Offline full-bridge SMPS… need help

Krisfr
If you get the pcb of Luka's supply from this thread and print it on a laser printer,take to any printshop,Kinko's or other even your high school photolab can make you a negative or a postive transfer(make sure its a mirror image)Hold the pcb up to a light source to check.Then you can process it and start soldering parts.
Thanks for that tip, I will try it today...

That makes a lot of sense...
 
A 110 Volt design with ETD49

I am going to do a 110 volt design slowly and build it with the idea of incorporating ALL that I have learned by reading this thread. I have some parts but want to do a LTSpice to make sure I am in the correct ball park.

I have time and I am not in a hurry. Just want the ducks in a row before I shoot off my gun.

MANY thanks to ALL for input and advice...
 
Luka"s SMPS

Hello from Croatia! I am new here and quite interested about Luka"s SMPS. It is great design, as far i can uderstend SMPS, and i will like to make one as PSU for my audio amplifier. Amp is build around TDA 7294, so power demand for that type of amp will be +/- 30-35v and 10 A peak(in stereo). I explored this topic from very beginning but i still have some questions about output transformer. Well, what tipe of awg wire to use for pri & sec windings, and how to winde then, and how many turns. for example:

pri: 26 turns of 3x19awg
sec: 2x12 turn of 5x19awg
layers:

1/2 prim + 1/2sec + 1/2prim + 1/2prim

or

1/2prim + whole sec + 1/2prim.

Is there any kind of formula so i can calculate for my self?

Tnx guys in advance...

P.S.

and of course i will post pic"s of my work here, as soon as i get it done.
 
Last edited:
26 turns for etd44 is perfect, I have this right now (making is regulated finally, for same output voltage as yours)

12 turns for secondary is too much, I have 6 and I get something like +/-35 (I will use pfc for my).. I would use 7 for each sedondary

You do your trafo like this: 0.5 primary, both secondarys(or more if you have them, all that are high current) and then last part of primary

So this looks like this, 13turns, 2x 7, 13 turns

there are a lot of equations out there, just be sure to find one that is right, and you don't mix american and our units

i'll be waiting for pic's 🙂
 
Great!. Tnx Luka. I have it clear now. One more question😀, i notice on asembled one a little trim pot standing on that little external board (near that TL). Is that R3 - 56k? What are you adjusting with it? You mentioned PFC as you upgrade, look in this PDF maybe you find somting interesting. 🙂
Cheers....

Download PFC.pdf, upload your files and earn money.

That whole board is feedback, which was not in use most of the time, while it was, that pot and yellow cap there, were not in use

I already have PFC, has enough power for me, but thanks
 
Update!

So I've been trying to make it regulated. I changed few components, add back coupled inductor, added filters, primary and seconday to kill the ringing as best as I could with components at hand.

And result is... around 50mV ripple looking both rails at the same time, that is ~+/-35v (load independent!!)

Load step from about 80w to 377w with probably only 0.1v drop and I say probably coz its that low. when on AC coupled, there this could be seen, but on 5v/div, and still seeing 70v there is no visible drop, rock solid.

So i am 😛 right now
 
Last edited:
hi luka,

i am planning to build your smps but with etd49 core.how did you calculate
your primary turns.iam coming with different results.here are my calculations.

input voltages =180-270 vac
output voltages =100 volts @ 5amp

output ripple =50mV p-p

power output =100x5=500w

power input =po/eff=500/0.8=625w

dc input vots

vin low =1.414x180=254 vdc

vin high =1.414x270=382

av input current

highest average current =500/254=1.96 amp

lowest average current =500/382=1.30 amp

max peak current

Ipk =2.8x500/254=5.51 amp

no of primary turns

npri=vin(max)x10^8/4xfxbxAc=382x10^8/4x50000x1600x2.11

=57 turns

these all are taken from marty browns 2nd edition

i have read bmax of 1600 is safe value i think its in pressmans book.

how do we find the safe value of bmax.i have read the book but some stuff
is going out of my mind.what value of bmax you are using

bsat is determined after finding the no of primary turns .that means we already using the bmax value.

i have many questions .i cant remember right now.i have read the thread
twice.help me


thanks in advance

ravs


be a vegeterian.....
 
Bmax

hi luka,

i am planning to build your smps but with etd49 core.how did you calculate
your primary turns.iam coming with different results.here are my calculations.

input voltages =180-270 vac
output voltages =100 volts @ 5amp

output ripple =50mV p-p

power output =100x5=500w

power input =po/eff=500/0.8=625w

dc input vots

vin low =1.414x180=254 vdc

vin high =1.414x270=382

av input current

highest average current =500/254=1.96 amp

lowest average current =500/382=1.30 amp

max peak current

Ipk =2.8x500/254=5.51 amp

no of primary turns

npri=vin(max)x10^8/4xfxbxAc=382x10^8/4x50000x1600x2.11

=57 turns

these all are taken from marty browns 2nd edition

i have read bmax of 1600 is safe value i think its in pressmans book.

how do we find the safe value of bmax.i have read the book but some stuff
is going out of my mind.what value of bmax you are using

bsat is determined after finding the no of primary turns .that means we already using the bmax value.

i have many questions .i cant remember right now.i have read the thread
twice.help me


thanks in advance

ravs


be a vegeterian.....


Bmax value for ETD49 core is 1600? how is that?
 
The Bmax value is determined by the type of material the core is made from, the ambient temp of the core in full load, etc.
all of these you can find and calculate if you search the datasheet of the material.
for example 3C90, 3F3, etc

regards,
savu
 
123-1.jpg


Vp min is half of the min DC voltage on input 128v or so.
f = freq. in Hz
Ac = core cross section in cm2, 1.73cm2 for etd44
Bac is let say half of the Bmax, in Tesla's, so 0,2T = 2000Gauss
Kf = 4 for squere waveform, other for others


For my was:

193v*10^4/(50000*1.73*0.21*4)= 26.56T, used 26