Flyback uc3845 transformer core hating up.

fifi_22

New member
Hello!

I had recently build flyback SMPS based on UC3845 chip. It operates correctly - no noise, stable output voltage, full power capable (24V/100W). I have Just one problem: Transformer core (not windings) is heating up (too hot to keep my hand on it) even at power of 50W.

operating freq: 80Khz

Details of transformer are calculated using Flyback 6000 software (which I found here) - I used EI33 core from ATX supply - rewound transformer and added gap (put 6 pieces of paper between halves of the core) to achieve 1700uH primary inductance with 81 turns.

Calculated data:

flyback.png

Schematic in attachement.

If any additional data needed - I will add.

Thanks for help!
 

Attachments

  • flyback.pdf
    75.7 KB · Views: 12

Silvio

Member
Hello there, I am just seeing your post about this flyback smps. One thing that comes to my mind that your core is heating up is the core material may not be suitable for such a high frequency. Why not try a lower frequency for example 50Khz and reduce the gap accordingly for the required inductance. I see you are using Excellent IT for your calculations. You can adjust accordingly to the new switching frequency.
You also have to keep in mind that for a flyback purpose EI33 cannot give more than a 100 watts or so. The reason is now being a flyback topology instead of a Half bridge only one quadrant of the flux is used thus the power is more or less divided by 4.

Good luck regards Silvio
 
Last edited:

fifi_22

New member
I tried different frequency (~55Khz) and result is exactly the same. I think it is not possible to get any better with heat from transformer... I measured efficiency of ~83% is that acceptable? Also - what themperature rise of transformer is save?
 

Silvio

Member
I tried different frequency (~55Khz) and result is exactly the same. I think it is not possible to get any better with heat from transformer... I measured efficiency of ~83% is that acceptable? Also - what themperature rise of transformer is save?
well it depends how much the transformer heats up. It can go up to 80 degrees Celsius with no ill effects as long it does not go beyond 100 degrees. What heats up the transformer is the amount of magnetic flux and switching speed. It may be most likely that the flux density is a little high in your case. This will require more inductance in the primary winding to cure it. Do not over do it because then you will loose power from your transformer.

As a rule of thumb when measuring temperature with your hand or finger keep in mind that anything hotter than 37 degrees will feel hot to the touch. At 45-50 degrees you will not bear to keep your finger there and you will remove it. This is to give you an idea of what you may be feeling. If you have a means of measuring the actual temperature then your guessing will end :)
 

liviu

New member
Your calculated frequency = 78 kHz. UC3845 is "50%" duty cycle. Output frequency in this case is 78/2=39 kHz.
Use in Flyback 6000 40kHz
 
Top