vIPER22a MULTI OUTPUT

[SATENDER]

i am using viper 22A. having three outputs. one output is regulated and other two are non regulated.

I am having below problems

1. i keep fan load of 120mA, output is regulated but as soon as i connect other load , all voltages dips including regulated ones.

2. if i add more load on regulated winding, my voltage drops on regulated ones.

3. if i increase grid voltage beyond 210v, ic enters into some kind of protection. it turns on and off.

i have attached the transformer  and circuit diagram.

It will be very helpful if anybody could help

  

[Peter BENSCH]

Welcome @SATENDER, to the community!

From the symptoms it looks as if the supply is simply leaving its regulation range and the VIPer22A is going into protection / restart.

Key things to check:

1. Transformer design: If the primary inductance or peak current are wrong, the converter will saturate or hit current limit when you add load or raise the mains

2. Cross‑regulation of the three outputs: Only one secondary is in the TL431 feedback loop. When you load the other two windings, the energy is redistributed and the regulated output will also drop. For three independently loaded outputs you normally need post‑regulators or a different topology

3. VDD / auxiliary winding Check VDD at high mains with an oscilloscope. If it goes above the OVP clamp or below UVLO, the VIPer will start and stop periodically (what you see above 210V)

Practically, the easiest way forward is to take an ST reference design for VIPer22A, copy the transformer and feedback network exactly, verify it works, and only then add your extra outputs with very tightly coupled windings.

Hope that helps?

Regards
/Peter

[SATENDER]

Hello peter ! thanks for your prompt reply.

There was mistake from transformer winding side. I changed the winding
polarity, it worked and everything is working fine now.



Why we are not implementing type II compensation network. Please support
with some reference design.

[Peter BENSCH]

At the output power you specified for the flyback, it operates in DCM at 230V and on the boundary between DCM and CCM at 115V, so a type-II network should indeed improve stability. You can start with 22k+10nF (fz~3.4kHz) in parallel to the 10k between Vo and Vref; if necessary, 4.7k+10nF (fz~725Hz) might also be sufficient.

Regards
/Peter

[SATENDER]

Greetings peter!

Do you mean I should connect 22k+10nF (fz~3.4kHz) in parallel with C8
capacitor.





Also , I have few queries,

1. What if I use 50V capacitor in RCD snubber in place of 2kv as shown
in figure.

2. What if I vary the R value of RCD snubeer.

[Peter BENSCH]

Since you have not assigned a component number or designation to a single component in your schematics, I do not know what you mean by C8. Therefore, I believe I clearly described where you should connect the RC series circuit in parallel: between Vo and Vref.

Regarding your further questions:

1. It would be extremely wise to use a capacitor with sufficiently high voltage rating for the snubber. After all, it fully gets the turn-off voltage of the flyback transformer.
   By the way, you can simulate this with eDesignSuite. Specifically, in response to your question: the 50V capacitor in the snubber will immediately blow up or at least smoke after the first power-up.
2. You can also simulate changes to the resistor in the snubber with the aforementioned eDesignSuite.

One more note about the snubber diode: a 1N4007 is completely unsuitable for this purpose, as it is far too slow with a trr of 1000...1500ns. You should use an ultrafast diode with significantly less than 100ns instead.

Regards
/Peter

[SATENDER]

Actually, by mistake I used 50V ceramic capacitor in RCD circuit. This
capacitor is safe but I lost my 4 viper IC.

Same circuit except RCD snubber cap is working fine on zero pcb, obviously
with 1nF, 2kV ceramic one. So I assembled it on my pcb, I used by mistake
50V ceramic capacitor in RCD circuit. I damaged 4 viper IC. Source pin (1,2)
gets damaged every time.

It will be very helpful if you could suggest something. Diode in RCD is
UF4007 in both condition .

[Peter BENSCH]

The hand-drawn schematic still shows the 1N4007, but a UF4007 should be sufficiently fast.

[SATENDER]

But any explanation of <>
VIPER22ASTR-E getting damaged, is it due to 50V ceramic cap. Two circuits
and only this difference (50C ceramic, 2kV ceramic).

Also, I could not find design tool for
<> VIPER22ASTR-E.

[Peter BENSCH]

If a 50V ceramic capacitor is exposed to a high-voltage pulse, what happens to it depends on the manufacturing technology: it either forms a short circuit or completely loses its capacitance. In both cases, the VIPER22A is no longer protected and will also emit smoke signals.

Regarding simulation: indeed, the VIPER22A is actually not yet represented, since his birthday was already a very long time ago in 2006. The VIPER222LSTR (60kHz) is the most suitable substitute to use for a simulation of the snubber in the eDesignSuite. The snubber values obtained with VIPER222L will be a very useful starting point, and can then be refined by oscilloscope measurements on your real VIPER22A board.