LinkSwitch-TN optocoupler feedback

Posted by: qa9b on 02/22/2013

Hi all,

I'm putting together a design with the LNK305 and was curious about the optocoupler feedback method. A schematic from the app note is attached.

When the zener diode is reverse biased, its zener voltage is present across its terminals. When Vo is sufficiently above Vz, the optocoupler LED is forward biased and current flows through the zener diode to ground.

I'm confused as to how this is not horrible for efficiency. Is the LNK305 feedback mechanism so fast that it prevents any appreciable current from flowing directly to ground?

Furthermore, are there any important considerations to be made when selecting an appropriate optocoupler? Such as Vce, max of its output transistor?

Thanks in advance!

Files

optofeedback.JPG (33.78 KB)

Comments

LinkSwitch-TN is an on/off control device. When the current into the feedback pin reaches a certain threshold, the devices stops switching. With the zener/opto combination, as the output voltage rises above Vz, the zener begins conducting. As the current through the zener increases, the voltage across Rz will also increase. Once Vrz is high enough to forward bias the optocoupler diode, current will be fed into the feedback pin from the emitter of the optocoupler. So in this setup, the on/off control is regulating around a value approximately 1V above the zener voltage value and the opto is not conducting all the time.

Regards,

-The Traveler

Thanks for the input!

Still, I'm not completely satisfied. It's my understanding that the on/off control will not turn off in the middle of a conduction cycle... assume MDCM mode. If the feedback current is sufficiently high, the LNK will not turn on at the beginning of its next clock cycle.

Could you imagine a case where the optocoupler diode begins conducting immediatley after the LNK turns on? As I said before, I don't think the datasheet claims that it will shut off mid-conduction. It waits until it reaches it current limit. Then it would decide to stay in the off state during the next clock cycle.

This means that current could flow to ground through the optocoupler for the majority of a switching period, especially with larger inductors which would sustain that current flow.

Am I missing anything?

Thanks again!

There is always going to be some conduction (on average at least) through the zener and through the diode of the optocoupler. Without current flowing through these devices, they wouldn't be functioning.

When LNK-TN switches and power is delivered to the output, the output voltage will rise. If the output voltage rises sufficiently, the optocoupler will conduct and feed current into the feedback pin of the device. If the current into the feedback pin reaches a specific threshold, the NEXT switching cycle is skipped. The device will not stop mid-cycle.

If you are concerned about the the power dissipation through the opto/zener combination, you can experiment with adding a series current limiting resister.

-The Traveler