LNK302 isolating power supply

Power supplies requiring isolation arn't really the target application for LinkSwitch-TN. It can be done but you're going to run into other problems first.


Two problems you'll run into with LNK302 and your design:
-Your output power exceeds the device capabilities.
-You input voltage is right on the edge of what is required to start-up IC for proper operation.


One of our products is a bit better suited for your application though:
http://www.powerint.com/en/products/dpa-switch-family/dpa-switch


DPA Switch is intended for low voltage DC input applications and can have some very nice operating characteristics due to some of the unique features of DPA Switch.


I would recommend downloading our PI Expert software (if you haven't already) and create a DPA Switch design using your specs.

-The Traveler

OK I understand this chip is not designed for that.

The 12V/4A circuit just for reference, adding external MOSFET and using isolated gate drive.

The LNK302 circuit only for what the LNK302 can provide directly.

When you refer to input voltage did you read it like I wanted to build 60V input, 12V/4A output using LNK302?

Grid voltage input of course, and only test circuit/for reference, output = 3 volts or 5 volts, using a small ringcore eventually as tranformer.

What I want to change is from using a coil + no isolation to using a small ringcore (like seen in commercial 12v halogen transformers), providing isolation.

If this makes any sense or is easy to do?

The software I am downloading now.

takao21203 -


It wasn't clear from your original post what you intended to do. If you can post your power supply specifications, I can help point you in the right direction. I need to know: AC or DC input voltages, output voltages and currents, etc. Any other requirements that are specific to your design would be helpful as well. For example, what is the intended application for your design, operating temperatures, ripple requirements, etc?


Regarding your question about a ringcore. I'm guessing you are referring to a toroidal core. Typically, these types of magnetic cores don't make for very good power transformers. Inductors, yes. Transformers....most of the time, no.

-The Traveler

I have ordered 2x LNK302 together with other parts, have encountered information about this component a few months ago.

Typical application for instance to power 0.36" serial LED display + small controller.
It is here powered from 12v el. transformer + 7805 which only slightly heats up, so current might be around or less 100mA. The display refresh will relate to pulsed load current not static, so maybe the peak current is more than 100mA.

These can work from 2.5V to 5V.
The accompanying controller needs 10 or 20mA as well.
I have attached a photo, where this circuit is built to measure temperature + control cooling fan.

This together eventually should become powered directly from 240 volts AC, using as less components as possible.

But isolation would be desireable, if possible.

And why are toroids (toroidal cores) less considerable/good to use for transformers?
Because there is no gap? I don't have full background for the formulas but usually I get along with inductors, also have access to related literature.

They are used inside CFLs as control transformer, and even directly inside some 12v halogen transformers (I don't have any of these here currently for reference). Most el. wall adapters use a ferrite transformer core (rectangular/strctured), have 3 here for reference.

Or is it that the power levels that are possible using a torodial transformer are lower? Please explain if possible. I am recently increasingly interested in these topics.

I saw one book text where a small gap is added to a toroid, in the explanation the author explains this would increase saturation so the toroid would saturate earlier. Well this won't be of much use if it is used to transfer power?

If using a torodial core only results in some degree of efficiency loss this is no problem at this power level, I would prefer to use ringcores.

OK I have tried the design software.

I now understand the LNK302 is for non-isolated supply using inductor only,
and the LNK362 would be matching for flyback transformer.

Is it possible in general, to build the LNK302 into a transformer circuit?

Like the circuit I built with LM2575, also outside specification?

Efficiency is not so important, only for reference/also possible to change over to LNK362.

Also transformer ferrite core, the smallest in the software is for 2.9 Watts.

I would rather prefer to ask (before experiment) for instance here on the forum, because grid voltage is dangerous, not just for human I also prefer to keep the LNK302 intact.

One LM2576 (pin compatible) and 6 or 7 adjustable resistors were consumed for the circuit in the attach...

So I would need about 2mH on the core, using the LNK362.

Toroidal cores tend to be used for inductors but not for transformers. This is due to a few issues: Toroidal cores only come in discrete gap sizes which limit the possibilities of inductance in any given applications. When used as transformers, toroidal cores tend to create transformers with high-leakage inductance due to poor coupling between the primary and secondary windings. Also, from a manufacturing standpoint, toroidal transformers are more difficult (read --> $$$) to build.


Toroidal cores are often used for filtering applications. They make great differential and common mode inductors, have a distributed gap which reduces radiated EMI and proximity losses and depending on the material, can have nice dependable saturation effects. This lends them for use in mag-amp power stages.


-The Traveler

It's still not clear what specs you're trying to design for.


From your sketches, I'm seeing inputs of 195-265 Vac, 60V DC, 12V DC and outputs of 2.5-5V.


Can you clarify your requirements? From what you've written, it looks like you're trying to do something like this:
Power Supply Input: 195-265 Vac, 50 Hz
Output 1: 2.5-5.0V @ 150 mA


Have you taken a look at our reference design section yet? There are a number of designs that are ready to use that already come close to meeting the specs it sounds like you're asking for.


Here is one good example to start:
http://www.powerint.com/sites/default/files/PDFFiles/di141.pdf

-The Traveler

The 60V circuit only for example using a component outside specification/intended application.

I have looked at LNK302/LNK262 functional block diagram, and they seem to be nearly identical.
LNK304-306 even seem to be 1:1 identical to LNK302 in the functional block diagram.

For the actual prototype, 4V / 100mA would be sufficient.

Various circuits around here powered by batteries, I use 12V electronic transformers as well + switching ICs to convert to 3.3 volts (sometimes 5V as well).

7805 only used for test circuits.

Some of these have low current requirements, in the 10mA to 50mA range but isolation is required.

If I can get suitable enclosures, and total bill of materials is less than $6, it would be interesting to use LNK chips instead of electronic transformers.

Sometimes I develope circuits for 3rd parties, for this reason also interesting to build for reference. I had one case where using a small 50Hz transformer for AUX supply was prefered to el. transformer module from China, or generating AUX voltage directly from high voltage DC (high voltage = more than 48V). I like to use these el. transformers for myself but can not verify/guarantee relieability. They are usually relieable, I had one here which simply made "pop" after about 3 weeks, by random I actually observed it.

takao21203 -


I sent you a private message. Wanted to ask you a couple quick questions. Please respond when you get a chance.

-The Traveler