So I actually built one of the boards now. Only tested so far that it outputs 5V (more tests later). Still waiting on the common mode inductor (so jumpered so far). It was hard to find, but the one I designed it for is available at Digikey, part number 817-2151-ND‎ (manufacturer Schaffner RN202-0.3-02-12M‎).

One tricky thing I noticed about the original design by sundberg84 is that he incorporates a thermal fuse. These are really tricky to solder and are normally crimped in production devices, e.g. for fans and such. If you hold the solder iron more than half of a second to the leg you will pop it (you will hear a tiny click when the spring releases) and you have to discard it. Now it isn't absolutely needed on this board, you could just put a jumper in place. But here is a trick if you want to solder a thermal fuse.

Don't solder any other components before the thermal fuse (makes it more easy). Take a small bowl of water and submerge the fuse in water while soldering it. Keep it submerged a while after soldering is ready, to let it cool down completely. This is a 100% reliable way to not destroy the fuse.

Here is a picture of the thermal fuse being soldered:

So I actually built one of the boards now. Only tested so far that it outputs 5V (more tests later). Still waiting on the common mode inductor (so jumpered so far). It was hard to find, but the one I designed it for is available at Digikey, part number 817-2151-ND‎ (manufacturer Schaffner RN202-0.3-02-12M‎).

One tricky thing I noticed about the original design by sundberg84 is that he incorporates a thermal fuse. These are really tricky to solder and are normally crimped in production devices, e.g. for fans and such. If you hold the solder iron more than half of a second to the leg you will pop it (you will hear a tiny click when the spring releases) and you have to discard it. Now it isn't absolutely needed on this board, you could just put a jumper in place. But here is a trick if you want to solder a thermal fuse.

Don't solder any other components before the thermal fuse (makes it more easy). Take a small bowl of water and submerge the fuse in water while soldering it. Keep it submerged a while after soldering is ready, to let it cool down completely. This is a 100% reliable way to not destroy the fuse.

Here is a picture of the thermal fuse being soldered:

That's some impressive research. Personally I think you don't have to be that hard on yourself when creating hobbyist equipment and devices. With that I mean you might not have to build the device according to the most rigorous safety level. It of course depends on what you are building and how it is going to be used. (Disclaimer; I have a degree in electrical engineering, but I haven't been working with this stuff actively, only as a hobby).

Now there is another aspect also (that is mentioned in this thread). There are competing standards that might or might not be applied in various countries. And most importantly, if you build a device for commercial use, you have to follow the standards in your own area. This might take years of research and work and isn't possible for the hobbyist.

There are of course the general IP classes. If you build a device for IP level 20, the standard says:

"Parts in mains voltage potential should have at least a 6 mm aerial distance and 3 mm insulated distance to conductive parts that the user of the appliance can touch"

This means for instance the distance between a circuit board with mains voltage and the chassis.

Then there is IPC-9592 for power conversion devices. It defines a safe creepage distance as:

0.6 + voltage * 0.005 = creepage distance in mm.

For instance for 250V that would be 1.85 mm, which I think is quite a good rule of thumb for hobbyist applications. That is around Pollution degree 2, II, in your table (which was from DIN EN 60664-1?).

Then there is UL 6500, which is an older standard for audio equipment. They define "The least safe distance in mm between the two differentials in voltage" as:

d = 10^(0.78log(U/300))

Where d is in mm and U in volt. Thus 250 V would be 0.87 mm. That seems rather low, though, if comparing with other standards.

As has been already said in this thread, you have to look at what standards you need to adhere to. But for hobbyist stuff and for devices you build for yourself, you yourself ultimately have the responsibility. And maybe you don't have to follow the most rigorous standard.

Hopefully people that build commercial devices have this all covered. I couldn't imagine that they copy a design from a hobbyist board and create a commercial device based on that. Not when it involves mains voltage. Low voltage stuff is maybe another matter.

I've modified the PCB and added support for the 5W module (you can select to use either the 3W or 5W on same PCB). I've also added screw holes. Here is the list of changes:

• Imported to KiCad 5.1.8
• Added M2.5 screw holes
• Added EMI/EMC filter
• Added HLK-5MXX 5W models
• Board size: 72 mm * 42 mm

I've also changed track width to a more proper one for the respective voltages (secondary has larger current). You can leave out the EMI/EMC filter (add jump wires instead of the inductor) as it is not tested. I have not updated the BOM, but e.g. the inductor part id is mentioned in the circuit and the X2 cap is of common size.

You can download the updated design here: https://johanh.net/hlk/HLK-5M12_Breakout_board.zip

Comments are appreciated. I'm new to KiCad, so I might have done some mistakes. The KiCad project should maybe have been named as it was originally, but I renamed it to keep it separate.