Based on my initial calculation, the primary inductance of my transformer will be around Lp = 351uH. This 351uH is designed with switching frequency of 55kHz for AC-DC flyback converter (DCM). By using LT SPICE, I will get result as below:
- Primary inductance, Lp = 351uH
- Switching frequency, fsw = 55kHz
- Primary peak current, Ipri pk = 2.30A
- Primary RMS current, Ipri rms = 0.91A
- Secondary peak current, Isec pk = 8.58A
- Secondary RMS current, Isec rms = 3.42A
From the initial Lp, I calculate minimum number of primary turns using below equation:
From above, I will get Np = ~ 24 turns (assume Bmax = 0.3T, Ae = 0.000120)
From Np = 24 turns, I calculate the secondary winding using below equation (assume Vout = 24V, Vdiode = 1V, Vr = 90V)
From that, I will get Nsec = ~ 7 turns
Therefore Npri = 24 turns, Nsec = 7 turns.
Then suddenly I found this document AN-4193 Design Guideline for Flyback Travel Adapter using FAN602 and it stated that:
Based on that document, I checked Np for both during (a) primary peak & (b) during transient/fault condition. By that I will get different number of turns with the same turns ratio as below:
- By using equation 14, Np will be ~ 24 turns, Ns will be ~ 7 turns, Nsp=0.2916
- By using equation 15, Np will be ~ 30 turns, Ns will be ~ 9 turns, Nsp=0.3000
- For equation 15, I'm using Bmax = 0.41T & IDS LIM = 4.09A (IDS LIM is based on current limit setting 0.9V/0.22 = 4.09A)
- My transformer has an air gap ~ 0.247mm
- As you can see above, both transformer A & B will have same turn ratio with different number of turns between primary & secondary. Will this impact the transformer performance? (ex. saturation level, peak current)?
- As you can see, my secondary peak current is 8.58A & secondary RMS current is 3.42A. Will this secondary RMS current 3.42A is able to go up until 8.58A without saturation? If not, what parameter should I improve?
- If this power supply is suddenly loaded with 3.7A for short duration, can this transformer handle it without saturation (same question as #2)
- Can problem #2 & #3 be improve by changing topology from normal flyback DCM (fixed frequency) to quasi-resonant flyback (frequency change based on output load)?