Designed for +/-3.3V output but shows only +/-2.6V
I followed the AN37 application note for the LNK306 to build an AC to DC converter with an output of +/-3.3V DC (page 10, Fig. 5). I've also attached the schematic here. Everything works except I see only +/- 2.6V on the output. I used PIXIs Designer software to calculate component values for an output of 6.6V. Am I correct in designing for this output voltage? The suggested value of the feedback resistor is 5.67K and the one I found was 5.62K. Any help would be appreciated. The PIXIs Designer file is also attached here. Thanks.
Comments
@Bouwser, Take a look at Fig.5 on page 11. I've wired my schematic exactly like that. Not able to interpret this the right way. Should the output voltage for design calculations be 6.6V or 3.3V?
Yes, I see now, what you were basing your design off of. You should design your resistors for 6.6 V like you did.
1. Can you just adjust one of the bias resistors until the voltages are what you want?
2. The fact that you are measuring 2.6 V, which is exactly one diode drop below your intended voltage is suspicious.
3. Are you sure the ground point for your measurement is correct? Could there be a diode between that point and the actual ground of the circuit?
Yes I'll try adjusting one of the bias resistors. Nope there is no diode between the ground and the measurement points. I wanted to make something really small and hence there are just the components that are needed. The ground point is essentially the negative terminal of the bridge rectifier.
OK did some more testing and here is the data:
Put in a trimmer in place of the Rfb and I finally saw 3.3V at a resistance of about 8K. At this value the voltage on the feedback pin was 7.4V. Also I now hear buzzing.
Attached an adjustable electronic load to see if the output remains stable and here are some values from that.
| Current in mA | Output in V |
|---|---|
| 150 | 3.3 |
| 160 | 3.3 |
| 190 | 3.3 |
| 200 | 3 |
| 250 | 2.9 |
| 300 | 2.8 |
| 350 | Shutdown |
At 300mA load current, the chip becomes hot. I measured 69 deg C.
Something is wrong (since the values do not match the design guidelines). Not able to pin point the errors.
Hi electrophile,
Thank you for your interest in using power integration parts.
Looking at your schematic and PIXls design, what I think the problem is your Rfb resistor. Your PIXls Rfb computation is based on 6.6V output in reference to the GND but looking at your schematic the Rfb should be computed only based on 3.3V output. Your transformer design is good for your required power. Your Rfb should be 1.89K (try nearest available value of 1.87K or 1.91K) based on +3.3V output in reference to the ground where your Feedback network (R2-D2) is connected. Hopefully this helps on your design. Thanks.
HI Mallora,
OK I tried it with 1.87K as the feedback resistor and all I see is 2.1V on the output now. I even tried switching the chips (thinking I might have damaged them earlier) but its still 2.1V on the output.
Hi Electrophile,
Do you have reference for that design? I'm looking at the schematic and operation wise, Im wondering how we can have a good regulation at the negative output. During turn-on, your loop will be L2-C5 then to the GNDA. Then during turn-off the inductor will have a discharge loop of C5-C6 then D1. -3V output only depends on the C6 energy provided by the inductor discharge during turn-off. Unlike +3V, it is always part of the loop.
Hmm. The reference is the AN-37 - LinkSwitch-TN Design Guide. My design is listed on page 11, figure 5. Can you recommend any solutions or alternatives?
You can refer to RDR-138 for the reference design with dual output using LinkSwitch-TN. Thanks.
I'm not an FAE with powerint, but just looking at your schematic, it looks like the anode of D1 needs to be connected to Gnd. The trace passes through the Gnd net, but does not appear to be connected to it, like the reference design says it should. I'm guessing the only reason it works at all is because of the 3.3 V zener diode you have. (D3)