Repairing a flyback - regulation fails under load

zvir

New member
Good evening, friends.

As a non-professional with not much to contribute, I've spent many hours only reading the forums and learning new and interesting things. Now, the time has come when I am facing a wall and I need advice from people more experienced than myself. Please forgive me if a problem I am facing is not challenging. :)

I am trying to repair a small (50ish W) SMPS. I don't believe it matters, but the thing is used to power a set of speakers (that's why I'm posting in Amplifiers, if the moderators have other suggestions, please move the thread to where it would be more apropriate).

When I first disassembled the power supply, I found that the MOSFET and the current shunt resistor were blown. I replaced those, but to no avail - I could not see anything happening on the primary side. After a bit of poking around with a voltmeter, I came to a realization that the PWM controller is also dead. It's a tiny SOT23-6 package, and it had a couple of letters on it, but Googling gave no result, so I went on to reverse engineer the important parts of the circuit. I came to a conclusion that the supply is of flyback topology, and by pinout and package (and considering local availability), I decided to replace the IC with an NCP1521. The partial schematic I reverse engineered seems almost identical to the suggested schematic in the NCP1251 datasheet.

There was also a small two-transistor circuit that was used (if I am interpreting it correctly) to block the gate drive before Vcc reaches a certain level. Both of those transistors were giving me a dead short, so I just removed them and connected the gate drive pin of the controller to the MOSFET gate.

Success! (Sort of.) I now have correct 27V on the output, the LED comes on (it draws power from the secondary), and it's stable while not loaded. However, if I load the output with 27R (to get 1A out, which should be perfectly fine, since the supply is supposed to be able to put out 1.5A), the output drops down to zero. If I load the output with 4k7, it drops down to 10-15V and begins doing some strange stuff. The controller seems to try to activate, but can't, then goes to some kind of reset state, and then recovers. I am measuring very large ripple on the Vcc pin of the IC, (something like 10VDC + 10Vpp, with no ripple in the unloaded state). The FB pin shows rectangular signals, with period correlating with Vcc ripple and output instability. No respectable current can be drawn from the secondary.

I am aware that the IC I decided to solder in is probably not the same one as the one that originally was there. This was the closest match I could find. I assume symptoms like these might be caused by a faulty regulation loop, but also by the auxiliary winding not being able to provide enough power (voltage?) to take over after the initial pulse. However, I am not brave enough to try and calculate new dividers to power the IC, I would first like to hear ideas from people who might have seen this behavior before.

If any more information is needed, I can provide measurements, photos, partial schematics, o'scope traces, whatever.

Thank you in advance.
Z.

P.S. I know it might be much easier (and perhaps cheaper) to just buy a new power brick and be done with it, but I'm a cheapskate. :D And I also like tinkering with this stuff, I consider it brain exercise. We all learn from new experiences, and seeing behavior like this is a new one for me.
 

MicrosiM

Administrator
Staff member
@Zvir

Please share some photos of your board, I am not sure if the part you have chosen is the correct one.

Do you have the model number of that power supply?

Or IC number?
 

zvir

New member
Hi MicrosiM,

I am uploading three photos - front and back of the PCB, as well as a photo of the cover, where PSU model is clearly visible.

Unfortunately, I threw away the burned out IC, and I seem to have misplaced the bit of paper on which I noted the IC markings. :(

As I mentioned before, I am almost certain that I did not choose the correct IC, but this was pretty much the only thing with the correct pinout locally available. However, there seems to be a fair amount of flexibility in the design of these power supplies. I remember fixing a similar one by just replacing the (unmarked) IC with the same one I'm trying now. That's why I was hoping the problem is not the choice of IC itself, but I just can't find anything else...

Thank you!
 

Attachments

  • IMG_20180714_130247.jpg
    IMG_20180714_130247.jpg
    1.5 MB · Views: 18
  • IMG_20180714_130302.jpg
    IMG_20180714_130302.jpg
    1.8 MB · Views: 15
  • IMG_20180714_130320.jpg
    IMG_20180714_130320.jpg
    1.5 MB · Views: 9

Silvio

Well-known member
Show us the schematic that you reversed engineer at least we could see what are you talking about. The two transistors removed could be an amplifier to drive the gate of the switching transistor or some kind of short circuit protection or for feedback regulation and maybe another thing that I did not mention.

Having a lot of ripple on the vcc pin is not good as this should be smooth as possible, the feedback winding should supply enough current for this purpose. Check for a faulty diode supplying this line and also smoothing capacitor. from what I can think it seems that the feedback loop is not acting right. Check also if there is a TL431 in the feedback this is a reference chip and if so see how its behaving under load conditions. Lastly check also the input capacitor and see that it is still in tolerance of the stated capacitance.

Good luck Silvio
 
Last edited:

kanon

Member
You should check, the small 6 pins Ic

tend to give many oscillation faults.

check if you have 12 volts in any pin, check if on the gate pin of the mosfet, see if there is voltage variation, If not, there is your ic is damaged

sample schematic.NCP1251-Pin Description.jpgadaptacion-ic-serie-ka5q-fuente-tv-lcd-2.gif
 
Top