tweaking LNK626 flyback circuit
I recently built up a DC DC flyback converter using the LNK626 it takes in 60V-100V dc and outputs 5V at 1A(max)... efficiency is about 70%, I have 2 questions
under no load condition (only my output cap and inductor on the output) my voltage drifts upwards... to solve this I threw in a 1k resistor at the output, so no load the 1k is pulling 25mA, is there a more efficient way to handle no load condition?
I'd like to increase my efficiency, is there any additional circuitry that I could add to help increase efficiency or is 70% pretty typical for my power requriements
thanks
my circuit is essentialy the same as the basic one on the front page of the datasheet except I have a clamp on my primary and a pi filter on my output
Comments
thanks for the tips
how would I find the right pre-load resistor value? I can sit at the bench and increase it until I notice the output votlage start to rise... but that doesn't seem like a questionable approach
lozmon -
The value of your pre-load resister will ultimately depend on your application. If no-load input power and efficiency are important for your application but you're less worried about output regulation at no/light-load conditions, you can use higher value resister for your pre-load.
If no/light-load regulation is ultimately more important, then you might end up using a lower value resister for your pre-load. This will give you better output regulation at no/light-load at the expense of efficiency and no-load input power.
-The Traveler
I'm not sure what you mean by bias supply winding, I have a bias winding sending my feedback to the chip.... are you talking about something different?
lozmon -
Again, it really depends on your application and the specs for your power supply.
If you have very strict requirements that the output voltage stay within regulation limits, even at light/no-load, then you will need a larger pre-load (smaller resister value).
If your power supply specs for output regulation can be loosened at light/no-load, then you can use a smaller pre-load (higher resister value).
Once you know exactly what your requirements are, it's a fairly simple matter of experimenting with some different values of pre-load resisters until you meet your desired performance.
-The Traveler
No-load output voltage rise in the LinkSwitch product family is fairly common. With the on/off control of the internal state machine, the output voltage is sampled less frequently at no-load than full-load. This coupled with not using a pre-load resister on the output can cause the output voltage to rise to fairly high values.
The simple workaround is the solution you're using...a simple pre-load resister. You'll need to find a value of pre-load resister that balances the power loss of the pre-load with your no-load input power requirements, efficiency requirements, etc.
Regarding efficiency, there are a few things you can do:
- if you're not already doing so, consider adding a bias supply winding.
- evaluate your transformer design and work on minimizing primary leakage inductance as well as your winding losses.
- experiment with different clamp circuitry components. With only a 100V DC input, you have a lot of room to be flexible here. Make sure you check your feedback winding waveform for ringing during MOSFET turn-off as this can cause regulation problems.
- With the 1 amp output, consider a different output diode setup. You could parallel two diodes or find one with a low forward voltage to minimize losses.
-The Traveler