AC-DC 100W Flyback SMPS

[gigabyte091]

Hello everybody,

after i was offline for quite a long time i decided to build flyback SMPS for my battery charger.

Here is some specs i want to achieve:

Input voltage: 230V AC (+/-10%) (about 325V DC)
Output power: 100W
Output voltage: 30V DC (adjustable from 20 to 30V if nessesery)
Output current: 3A DC (90W on main secondary, 10W on aux secondary for powering IC and cooling fan)
Fully isolated output (Feedback trought optocoupler)
short circuit protection.
Switching frequency: 70 to 75 kHz

Controller IC i will use is UC3843 because its cheap and avalible in my country.

Here is schematics, its not fully completed, there is some calculation on feedback loop, frequency compensation of the TL431 as well as freq.comp. of the controller IC itself. Also at the moment im searching for optocoupler and a MOSFET as IRFP460 would be overkill for my aplication (only about 130 to 140W on primary side).

I've done transformer calculation, choose winding method.

Here is calculation:

Core type: ETD34, material: F44

F44 max. induction @ 25°C: 500 mT; @ 100°C: 400 mT

Thermal resistance: 19°C/W

-------------------------------------------------------------------------------------------------------------------------------

Power loss on transformer:

Pv=delta(T)/Rth=40/19=2.1W

-------------------------------------------------------------------------------------------------------------------------------

Primary turns: Vimin*tonmax/B*Ae=284*6.66/0.4*113= 42 turns (Where Vimin is minimum input DC voltage; tonmax is maximum on time of the switching MOSFET; B is max core induction and Ae is core area)

Main secondary turns: (Vo+Vd)*(1-tv)*Nprim/Vimin*tv= (30+2)*(1-0.5)*42/284*0.5 = 5 turns (Where Vo is output voltage; Vd is voltage drop on rectifier, tv is maximum duty cycle, Nprim is number of turns in primary)

Aux secondary turns: 3


If there is any sugestions please be free to share it with me.

 

[MicrosiM]

Things looks good for me, This is interesting project indeed

Please keep us posted

 

[gigabyte091]

I will keep you posted, and at the end i will post complete documentation, who knows, maybe someone would need my power supply or it will be base for some improvements.

But one thing is bothering me, and that is transformer calculation, i think i use to big induction...

And im still searching for good optocoupler and MOSFET. Im very limited with component choice in my county.

Im modifying schematics in the moment, i added 1nF 400V capacitor in parallel to zener diode in snubber circuit to make it faster, also i choose MUR460 for diode in snubber, it's one of the faster diodes available to me. Also, i increased startup resistor to 330k ohms, decreased IC capacitor to 100 uF because 1000 uF was too big and in the case of short circuit there would be failure of the power supply (high capacity capacitor would prolong time requierd for hiccup mode to start and high current in secondary could damage rectifier). Added 1N5819 will rectifier aux power for IC.

Also, im thinking about simple linear regulator for FAN, something like 7812 or simple one transistor regulator with zener diode.

But primary goal for now is to calculate and wind transformer.

 

[mkuczma]

Here you are. Schematics and PCB working excellent flyback power suply build with ei33 core from PC PSU with 1 mm gap working about 120kHz. primary induktance only 650uH. primary turns: 80, Bias turns: 7. For 30V out make 16 turns
View attachment Zasilacz flyback 280W.pdf
View attachment Zasilacz flyback 280W PCB.pdf

The most important thing in flyback is makeing a good transformer. You must fill all space in core to make the lowest parasitical inductance which heats snubber. In my PSU in 200W continous power reseistor heats up tu 60 degress( I use only one 47kom) and transil get over 50 degress.

 

[gigabyte091]

@mkuczma

This seems like a nice supply, if you could give me website so i can see transformer calculations.

 

[mkuczma]

http://schmidt-walter.eit.h-da.de/smps_e/spw_smps_e.html
ei33 is like e etd39, the same Al and Ain

 

[gigabyte091]

Did you make transformer using this calculator ?

I just have recalculated my transformer and i use 0.2T induction and primary turns is about 85

 

[mkuczma]

this page shows you a value a current and a viltages, but you must know what you do. In flyback the opening time transistor compared to the period fixed voltage value that is added to the input voltage, and causing the loss in snubber. I use 30% Ton to T in full power determined by transformer ratio. Lower Ton make lower losses in snubber but higer Ipk in transistor and needs bigger transformer core for the same power.
Yes i make a several transformer using this site and always works good.

 

[gigabyte091]

i think my calculations are good now

 

[gigabyte091]

Here is updated schematics.

If someone have some material on calculating freq.comp. for TL431 i would be gratefull.

 

[mkuczma]

Transistor irf840 is not used in flyback. Needed >600V Uds.
D3 is in uc38xx.
R4 too big value. I always use 120-150kom 1W and it heats up to 35 deegres.
Correct bias shows picture http://img03.blogcu.com/images/p/w/m/pwm/31f6cf66426719c996de25677e0be15e_1267146755.png
Why you use tl in stabilization? Zener diode dont make any problems with compensation.
D2 it must be transil p6ke or p1.5ke on a right voltage(120V you used might be to small)
R3 C9 1k 1n
Bad connection slope compensation: C to vref, B to ct, E through resistor to isence
D6 change to resistor about 20k - uc have limiter voltage on output to about 16V
Why R6 have 5W?? only need to 0.125w
you need capacitore parallel to r7 about 100-220pf
All this things is in the datasheat, just read it.

 

[gigabyte091]

R4 is choosen just to provide startup current for IC, and it's calculation is based on IC startup current. D3 is here to stabilize supply voltage for IC and cooling fan, zener inside IC is 30V ones. Choosen zener snubber should do the job, i don't have choice, im buliding it from parts easy avalaible to me.

shunt should be 1 k ???

D6 is changed to schottky diode to protect fet from negative voltage, it was mistake. Slope compensation changed.

But i dont know if there is 600Vds mosfet avalaible

 

[mkuczma]

Look at the texas instruments datasheat how bias should be build, fan suply from any other winding.
mosfet try 2n60, it is often in pc standby psu.

 

[gigabyte091]

I choose MOSFET, it's 2SK2611

 

[vorezky04]

Nice information..

 

[lynxlynx]

Hey gigabyte091!

Did you succeeded with your schematic?

Just because I tried similar posted here http://www.diysmps.com/forums/showt...3843-smps-remake&p=11902&viewfull=1#post11902 and it unfortunately did not work. Just because I follow same requirements but I only need lower adjustable voltage in ranges from 4-5V to approx. 20V.

 

[lynxlynx]

Nevermind, I am wrong at transformer calculation. I once got it working nicely although with wrong voltage. I'll continue research.