output voltage is as follows :

    Loading = 0A, output voltage = 12.2V

    Loading = 1A, output voltage = 12.0V

    Loading = 2A, output voltage = 11.8V ( ?? )

    Loading = 5A, output voltage = 10.6V ( ??? )

The Rb = 2.2 k is the best value I tried.

So could you please advice me How to increase the load regulation ?

Hi,

Firstly, I would like to confirm with you:

Did you measure this voltage right on the PCB board or at load side with long cable connected? I just want to confirm this voltage drop is not due to the cable drop voltage.

Thanks

Thanks Terry,

You are right. I measured the voltage on PCB board, it was about 11.9V when loading 5A, but at the load side it was only 10.8 V. I have to change the output cable with the bigger size and now it is working perfectly.

Now I want to design a converter for output 12VDC-10A. But how to wind the secondary with diameter so big as ~ 1.8mm.

If using 4 parallel wires, it must be ~1mm / each wire, it will be too hard to wind.

In my market, it is difficult to find the foil. May I separate to many 6, or 10 wires with smaller diameter ?

Could the leakage inductance increase or not ??

Hi,

If you can not find the foil, you can use liz wire which is bunch of small wires. If you  can design the transformer with good coupling, spread the winding properly, you could get a good disign with small leakage inductance.

Thank Terry,

I have finished building the transformer with Litz-wire as attached picture. I used 2 litz-wire like that, each wire has 7x0.6mm so it forms 14x0.6mm.

The leakage inductance looks so good, only 6 uH.

Now I am assembling the board and whenever finish, I will tell you the result.

Once again, thanks.

Hi,

I design another converter with 3 output as follows :

+12V : 10A (Main output)

-12V :  5A

+ 24V  :  2A

The feedback circuit and transformer construction as in the picture. The minimum load is the cooling fan connect to +12V.

The problem is : 

**   No load : +12V, -12V, +25V

** Loading +12V  =  1Amp :  +12V, -14V (no load), +28V (no load)

** Loading +12V  =  2Amp :  +12V, -17V (no load), +31V (no load)

I think all output winding are well coupling as in the picture,but why the output voltage vary too much ?

Hi,

This is actucally cross regulation which is very common in multi output fly back power supply. there are a lot of facts that can impact cross regulation performance.

Are the out put -12V and 25V always at no load condition at the real application? Are the no load output voltage (-17V or 31V out of your design spec?)

If you do not care much about the no load consumption, there is a very simple way to solve it: Put some dummy load (1kohm or higher value resistors on the outputs.)

Hope this will help you.