Page 1 of 1

Do small SMPSes really qualify as double-insulated?

PostPosted: October 18th, 2014, 10:34 pm
by LongRunner
As they are currently built, I think not; between the potential risks of a breach posed by tin whiskers or capacitor remains (foil or electrolyte), and the equipment-killing misuse of Y1 class (or series pair of Y2) capacitors, I'm not really satisfied with them. They're not as safe or reliable as the old linear plug-pack supplies, that much is certain.

Personally, the only way I would qualify an off-line switching supply as "double insulated" would be to have the entire primary side on one PCB and the entire secondary on a separate PCB, or some other form of isolation barrier that can't easily be breached. But what do you think? It would certainly help to find a replacement for electrolytic capacitors (not that it's going to happen any time soon).

Re: Do small SMPSes really qualify as double-insulated?

PostPosted: October 20th, 2014, 4:51 pm
by c_hegge
I'd say they are fine if they don't use Y caps on them. The other potential failure points (tin whiskers and capacitor foil) exist on just about all power supplies - double insulated or not.

Re: Do small SMPSes really qualify as double-insulated?

PostPosted: October 20th, 2014, 7:47 pm
by LongRunner
But it's the isolation barrier at stake. Shorting across that could very well have disastrous consequences. In linear supplies the isolation barrier was in just the transformer, and was virtually impossible to breach outside of overheating (which a thermal fuse guards against) or a direct lightning strike (which will destroy any insulation, anyway).

Anyway, at which point can you consider the insulation "safe enough"? 1 failure per billion units? 1 per 10 billion? I don't see any current SMPS being that safe.

Re: Do small SMPSes really qualify as double-insulated?

PostPosted: October 20th, 2014, 8:33 pm
by c_hegge
Tin whiskers can still form in a linear PSU (especially if RoHS solder is used), and caps can still explode in a linear PSU. The only difference is that the physically larger transformer required by a linear PSU requires a larger PCB and thus more space between the two sides, although I guess it is true that simply having more space between the two sides would reduce the likelihood of something shorting the isolation barrier. You could probably work around that by installing a peice of plastic or rubber in between the gap in the PCB and the case, though, or by covering the solder side of the PCB (at least the primary side) with silicone or rubber.

Re: Do small SMPSes really qualify as double-insulated?

PostPosted: October 20th, 2014, 9:25 pm
by LongRunner
That's if the whole linear PSU is built on a PCB, of course.

Further up in the article linked, you can see that the example AC-output supply has no PCB at all. The DC-output version shown has just a small PCB with the diodes and capacitor. Unless integrated into the device it powers, building the whole linear PSU on a PCB would be a needless expense.

Besides, if there's a single (plausible) failure that can breach the isolation, then how can you continue to call it double insulated? :s