SG3525 ferrite transformer calculation

Jump to Latest
roomi

roomi

Member
Joined 2008
#1
I am designing a 300W car amplifier using TAS5630B. To power the chip I am stepping up the 12V input to 40V DC using push-pull topology. My controller chip is SG3525 and I am using a PC40 material T28X16X16 size torroidal transformer.

For SG3525, I choose Rt= 22K, Ct= 1nF and Rd= 56 Ohms. Which gives oscillator frequency of 64KHz. So each MOSFET is switching at 32KHz.

My question is during calculation of primary turn of torridal transformer which frequency, f I have to consider in the formula Npri= (VinX10^8)/(4XfXBmaxXAc) ? 32Khz or 64 Khz ?
 
#2
Your formula looks strange, for me it's Np = V/(4*B*F*A), with all units in a coherent system, for example V in volts (across one half primary), B in Tesla, F in Hz and A in square meters, but maybe it's just the way it is written without a formula editor.
The frequency you need to take into account is that seen by the winding, in your case 32KHz.
Note that this formula is only valid for steady-state conditions. Without any particular precaution, the core will see a ~doubled induction during start-up, which might kill the active switching elements
 
#3
Elvee said:
Your formula looks strange, for me it's Np = V/(4*B*F*A), with all units in a coherent system, for example V in volts (across one half primary), B in Tesla, F in Hz and A in square meters, but maybe it's just the way it is written without a formula editor.
The frequency you need to take into account is that seen by the winding, in your case 32KHz.
Note that this formula is only valid for steady-state conditions. Without any particular precaution, the core will see a ~doubled induction during start-up, which might kill the active switching elements
Click to expand...

Thanks Elvee, both formula is correct as in my formula B in gauss, A is in sq cm.

For my application (push-pull), Dont you think the winding is actually seeing 64Khz as it will see both switches, each running at 32 Khz ?

Can double induction could be prevented using soft start ?
 
#8
roomi said:
The two switches are creating flux alternately, each at 32Khz. So the core itself sees 64 Khz. Appreciate if you explain why its 32Khz. 🙁
Click to expand...


the cores flux density is bi-polar, so the max level is calculated for each half cycle. volt *seconds per turn / core leg cross sectional area. so at the end of a complete cycle eg 64KHz, the core flux is reset to zero.

the key is to think in terms of volt*seconds instead of plain old frequency.
 
Last edited:
#9
infinia said:
the cores flux density is bi-polar, so the max level is calculated for each half cycle. volt *seconds per turn / core leg cross sectional area. so at the end of a complete cycle eg 64KHz, the core flux is reset to zero.

the key is to think in terms of volt*seconds instead of plain old frequency.
Click to expand...

now it makes sense! So I should use 32Khz in equation.

BTW, I didn't understand the equation how you calculated Bmax for each half cycle. Is it, Bmax=(V*Time)/(Npri*Ac) ?

What is the units of each of them?
 
#10
I just remember it that way,
iregardless of any constants, waveforms, etc

the flux change is proportional to the
integral volt-seconds per turn (directly equal in the SI
system). Faraday's Law operates bilaterally -
Click to expand...
from http://www.ti.com/lit/ml/slup123/slup123.pdf
Re units, I just use faradays law and the integral > area of the voltage drive
thus units come into play only for Bmax and Ae from the cores data sheet.

now for core losses, use double Bmax for one complete cycle > delta B
 
Last edited:
#16
You must log in or register to reply here.
Top Bottom