3.3V design with LinkSwitch-TN device
I have implemented a 3.3V with LNK306. I have done the design with PIXIs Designer software. However, I am not able to get the desired 3.3V output. Once I apply a load, the output voltage goes to 2.5V.
Has anyone being able to get a reliable 3.3V output using the Buck converter topology? Does the LinkSwitch-TN devices support voltages as low as 3.3V? The software did not give me any warning or any other indication that the output voltage is too low.
Thanks
Comments
I have attached the design. I am using the 470uF inductor as recommended by the software. Coilcraft's RFB0810-471L.
Regards,
GYH
The schematic looks OK. I would try changing D6 to a UF1007. The faster diode will track D10 better.
As you change the load, look at the voltage across C4. This voltage is what the IC regulates. It should track the output voltage.
For better no load regulation, you may want to reduce the value of R26 to 470 to 1000 Ohms.
Z. Cochrane
I have ordered the diode and do the change as you have suggested. However, my question is: since this diode has Trr=75 ns, how is it faster that the BYV26C with Trr= 35ns.
Also, most of the diodes that are recommended by the software and in AN-37 are hard to source (multiple week lead time). Do you have alternatives to these diodes that can be easy to get.
Regards,
GYH
The Trr of D6 (1N4007) is not defined. A 1N4007G has a Trr of 2 us. Any fast or ultra fast diode will be a closer match to D10 compared to a 1N4007.
You can use any diode with the same or better specification (forward current, blocking voltage and recovery time) to substitute for the suggested diode.
Z. Cochrane
I replaced D6 with an NTE575 (Trr=75nS). The preload resistor was changed to 1K as well.
With the preload only, I get 3.3V on the output.
I added a 100 Ohm load, the output voltage dropped to 2.5V; I measured the voltage across C4 and got 3.3V.
GYH
It is unusual that the output is not tracking the voltage on C4. The LNK306 is correctly regulating the voltage on C4. The power supply's output regulation depends on that the output voltage= the voltage across L2 = Voltage across C4 when D10 conducts.
You may want to verify that the voltage across C4 is not peak detecting any stray inductive spikes. Verify that the connection of D6 anode is close to C5 pin 1. As you increase the load from 100 Ohms to 20 Ohms, does the power supply regulate at 2.4V?
Z. Cochrane
I put the 20 Ohm load resistor and got the same 2.5V.
I do see spikes across C4. I am not sure if they are stray spikes or they are there due to the switching of the chip.
We have followed the layout recommendation and D6 anode and C5 pin 1 are close; they are connect via a plane.
GYH
LinkSwitch TN will support a 3.3 Volt buck design. I will need to know which buck configuration you are using (high side with direct feedback, high side opto feedback or low side).
Below are some ideas to help you troubleshoot the power supply:
If you are not using an optocoupler feedback (high side buck direct feedback), you will need a minimum load to maintain regulation.
It sounds like you regulate at no load but loose regulation when a load is applied. Verify that the inductor is not saturating. The inductor needs to be able to support the peak drain current of the LNK306 (647 mA). Use a current probe and a scope to monitor the source pin current. When the switch turns on, you should not see any curvature in the current ramp waveform.
Z. Cochrane