Peak currents in LNK3205 design

Posted by: greanie on 04/22/2024

I'm designing my first circuit with the LinkSwitch-TN2, in buck and MDCM configuration, for a line-connected AC/DC application. I have read through the data sheet and app notes AN-70 and RDR-721. I've selected the LNK3205 based on maximum current consumption requirements, up to ~ 150mA with everything turned on. I have used the online tool to size the components and spice to model the suggested circuit to estimate maximum currents. I read that you do not offer spice models for your components so I represent the LNK3205 with a pmos switched at constant duty cycle (with 14.3us being the smallest time increment possible with a 70kHz switching frequency), and my load as a resistor. I understand that the result is not accurate, I hope it is nonetheless representative.
Based on the simulation results, I have a few questions :
1. What is the purpose of the LC filter between the output of the rectifier capacitor and the input of the LNK3205 ? The cutoff frequency is pretty high for this application, no ? I understand that the inductor is more efficient than the resistor at higher powers, but why not just add more capacitance ? The current into the LNK is much cleaner with a resistor than with the inductor. Do you recommend for lower noise, a resistor or an inductor ?
1. I have a 230VAC line voltage input and want 6V output. The capacitor-filtered, rectified voltage swings between 320 - 240V. That should result in 1.8-2.5% duty cycle. My simulation shows a duty cycle of 0.3%, set with an on-time of 14.2us (based on 70kHz switching frequency) and 8.5m period. It this a duty cycle reasonable ?
3. At fixed duty cycle my output voltage swings from 5 - 17V. Your design tool says I will get a better result, very good. However, I get a peak current of 3.5 - 5A out of L2 (output inductor), depending on its inductance, following the brief on-time. This seems really high. Can you confirm that the output current will be in this order of magnitude for my implementation ?
Thanks in advance for your clarifications. The LNK3205 seems like a great component, I'm looking forward to getting to understand it better.

Files

Attachment	Size
Simulation Model	66.55 KB
Schematic	69.03 KB

Comments

Hi Greanie!

The capacitor-inductor-capacitor network after the rectifier forms a pi filter used as EMI suppression. CLC configuration is used to better filter high frequency noise. Usually CLC filters can more efficiently filter ripples and noises without heating up compared to a CRC filter. CRC and C filters can be used but and the design of this will depend on actual application especially if certain industry standards are to be met.
Due to the high difference of input voltage and output voltage, it is expected to see a low duty to achieve regulation. LNK-TN2 operates using an on-off control scheme. With that, it will switch at a fixed current limit with a duty no less than the minimum on-time stated in the datasheet. Since switching at a fixed current limit constantly will cause the output increase, the LNK-TN2 (with its on-off control scheme) will remove pulses some pulses to be able to regulate the output. You can see a good explanation of this scheme in AN-70. From my understanding, your simulation assumes a constant PWM signal at a fixed duty and frequency to achieve regulation which is why the expected duty is very low.
The duty output of the LNKTN2 is dependent on the loading conditions so fixing the duty cycle and operating in open loop configuration can cause discrepancy with actual operation of the IC. The average output current of the inductor is equal to the output current. Computation on inductor current output and ripples can be found in AN-70. This should help you estimate expected magnitude of currents.

Thanks for looking into using LNK-TN2. I hope that it will be suitable for your needs.

PI-CookieMonster