Choosing transformer core

Y

ytsoc

New member
Hello everyone

I need to build a smps for a car amplifier(class D)
The amp requiers a single rail of 50v and about 6 amps .
Therefor the smps needs to deliver 300w, and it will be a classic push pull topology.
Problem is i do not know how to design the transformer, If for the turn ratios i did find some formulas(altough i do not know if they apply to push pull top) i have no ideea how to determine what the max powwer for a given core can be.
In here they use a 29mm ring core for 500w(same core for 1kw??) . In the datasheet for that core the core losses are 614 mW(112 mW/cm3).
I want to use this core(3F3 material) Will this core be able to handle continous 300w? How do i determine the max power for said core?
Thank you!
 
W

wally7856

New member
The formula for the number of turns does not have wattage in it, although the parameters will affect power. At 50khz i would not exceed 160mT. This will increase the number of winding’s, which is good because 2 turns is to low anyway to realistically wind.

Frequency directly affects power. According to Pressman’s book that toroid is good for
320W at 60khz at 160mT.
283W at 53khz at 160mT.

To get 300W out of it at 50khz you can increase the flux level although 340mT seems excessive to me. I personally think that 340mT level is to high and you will have an unstable push pull.

With that said i think you should run 53khz and wind your transformer as follows.
pri = 5 turns of 10 strands of 24awg. Equally spaced all around the core.
Then your secondary.
Then your 2nd pri.

The 5 primary turns will be just over 160mT.
 
W

wally7856

New member
For the secondary if you only need one rail of 50vdc, i would do this. One winding to a full bridge rectifier.

Sec 26 turns of 3 strands of 24awg. 3 strands is just enough and 4 strands will not fit.

This was calculated assuming your battery voltage could drop down to 10vdc and you have 2v drop on bridge rectifier.
 
Y

ytsoc

New member
thank you wally,

that flux(340mT) is achieved with a 42 degrees C increase of temp (above ambient tremp)
But i don't want to use that so i'll just increase number of turns in primary side to decrease Bmax.

I was afraid that Bmax affects the power output of the core.
 
W

wally7856

New member
Ytsoc, can you buy that Ferroxcube toroid. Usually they are only sold by distributors who have $75 to $100 minimums. Let us know if you have a good source to buy them or others.
 
Y

ytsoc

New member
do you mean the TN32 core?
I'm from europe and i don't know what stores you have there.
In here it's found either at farnell or TME via a local store that orders them from those 2 suppliers. Price/unit is about 7 dolars or even lower, i'll see when i buy it.
Until then i'm gonna wind a EI(from an atx psu) core just to see how it works
 
Last edited:
W

wally7856

New member
Yes the TN32 core, I am so used to not finding what i want at regular electronic distributors i did not even check. I looked that core over a little more now and i see it is a “gapped” toroid core. This is confusing me a little because the 3F3 is a power ferrite but gapping is usually done for an inductor.

Good idea on practicing on a EI from an ATX supply. You have to be a little careful though. Some atx were forward converters that use a low permeability core’s, and will work poorly for a push pull. If you still have the atx supply, try to trace out the circuit and see what you have.
 
Y

ytsoc

New member
yes, i see, strange, I sent an email to see how much the tx40 will cost, but maybe an inbetween solution is a tx36 3c90
 
Y

ytsoc

New member
well that seems the best option , what is Bmax for 30Khz?
From their software is 215mT at 5turns
 
Last edited:
You must log in or register to reply here.
Top