I also got an email from Tom at smart-sense.hr saying that he had used one of the HLK-PM01 modules, and his emission graphs were similar to the ones I produced in the last post. He had also been in contact with people at Hi-Link, the manufacturer who sent him the following solution.
I don’t have any 10 mH chokes (Well I probably do, I just have no idea which ones they are) however I have some 20 mH chokes, and a Schaffner RN102-0.6/02 which is 2 x 4.4 mH chokes in common mode (Farnell number 2097049).
I don’t use Inductors all that much as a component that I spec by value, but on paper Serial and Parallel Inductors behave nicely, so lets have a look at doing that:
At work we have two LCR meters, a Wayne Kerr B424/N and an Sourcetronic ST2830 (ST2830 LCR Meter Manual), so setting the RN102-0.6-02 up to each of them in a number of configurations I managed to get some values (easer than the maths, and less likely to be wrong):
|Wayne Kerr Meter B424/N||Sourcetronic ST2830 LCR Meter|
|5.39 mH||7.59 mH|
|73 μH||50.82 μH|
|23.9 mH||27.24 mH|
|6.51 mH||8.75 mH|
|27 μH||12.78 μH|
Well there goes that cunning plan…
It turns out that if you put two inductors in series, that while yes you double the turns, and double the inductance, this only works if you are not doubling the turns on the same core! just doubling the turns on a core quadruples the inductance, as we can see from above.
Lucky for me, I had a Schaffner RN114-1.2-02 2 x 10 mH 1.2 A laying around, so I will use one side of that.
Looking back at the original results, which were performed with the following setup. We can see that it is only the conducted emission we have to try and win back:
So lets look at the results using Filter 1:
As you can see while there is some improvement of the Live conductor however the Neutral conductor still lets the circuit down. unlike last time I am not looking at the Class A limit lines due to BS EN 55022 “Information technology equipment. Radio disturbance characteristics. Limits and methods of measurement” [pdf] no longer being a current standard, and its replacement BS EN 55032:2015 “Electromagnetic compatibility of multimedia equipment. Emission Requirements” further restricts the get out clause for using Class A limits:
However also in Tom’s email was a solution that he had come up with that has slightly more components (And lucky for me the RN102-0.6-02 is the correct component):
Building it up out of components and wire I had laying around, I think it goes without saying that I can do daft things because I know what I am doing, but here it goes anyway. Please don’t use this size of wire when you do anything like this!
This is a common mode filter, any common mode noise should be removed by the choke, the additional capacitance (and ground wiring) should help to take out differential mode noise. lets see how it goes…
Ok so the results from testing:
Some how I have managed to kill the HLK-PM01, as I don’t trust the results I have got from the Neutral line scan, however the live side scan does show a close pass to Class B limits.
While I would not be happy using this Switched mode PSU to make anything I was going to place on the market, I may consider it!
If you are making a product that users are required to plug in, Please, please send CASS Industries an email to ask about testing!
The test results on this page are indicative, and in no way constitute evidence of a pass! and as the Irish or Welsh would say “Well sir, if I were you, I wouldn’t start from here” I would go with something like a VIGORTRONIX VTX-214-005-105 AC/DC PCB Mount Power Supply (Farnell #2401040 ), as for only a few quid more you win a switch mode PSU that stands a chance of not requiring external components to pass Class B!