Fine Tuning does not work properly

4 posts / 0 new

My design is supposed to have a 17 V Output. Measured Output Voltage of the prototype was 14.6 V.

Consequently I entered this value in PIXIs Designer 9 in [49] "Measured Output Value". The proposal for RLOWER_FINE was 6,5 KOhm instead of 5,5 KOhms accordingly. I adjusted the lower resistor and the result was an even lower output voltage of 11 V. The Decrease of the lower resistor to 4,7 KOhm lead a higher output voltage of 15.5 V, which were a bit more unstable than the 14.6 V before.

How come PIXIs Designer suggests a higher resistor which makes output-value become worse? 

Thank you.

FineTuning_LinkSwitch-CV.pixls45.5 KB

The increase in the lower feedback resistor is as expected. I would recommend you perform the following:

1. Put back the 6.5k resistor and you should see 11V. Look at the voltage on the feedback winding when the MOSFET turns off. Check if there is excessive high-frequency ringing (from the primary leakage inductance) on this voltage waveform.

2. What type of diode are you using on the feedback winding/clamp? Try using slow diodes if you are using fast diodes.

Let me know if you are able to solve the problem with the recommendations mentioned above.

Thank you very much for the answer.

In the picture attached, you can see what my Feedback-Voltage looks like. I used a superfast-diode (trr_max < 50ns). When I replace it with a slower one, the voltage-level stays the same but the ringig during the on-state disapears and also does the efficieny suffer. 

I haven't replaced the clamp diode (trr<75ns), yet. I will do that as well. 

I was wondering how a larger RLOWER (6,5 kOhms instead of 4,7 kOhms) should lead to an increase of secondary voltage. Using a larger RLOWER, a smaller induced voltage in the feedback curcuit is required to attain V_FB = 1.84V. The secondary voltage will be lower accordingly and I don't understand why the Fine Tuning Option suggests a larger RLOWER.

Feedback with super fast diode51.39 KB

The FB pin of LinkSwitch-CV looks at the DC amplitude non-rectified bias winding AC signal during the primary MOSFET turn-off interval to calculate the output voltage. If the design turns ratio is fixed, the only handle that controls the secondary voltage is the feedback network. Keeping the Rupper fixed, increasing the Rlower increases the voltage seen by the FB pin at the sampling time. Since you are seeing a lower secondary voltage, the fine tuning section suggests a higher Rlower. Hope this helps.