Pulse Grouping

27 posts / 0 new

Hi everyone

I am designing a flyback supply following your steps in RDR-201....but i want 6W and 16V in the output with the lnk625 IC. I have followed the PIXIs program for choosing the components. The problem is that i have something similar to Pulse Grouping showed in Figure 11 of datasheet of lnk625, but is enormous, i can see pulse during 100ms and each pulse 20us. The result.......without charge i have 8V instead of 16V with ripple (it is like a sawtooth) and with charge i can see like in Figure 11 of datasheet of lnk625 where is all the time up and down like a hill.

I have changed the values of output capacitor to reduce ripple, i have changed the value of resistor divisor according to the PIXIs program to get 16V, i have put the RC net to reduce pulse grouping like you indicate in the datasheet........but i don't get supply works properly. The trafo works properly.

Any idea that what I can do

Thanks in advanced

AttachmentSize
schematic_flyback_16V_0.pdf103.39 KB

Hello rubenadd,

From http://www.powerint.com/sites/default/files/product-docs/linkcv_family_datasheet.pdf, page 9 the main change to deal with this problem is to use a zenner clamp, see page 4. The zenner is capable to eliminate ringing faster when the MOSFET turns OFF, and the critical condition is to have the voltage stable (with no ringing), see page 8. You can look at drain voltage to make sure you have no ringing after 2us from OFF command. In your case, you have to replace the discharge resistor with R1 and VR1 (from page 4).
The voltage on FB pin is sampled 2.1us after OFF command and this is the one used for regulation. That’s the reason you must make sure there is no ringing 2us later after OFF command. Most of the time you have to use this zenner solution, especially if the output voltage it is higher than 5V.

Cheers,
PI_Crusher

I have tried to find out how i can design a clamp net to choose correctly the values of component, but without success. Could you give some reference about this please? About what values and kind of features i have to be care.

Cheers

Hello rubenadd,

Please read the answers to your question:
From http://www.powerint.com/sites/default/files/product-docs/linkcv_family_datasheet.pdf, page 9 the main change to deal with this problem is to use a zenner clamp, see page 4.

Replace your D5, R2, R1, and C3 with D5, R2, VR1, R1, C3 from the example.
If you have trouble doing this, read throughout documents posted on our webpage, or contact our design consultants here:
http://www.powerint.com/en/community/design-consultants

Cheers,
PI_Crusher

Hello PI team

I have done the changes which you proposed me......and i have put a clamp zener with the same values components which are in AN, but the result it is the same. Certainly the transitions are softer than before, but i don't get nothing with which work. Could it have another posibility??

Cheers

rubenadd

Hi rubenadd,
Maybe it is a silly question to ask: have you redesigned the transformer for 16V/0.375A? The original RDR-201 is for 5V/1.2A and all I could get is that you used our program for choosing the components. To understand what’s wrong with your design I need to see your project files (PIXls Designer or PI Expert 7).
Cheers,
PI_Crusher

Hi Crusher

I have included the PIXIs document. On the issue of transformer, i asked for them in Power Integration too, according to the PIXIs design. I hope this can be useful for you to help me.

Thanks in advanced again for your time

Cheers

rubenadd

Hi rubenadd,

Please measure your physical transformer and tell me what Primary Inductance and Primary Leakage Inductance you have. I need to compare the transformer you use with the design parameters. Something is really out of nominal range; auto-restart like behavior is mostly related to transformer parameters, providing the nominal power is not exceeded.
Also, please measure the input voltage and make sure the input voltage ripple is not excessively high. You must have at all times more than 100Vdc input (after AC rectifier). If needed, increase the electrolytic capacitor value.
Let me know if you do not want to post your schematic here and I will send you a private message with an email address.
Cheers,
PI_Crusher

Hi PI_Crusher

The primary inductance is 1.6 mH and the leakeage inductance is 40 uH. The input voltage is around 350V dc. I have included the schematic where VR1 is a Zener (in the schematic it appears like a normal diode and R8b is for adjust values but actually it is a short circuit) and the BOM. Moreover around R9 i can put optionally and RC net with 47K and 680 pF for pulse grouping as indicate in datasheet. I have 6V in Bypass Pin.

Thank you for your time

rubenadd

Hello rubenadd,
The transformer measurements are normal.
Since it is difficult to read components designation and values from your schematic picture, let’s discussed based on RDR-210: http://www.powerint.com/sites/default/files/PDFFiles/rdr201.pdf
You have to identify the equivalent components in your schematic.
Make sure the BP capacitor C4 it is physically very close to BP and S pins on PCB. Try first to get the circuit running without extra current into BP pin. Disconnect R6 and R5 from RDR-210 (identify the equivalent resistors in your schematic); the circuit should be capable to get the necessary current from D pin. If the output is stable now and you need the bias winding for improved efficiency, change the “NO” option to “YES” for Bias Winding on the design file and double check the windings with this new option. Or bring back R6 and R5 and try to change the value for these resistors to remove fast oscillations on FB pin.
If the output is not stable without R6 and R5, the problem you have to test now is the signal on FB pin 2.5us after the internal MOSFET turns OFF. Check http://www.powerint.com/sites/default/files/product-docs/linkcv_family_datasheet.pdf on page 4 and try to add R5 and C5 from FB to S. Test with various values for R5 from 10k to 200k and see if you can clean the parasitic oscillations.
Make sure you follow the PCB guidelines. Can you post a picture with you PCB (placement and bottom side)?
Cheers,
PI_Crusher

Hi PI_Crusher

I have done what you have told me, i have removed r5 and r6 (both of them belong to feedback winding) and i have check the output, it's perfectly stable from 2us from the turn off. So i will follow the steps which you told me. But i don't know how i must put the bias winding. Could you send me a scheme of how i must connect the elements? Can i use the same transformer? The PIXls designer tell me that the number of winding is 1.

On the subject of guidelines i think i have follow all the advises. I include the footprint where it is not indicate the zener diode because i solded "in the air" but it is present and R3b is only for adjust the value exactly but in that moment is a short circuit.

Cheers

rubenadd

Hello rubenadd,

The output at low and high load it is stable without R5 and R6? Only if you have a perfectly stable output makes sense to bring back R5 and R6 (change the value for these two resistors, try with bigger R6 and maybe smaller R5) or to change the bias winding. Adding one bias winding is another way of getting bias current for BP pin. Adding one turn to the bias winding is illustrated in http://www.powerint.com/sites/default/files/product-docs/an45.pdf page1.
If you still have problems, show me the PCB you have.

Cheers,
PI_Crusher

Hi PI_Crusher

The output is stable but very low, i have changed the resistors in your way, but the behavioural was the same. I sent you a pdf with the pc in the last post. Maybe if you have anyt time, you could browse it.

Thank you for your attention

Cheers

rubenadd

Hello rubenadd,

For now just forget about R6 and R5. There it is nothing really abnormal with your PCB. Since the output is very low and it is nothing wrong with the transformer, all you have to do now is to adjust R3 incrementally up. After you get the desired output voltage you can go back to experience with bias support (add back R5 and R6, increase R5 as necessary to maintain good regulation under all conditions).
There are some RDKs available to buy here (Look for LinkSwitch-CV applications):
http://www.powerint.com/en/design-support/reference-designs/reference-design-kits

Cheers,
PI_Crusher

Hi PI_Crusher

I will try to explain you the situation: With r5 and r6 remove (resistors in bias winding) and i have put a short circuit in theirs places. From this point and without charge, i can see autorestart in the IC when r3 (r_up in the feedback) is under 100K because i see the hills in the output signal and the autorestart in the IC signal. And when i put r3 over 100K i can see a stable output signal but with 1.7V and i design the power supply to 16V, but it is impossible with any combination of resistor change this value of voltage.
On the other hand when i put the charge, independently of the value of r3, i can see the hills in the ouput signal and the autorestart in the IC. So it is clear it doesn't work properly even without charge, so it is clear i can't see a incorrect parameter in the design, because unless the IC should work properly. Does this information give you any idea?

On the subject of purchase, i try to find a power supply with 6W and 16V and in RDKs there is one of 6W/5V, in fact i based my design in these power supply but without succes. Could it be possible the PI team design to me one with my requeriments. And if it could be possible this, how much would it be?

Cheers

rubenadd

Removed means to take them out. Short-circuit is the oposit. You were not supposed to place something back, just to remove.

Yes but if i did this, i could see like the signal with "hills" but square signal instead of hills signal

Hello rubenadd,

Your best bet is to contact design consultants here:
http://www.powerint.com/en/community/design-consultants
Cheers,
PI_crusher

Hello PI_Crusher

I have been searching again about the topic, and i have been simulating in PSPICE with another IC, and i have noticed myself when the initial conditions are 0V in the output capacitors, the IC goes into auto-restart, but when the initial condition are 16V, the IC works correctly. I can imagine this is the same problem with lnk625 but i don't find out the reason. Any idea?

Cheers

rubenadd

What's about the possibility of the design the supply by your side? Or another possiblity maybe you could keep in touch with some client who uses this integrate for ask him how he could make it works properly

Could it be possible buy a power supply with this IC?

Sorry i am confuse. When i removed the resistors and i put a short circuit in them, in the output of 16V certainly i don't get a proper signal (nearly 700mV) but the drain signal is stable. But there is a detail, when r5 and r6 are in the pcb, the drain signal is stable 2us after the turn off too.

What kind of parameters should i change if the primary and leakeage inductande are higher??

Sorry, files are here.

Cheers

rubenadd

In my opinion, it is like Auto-Restart issue beacuse i can see switchings during 200-250ms and then it is without transitions during 1 second aprox. instead of 2.5 seconds like Aplication Note indicates. if the problem is this, what would be the solution?

I forget telling it that in BP pin i have 6 V constants and in the FB pin i have 0.7-0.8V in 2.5 us after the MosFET turn-off.

I mean........The value of the Zener would have to be such it can be resist the Vor in my case 90V, and one resistance that support the current in the Zener, aren't they?