RDR242 TOP266VG ac-dc flyback circuit - can't get it to work properly with a planar transformer

5 posts / 0 new

I tried to use this RDR242 design which uses TOP266VG offline switcher.

Right now my issue is with instability of the circuit. There's ~12VDC at the output, but it drops when I connect a load, and doesn't recover back to 12VDC fast enough when I remove load.
Also, not enough voltage at C7, which is near aux/bias winding, it's around 5.9V, instead of 9V.

This is the original RDR242 schematic. This is my schematic in Altium Designer. I changed a few components as not everything that's on the original old schematic is available nowadays. For instance, Shunt regulator U3, T1, C3 input capacitor etc.
For U3, shunt regulator ATL431 I used such values of resistor dividers R21, R23 as to yield 2.54V at its reference pin as this is the minimum ref voltage that it will work on, compared to LMV431 shunt regulator on the original RDR242 version, which was set to 1.24V at its ref. pin.
I put TOP266VG into 66Khz switching frequency between its drain and source pins, by connecting F pin to C pin, as suggested on p.3 in datasheet.

For T1 I used an ER25 transformer with the following parameters. Originally it was intended for Viper22A circuit, and it worked fine at 60Khz MOSFET switching frequency at primary winding. There was 12V at the output, and about 24V at auxiliary winding.

Yet when I try to use this 25W ER25 transformer in RDR242 circuit, I get voltage that's about 11.7-12VDC at the output (X1, X2), and drops when I connect a light load. C7 near aux winding shows about 5.9V at no load. I also noticed that the frequency at TOP266VG's drain and source pins doesn't go above ~2Khz (I monitor those pins continuously with an oscilloscope).

Here's my PCB design. I connected transformer to PCB with wires. As you can see from ER25 parameters, its aux winding would yield about 24V when its secondary is at 12V (since aux has twice the turns that secondary) and it appears that original rdr242 design was intended for aux/bias winding that would yield 14V, as VR3 zener diode would start reverse conducting at above 15V and the V pin of TOP266VG would receive too much current, prompting shutdown. I tried to use an LDO that would output 12VDC, but the circuit didn't work. However, I read that TOP266VG uses power from the opto-coupler's feedback path, and as long as C7 is above 9V, the U2 opto-coupler would be properly biased. Yet for some reason C7 doesn't go above 5.9V at no load. I also noticed that the 15V 200mA diode D7 would fry up quickly, so it appears that too much current is drawn from the aux winding, I tried lowering its theoretical 24V with an LDO or resistor divider network which would have also limited current, but ended up simply putting a resistor in series with D7.
Through experiments I realized that an SMD 0805 1/8W resistor with a value of less than ~30Ohm, can't handle whatever current is passing. So I tried replacing it with like 120Ohm to limit current and D7 was able to handle whatever current was passing (<200mA for sure), and the circuit kinda worked fine, except the voltage at the output would drop when I try to connect a load, such as Nanopc T4 single-board computer.

Why isn't this circuit working? Feels like I messed up transformer polarities, but I checked it and it seems fine. Perhaps I should fix the opto-coupler feedback network? That could be the reason my circuit doesn't work properly/has stability issues.

Over here a guy used TOP259EN controller with 12.2VDC, 9.4A output. If you look at his his circuit diagram you'll see that he used different resistors for biasing opto-coupler feedback network. There's such thing as a gain for opto-couplers, perhaps I should lower it for my circuit? I don't want to mess up feedback network though, so is there a tool that would allow me to calculate proper values of resistors and where to put them at the feedback network? I don't think just copying his opto-coupler resistor feedback part of the circuit is a good idea as my current limit is not 9.40A, right?

Anyone here?

The manufacturer's application note an47 doesn't go into detail about proper formulas unfortunately, on p. 17 they just lightly touch upon the subject, and use fig. 15 as an example.

So the general principle is that at 66Khz frequency mode, TOP266 works in CV mode (varying current changes PWM of controller) reduces duty cycle to a mininmum/shuts off when receiving about 4.6mA into control pin.
I_B = 1.5mA, I_C(OFF) = 3.1 mA (refer to p.19 and p. 24-25 in datasheet)

ATL431 that I used for U3, starts "sinking" more current at its cathode pin when the output voltage increases, thus increasing its ref pin voltage above 2.5V, thus more current will flow through opto-coupler U2A, I_f, causing more current at bias winding/control pin node, which means that duty cycle of TOPSwitch controller will decrease, correct?

Also, is the reason there's only 5.9V (or 5.8V) on C7, is because TOP266 is in standby mode? Or is this how much voltage should be there during normal operation? Otherwise, the aux/bias winding of my planar transformer should have double voltage of secondary winding, so 24V.
I'm guessing TOP266 internal shunt makes sure that there's a maximum of 5.9V at C7??

R19 calculation:

12V (positive node of output rail X1) - 1.2V (V_f of U2A see p. 10 in datasheet) - 2.5V (drop between U3's cathode of shunt regulator and its anode/GND) = 8.3V

For a 100% CTR in opto-led, R19 = 8.3V/4.6mA = ~1.8K

Is that correct?

I also read about biasing opto-coupler with TL431 using Chris Basso's guide (start from p. 135)

He recommends for opto to have a gain of more than 7, or more than log_10(7) * 20 = ~17db
Well, in that case:

My R_led, which R19, was left at 470.
And through experiment, I determined that R_aux (the resistor I put in front of D7) value of 197ohm resistance limits enough current to ensure there's about 12VDC at the output (X1, X2).

However, that voltage drops immediately if I connect a load that attempts to draw about 500mA (NanoPC T4 single board computer), and doesn't go back up, feedback loop doesn't seem to bring it back up.

Could the issue be that I connected planar transformer with short wires to the PCB? (see above my PCB design in Altium Designer).

Should I worry about zeroes and poles compensation in the feedback loop as was mentioned in Basso's guide? Or is it somewhat taken care of by capacitors C10 and C20?

I tried lowering gain of opto, and chose R19 = 820 Ohm
Then my R_aux had to be about 65 Ohms in order for there to be 12 at the output.
If I put higher blocking R_aux resistor at the bias winding, then the output voltage is too high
if I put smaller value of R_aux, output is too low.
Yet once again, even if I as much as shake the PCB board, the output voltage drops to nearly 0 and doesn't go back up.
So I have to unplug power from mains, and then turn the circuit back and only then I can get 12VDC again at the output at no load.

Also, could you advise how does one limit current at the output? I'm guessing R19 resistor will play a role in it, right?
For instance, the RDR242 circuit was designed to limit current to 2.5A @ 12VDC, how do I modify it to 1.5A @12VDC? Can you provide a formula/explanation?


What is the input voltage range for your design? Is it universal input or high line only. The reference design uses 132kHz as nominal switching frequency. In your design you are using only 66 kHz. The output capacitance also very low. Please re-visit your transformer design using Piexpert online tool. The tool still doesn't have ER25 core but you can add core details manually or consider the core area equal to ER25 . Your bias voltage seems very high, please follow design parameters form PIexpert tool.

Hi, 220VAC euro mains? What about my questions? Can someone help with them here? They'd help understand how this circuit works/what the issue is.

I don't think output capacitance is the issue here...
RDR242 uses total 1460uF at the output (C14, C15, C16)
In my test I have 1100uF total at the output.

As for bias voltage, I read here on this forum that TOPSwitch can work between 6V - 30V supply.
So what would you say the maximum bias winding voltage ought to be for TOPSwitch JX controller?
And if someone could answer why there's 5.9V on C7 and not ~24V when there's 12V at the output at no load that'd be great too.

You are correct. I overlooked the C15 value. The output capacitance is correct. The absolute control pin (C) maximum voltage is 9V.

- " I put TOP266VG into 66Khz switching frequency between its drain and source pins, by connecting F pin to C pin". Your schematic shows that F pin connected to source. That forces IC to 132 kHz operation instead of 66 kHz operation.

I recommend to follow below steps to diagnosis the issue.

- Use signal generator and apply sine voltage/100kHz frequency across the primary of transformer on bear pcb board to verify the transformer turns ratio.

- Check polarity of the capacitor (C7) connection on the board.
- Can you share drain waveform during startup and just after reaching regulation - check frequency during startup and after reaching regulation.