SMPS using UC3842 - And related problems

[Checksum8]

I think R2 and D5 may be part of your problem. It is only allowing about 15ma to to power UC3842 and drive the mosfet. Replace R2 and D5 with  a LM7815. Do what Faros suggested and that will verify if this is your problem.

[Faros]

the UC3842 is not powered through the R2 / D5 branch, it is powered mainly by the auxiliary winding on the transformer ...

[vantusaonho]

the UC3842 is not powered through the R2 / D5 branch, it is powered mainly by the auxiliary winding on the transformer ...

UC3842 is powered through the R2/D5 branch for start up because if UC3842 is not powered, auxiliary winding will not conduct.

This is a schematic for input 110-220VAC, output 12V


Uploaded with ImageShack.us

There are a problem, the voltage at R2/D7 branch is 18.5V when unplug UC3842 but is 10 -14V according sin wave when plug UC3842. Please explain this

[xenix]

This is a flyback circuit. There must be airgap in the transformer.

[vantusaonho]

This is a flyback circuit. There must be airgap in the transformer.

Yes, there is a air-gap about 0.25mm in flyback transformer but I don't know why happen the phenomenon like that

[Faros]

There are a problem, the voltage at R2/D7 branch is 18.5V when unplug UC3842 but is 10 -14V according sin wave when plug UC3842. Please explain this

When the chip is out D3 is reversed and the voltage at pin 7 socket is equal to zener voltage ~18V, when the chip is inserted, the R2/D7 branch provides >1mA (enough for UC3842 to start oscillating), once oscillated, it will drive the MOSFET gate through the 10 ohm resistor and hence consume more current, if the circuit couldn't oscillate correctly for any reason , the auxiliary will not be able to deliver the required voltage and the Vcc of the chip will drop... did you try using an external PS? 

[xenix]

Yes, there is a air-gap about 0.25mm in flyback transformer but I don't know why happen the phenomenon like that

Transformers do not store energy, they just transfer it.
Inductors store energy. In this manner, flayback transformer is more like an inductor rather than a transformer. They should store energy when the transsistor is turned on. When turned off, they should deliver the energy to secondary winding.

Magnetic energy stored in unit volume is E=B²/(2µ). Relative magnetic permeability of ferrite is in the range of 50-500. That means you cannot store energy in ferrite. Therefore you need airgap. Nearly all the energy is stored in the airgap. it can be formulated as
½×L×I²=AirGapVolume×½×B²/µ

[vantusaonho]

Bum!!!!! That is the sound have just happen.
I replace R2 with 27K to increase more current through pin 7 UC3842, the voltage not vary anymore but the frequency at pin 6 only 4Khz meanwhile it should be 100KHz. When I remove current of auxiliary winding from pin 7, the circuit have burned after 2 seconds.

[vantusaonho]

I noticed that Resistor snubber and Mosfet overheat. I feel that flyback transformer have not transfer all energy cause overheat on mosfet and snubber resistor

[vantusaonho]

I make a smps input 33VDC output 12V
I wound a smps in order as follow: primary winding 15 Turns wire 0.4mm, bias winding 11 Turns wire 0.4mm, secondary winding wire 0.6mm
All winding are wounded as clockwise


* With first scheme:
- Output voltage 18V, rectifier diode at secondary side overheat. bias voltage 60V



* With second scheme:
- Ouput voltage 10.5V, bias voltage 18.5V, mosfet and snubber resistor overheat


Please let me know which scheme right

By the way please confirm to me which schema for flyback transformer. I read many app note, I saw both schema so I don't know what is right


Uploaded with ImageShack.us

Thanks in Advance

[Faros]

I am afraid that I didn't understand that this BUM sound was related to which component?!! ...

[Faros]

you may find this site helpful ...

http://schmidt-walter.eit.h-da.de/smps_e/smps_e.html

[vantusaonho]

I just find out that the rectifier diode at secondary side is very important. In my schematic, I use FR207 (Fast recovery diode) which I thought it can be work fine in high frequency. But it seem not like that. The diode overheat due to not  conduct quickly and block quickly. I think it should be replaced by a schottky barrier rectifier diode as: SR340, MBR1060....
The mosfet also overheat due to diode. I think so.
What do you think? Please let me know. Thanks

[pickit2]

Bum sound = bum note. or maybe even a fart....

I hope he not going into full production anytime soon.
How many are you planing on building? 1 or 1 million?

Note: My place of work buy this Item for $1 in lots of 10.

[alexxx]

I just find out that the rectifier diode at secondary side is very important. In my schematic, I use FR207 (Fast recovery diode) which I thought it can be work fine in high frequency. But it seem not like that. The diode overheat due to not  conduct quickly and block quickly. I think it should be replaced by a schottky barrier rectifier diode as: SR340, MBR1060....

I agree that a schottky diode should be used for that purpose.

If your amperage is below 1A, then MBRS1100 could be an option. I used it several times as the diode after the secondary winding, without having any overheating problems up to 250mA.
http://www.alldatasheet.com/datasheet-pdf/pdf/3176/MOTOROLA/MBRS1100.html

For higher currents, you could also look into SS34. I have never used it for such purpose, but I think it could do the job if the output is 18V as mentioned.
http://www.alldatasheet.com/datasheet-pdf/pdf/34915/ZOWIE/SS34.html

[Gallymimu]

I just find out that the rectifier diode at secondary side is very important. In my schematic, I use FR207 (Fast recovery diode) which I thought it can be work fine in high frequency. But it seem not like that. The diode overheat due to not  conduct quickly and block quickly. I think it should be replaced by a schottky barrier rectifier diode as: SR340, MBR1060....
The mosfet also overheat due to diode. I think so.
What do you think? Please let me know. Thanks

I can't think of a time that you WOULDN'T use a schottky in an SMPS.  You really can't do better than a SCHOTTKY for fast switch times and minimal switching and conduction losses.

[vantusaonho]

I replaced FR207 rectifier diode with MBR1060. But there is still a problem. The ouput voltage is 18V meanwhile the steady voltage is 12V. Perhap I biased for TL431 not properly. I read some paper about bias for TL431 but I don't understand because they present so complicated. I need a simple equation to calculate Rup, Rlow, Cf, Rled and Rbias. it would be good if people could help me in this problem.
 Thank in advanced

[Faros]

trying to design an SMPS without a scope is like trying to put together a 1000 pieces puzzle with your eye blindfolded.
you will keep having signs that you will not be able to interpret simple because you don't see what is going on.
take your circuit to a friend who has a scope and then you will be able to judge for your self what a 42 reply so far couldn't .
 

 

[Ichan]

All winding are wounded as clockwise

I have a batch of EE16 transformer which is useless because the maker wound the the windings wrongly in the same direction, where it should be in opposite direction - the result is the similar with yours, hot everywhere and i do not know why. I know a little about transformer things, but from that experience maybe a worthy try for you to wound the secondary winding in CCW.

@Xenix, looks like you know transformer well - on EE core, where should the gap made: on center leg, on the two side leg, or on all three legs? What is the use of copper foil that sometimes found on hf transformer?

-ichan

[vantusaonho]

Yes! Mosfet and snubber resistor overheat due to wrong direction, rectifier diode overheat due to wrong diode. All fixed now, I replaced FR207 by 1N5822. It' ok fully. I made successfull a smps input 33V output 12V 1A
And I now am making a smsp with input 110V - 220VAC ouput 12V 1 A

[zab]

Mosfet and snubber resistor overheat due to wrong direction

Wrong direction of what ? Please make it clear. so that the whole problem may understandable.

[vantusaonho]

Wrong direction in windings flyback transformer

[vantusaonho]

Hi everybody!
I have succeed with a smps input 220VAC output12VDC. But I still have a problem. In my design, I must use Mosfet 25N120 (25A, 1200V), the board work fine. If I use 7N80 (7A, 800V) , 4N60 (4A, 600V), the mosfet and fuse will be burned. I wonder that why some smps use mosfet IRF840 (500V, 8A) work fine?
Please explain to me.

[Ichan]

First make sure that you use "original" mosfet, not a "fake" one - they are too many on the street market.

-ichan

[Gallymimu]

Hi everybody!
I have succeed with a smps input 220VAC output12VDC. But I still have a problem. In my design, I must use Mosfet 25N120 (25A, 1200V), the board work fine. If I use 7N80 (7A, 800V) , 4N60 (4A, 600V), the mosfet and fuse will be burned. I wonder that why some smps use mosfet IRF840 (500V, 8A) work fine?
Please explain to me.

Based upon your previous issues with high ringing voltages you might check to see if you are ringing beyond the voltage capacity of the MOSFET.  Other factors such as gate capacitance, switching loses, Rdson could also be sources of problems if they are not specified correctly for your alternate mosfets.