Switched mode power supply noise test

[ahmedjouirou]

Hi , i m designing a product and i m currently using these low cost AC DC switching mode power supply (photo attached below). it's based on THX208 controller . The problem is when i test this PSU in a noisy mains AC line it starts to randomly stop the switching (like for 20 ms) and then on again . The thing that helped to prevent this thing from happening is by replacing the C3 on OB pin (which was 10 nF and replace it with 100nF) it seems to work fine.

But i'm not sure of it s performance in more noisy AC line (i m testing it on both 110VAC and 230VAC using an auto transformer). Is there a way to produce/ADD noise the mains AC to verify the PSU performance  ? How to generate that noise and superpose it on a 230/110VAC line ? thanks !

[EPTech]

Hi There,

It looks to me that Q2 does the actual switching and Q1 will ramp up current during soft-start. the soft-start timing is handled  by C3 and the series resistor on top. But it is also a low pass filter for the OB input and by making C3 higher you will filter out more lower frequencies which may mess up the start circuit int the THX208.

This aside we are missing a lot of critical components for a reliable PS, like a common-mode choke, a differential mode filter capacitor, eventually a common mode filter capacity, a Varistor. These are all components that make the supply more immune to mains disturbances.

This psu is good for reset insensitive devices like unmanaged chargers but not for reliable devices.

If you want to keep using it, I recommend at least a line filter in your device. They also come in the cheap flavor. 

[ahmedjouirou]

thanks blueskull and EPTech .  Actually when i changed the C3 at OB pin from 10nF to 100nF i didnt see the shut off afain (i compared it with orifinal one 10nF and a modified one 100nF on the same AC line , the one with 10nF stopped randomly and the modified one never stopped , but i dont really like "magic stuff", i need a reasonable realability) . blueskull  For the C6 which is 10uF electrolytic you suggest to add a film cap , how will this help ? .
EPTech , you are right , i have problem with physical space of the product i cannot really put a choke (maybe i ll put a ferrite on the wires) .
i dont really understand why when i replace C3 with higher value , the thing works ok (but i dont know at which max  noise level ...) , if C3 is a bit higher(10nF to 100nF) , the rise time on pin OB (1) will increase by 10 (t = R2*C3 ) , R2 = 2.2M, so the base of Q1 will or not even reach threshold  so the OC output (transformer switchin) will be more unpredictable .. but .. it kinda make it works better .. why ?:S

[T3sl4co1l]

"Power tube base", hmm, but where's the heater?

I find it noteworthy that there's no EMI filter on primary or secondary... this thing is going to shit out some noise if you aren't providing one externally.

Tim

[ahmedjouirou]

Well I didn find English data sheet so I just google translated the chinese one , any suggestion for minimal filtering ? Or think that just a ferrite on wire will do the job ?