Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Input Bridge rectifier Heating up in a Flyback converter

Status
Not open for further replies.
there is no phasing shown on the schematic - dot notation refers to the fact that when the fet turns off - the winding end it is connected to goes +ve, w.r.t. the other end - the output winding should now energise +ve such that the +ve end now drives thru the o/p diode - similarly for the aux supply - if the wgds are not conected the right way round - things will never work ....
--- Updated ---

you need a 16V 3W zener on the Vcc - to protect the chip from the 100k R connected to rectified mains ....
 

    S

    sabu31

    Points: 2
    Helpful Answer Positive Rating
Yes and also i hope your R10 "bleed startup resistor" is rated for the mains peak voltage at least
 

    S

    sabu31

    Points: 2
    Helpful Answer Positive Rating
Dear All,

I have replaced the IC UC3842 and also rewinded the Transformer. The turns ratio is now (60:9:8(aux)). This time I am keeping Vcc 18V. The magnetizing inductance is 850uH. However, the IC is not generating any gating pulses. What could be reason for IC to not generate any gating voltage. As this board has experience few burnt outs (gate resistor, diode bridge ect), may be traces/pth may be damaged. I will try soldering items on a spare PCB and check again.
 

sabu31 said:
Dear All,

I have replaced the IC UC3842 and also rewinded the Transformer. The turns ratio is now (60:9:8(aux)). This time I am keeping Vcc 18V. The magnetizing inductance is 850uH. However, the IC is not generating any gating pulses. What could be reason for IC to not generate any gating voltage. As this board has experience few burnt outs (gate resistor, diode bridge ect), may be traces/pth may be damaged. I will try soldering items on a spare PCB and check again.
Click to expand...
What value is the voltage measured at the COMP pin?
 

I'd say you've blown up a few parts & tracks - time for a clean build and correct phasing on the Tx wdgs ...
 

Akanimo said:
What value is the voltage measured at the COMP pin?
Click to expand...
The waveform measured in the COMP is attached. The comp pin waveform has a square wave shape. The DC input is around 140V.
When I am giving separate supply to Vcc (18V, from external PSU). One end of the differential probe is connected to Source Pin of Mosfet and if the other end is made to touch PIN 2 of UC3842, the system runs briefly for a second or so. Again If I repeat for a few times, it works then stops all together.
 

Attachments

  • scope_1.png
    scope_1.png
    27.4 KB · Views: 86

Easy peasy said:
perhaps we need to see your actual build ... ?
Click to expand...
I have attached the gerbers.
Please note that input rectifiers are now 1n5408
aux diode and snubber diodes are UF4007
Have removed the input CM choke, NTC and Capacitor.
Input I am giving through a variac
 

Attachments

  • IsolatedSupply.ZIP
    43.3 KB · Views: 60

Hi,

I don't have a Gerber viewer on my tablet..

Schematic, PCB --> PNG, PDF....

Klaus
 

KlausST said:
Hi,

I don't have a Gerber viewer on my tablet..

Schematic, PCB --> PNG, PDF....

Klaus
Click to expand...
Dear Klaus

I have uploaded schematic and layout. The input choke/ntc is not populated. The present issue is that the voltage is present across the dc link and have checked upto 300V DC, however the mosfet is not switching. The UC3842 has been tested with auxillary supply of 18V but it still does not switch.

When the Vcc is give through auxillary/ R10 resistor. The voltage is rising to 25V but it still does not switch. The IC has been replaced twice but no avail.

The transformer winding is having turns ratio of (60:9(output):8(aux)). The polarity has been checked. Similar circuit configuration is working in PSIM simulation.
 

Attachments

  • Layout.pdf
    24.1 KB · Views: 74
  • Layout2.pdf
    35.3 KB · Views: 69
  • Schematic.pdf
    46.8 KB · Views: 93

the way you have hooked the opto to the feedback on the IC is a little unusual - have you ever had this working - ever - before ...?
 

Easy peasy said:
the way you have hooked the opto to the feedback on the IC is a little unusual - have you ever had this working - ever - before ...?
Click to expand...
May I know whats the unusual part. I have used TI appnote (Pg 22 Schematic) for calcualtions (did not use slope compensation) .
 

Attachments

  • USED DESIGN_uc2844.pdf
    3.9 MB · Views: 69

Easy peasy said:
the way you have hooked the opto to the feedback on the IC is a little unusual - have you ever had this working - ever - before ...?
Click to expand...
It does work.

That's the CC configuration. I use it for DCM, with Ri and RF (his R15 and R16, respectively) having equal values of 47k to 100k so that the IC error amp contributes no gain (ie unity gain) but phase inversion. With this I set the gain using RLED, CTR and Re.
--- Updated ---

sabu31 said:
May I know whats the unusual part. I have used TI appnote (Pg 22 Schematic) for calcualtions (did not use slope compensation) .
Click to expand...
You do not need slope compensation. Your duty cycle is less that 50%. It's around 35%.
--- Updated ---

If everything is working fine, the voltage at the IC's COMP pin should be around 3.1V
--- Updated ---

sabu31 said:
The waveform measured in the COMP is attached. The comp pin waveform has a square wave shape. The DC input is around 140V.
When I am giving separate supply to Vcc (18V, from external PSU). One end of the differential probe is connected to Source Pin of Mosfet and if the other end is made to touch PIN 2 of UC3842, the system runs briefly for a second or so. Again If I repeat for a few times, it works then stops all together.
Click to expand...

The voltage at the COMP pin should not be a square wave. Seems like it's is starting up and then shutting down intermittently. I'd advice you use UC3844 and also make provision for soft-starting.
 
Last edited:

Akanimo said:
It does work.

That's the CC configuration. I use it for DCM, with Ri and RF (his R15 and R16, respectively) having equal values of 47k to 100k so that the IC error amp contributes no gain (ie unity gain) but phase inversion. With this I set the gain using RLED, CTR and Re.
--- Updated ---


You do not need slope compensation. Your duty cycle is less that 50%. It's around 35%.
--- Updated ---

If everything is working fine, the voltage at the IC's COMP pin should be around 3.1V
--- Updated ---



The voltage at the COMP pin should not be a square wave. Seems like it's is starting up and then shutting down intermittently. I'd advice you use UC3844 and also make provision for soft-starting.
Click to expand...
Thanks for the reply.

Since, I was not getting any breakthrough. I thought of testing IC based on the circuit mentioned in TI appnote using 2N2222 transistor. I am not able to getting PWM output pulses from IC. This i need to check the various potentionmeter values. My main question is

1. The ICs which were being used in circuit that did not work, when I am giving Vcc =18V, vref is nearly 0. However, it should be 5V. This means IC should be damaged?
2. One IC which I did not use, also is not showing Vref=5v when Vcc=18V. Another IC is showing Vref=5V, 18V, and also saw tooth waveform at RT CT pin. Though I am not able to get PWM output.
3. This is bit of digression , I tried simulation in LTSPICE and PSIM. But PSIM model is not working, LTSPICE is working. WHat could be the issue. I am attaching the spice and PSIM models.
4. I am also thinking of testing the circuit based on Dr. Ridleys article of replacing transformer with a load. But I am not sure what should be put at the FB, COMP side. I am attaching the article for your suggestion. Thanks
 

Attachments

  • Ridleyarticle_FlybackConverterTesting.pdf
    1.1 MB · Views: 65
  • PWMIC_TESTCIRCUIT.zip
    8.7 KB · Views: 55

Yes , If you put vcc = 18V into an otherwise unconnected uc3842, you should see 5v on vref and rt/ct pin should show sawtooth. If you ground the FB pin then you should see an output on the gate pin.

The attached is your flyback in basic form in LTspice.

1….Set your circuit up as in your last schem.
2…remove the series gate resistor (r13); Remove the startup resistor (r10)
3….Apply a 18V bias supply to the C6 (aux rail cap)
4…connect uc3842 FB pin to ground

Do you now see the gate pin going high?

If not, the UC3842 is blown or you have a short/open somewhere in your cct
 

Attachments

  • Flyback 20v 3a.TXT
    8.3 KB · Views: 54

    S

    sabu31

    Points: 2
    Helpful Answer Positive Rating
you may have blown up a lot of stuff, also the original pcb layout might have an error or two to find - and may now have blown tracks ...(?)
 

    S

    sabu31

    Points: 2
    Helpful Answer Positive Rating
Yes, can you take a pic of your PCB and send?
It sounds like you have been in the wars.......with High voltage possibly having seeped through the blown fet and trashed your control circuitry.

In some ways, after the raft of faults youve reported, i need to be at your side, looking at the pcb and probing it myself, to see what has gone on.....we've all maybe been there, a PCB that has been in a "war zone"....maybe you should start over, with an 18v bias supply at first...no bias winding and no bleed startup resistor.....otherwise, i think you will have to pick one of us, and get us a plane ticket to your nearest airport.
 
  • Haha
Reactions: sabu31

    S

    sabu31

    Points: 2
    Helpful Answer Positive Rating
Dear Thanks for help.
Yes I think have to restart over again. Only reluctance to to the pcb again is that unless I find out what is the issue in earlier PCB, it may get repeated in 2nd version. Also where all do I have to put test points for probing in PCB.
 

Hi,

The whole design neither follows the design recommendations given in the datasheet, nor safety regulations.
So it's no wonder it does not work, but more important: you may hurt/kill yourself or others.

And then you have to keep on EMI/EMC regulations. I see no way with your current layout to satisfy them.

Thus my recommendation: buy a cheap laptop power supply.

Klaus
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top