These things have a bridge rectifier straight off the mains going into a fairly large capacitor. Even though the power supply may be disconnected from the mains this capacitor will probably have sufficient charge to give an errant finger the most stimulating surprise of its little life. Take care!
On Tue, 19 Nov 2002 18:59, you wrote: > Zane Gilmore wrote: > > Treat mains with respect it can kill you. ( I've put myself across the > > mains a few times and it *hurts*) > > > > You should probably just use a transformer with a LM317 regulator > > circuit or something. At least this will provide some isolation if > > anything goes wrong. Nick E do you know a good safe cheap bench power > > supply? This is your game isn't it? > > PC power supplies are isolated. Well enough to get UL and CE marking so > there'd be a couple of kV (rms) isolation as a minimum. The actual > requirement depends on internal working voltages... my guess would be > around 2.5-3kV rms. > > Just don't open the box and you'll be fine. UL requires that you can't > fit your fingers into the slots, but the "standard test finger" is > slightly larger than my (very thin) little finger so YMMV. So don't try > sticking your fingers through the cooling slots either :) > > > The other thing with switch mode power supplies is that they tend to > > produce very "noisy" power. So if you're doing analog stuff then it will > > be affected. (though I haven't seen what the later stuff is like I have > > to admit.) > > I'd expect a PC power supply to be pretty noisy because digital circuitry > tends to have pretty decent noise rejection, and the major design goal in > these things is for low cost. The manufactured cost of PC PSUs is pretty > low considering whats in them. > > But I wouldn't be quite so fast to put all switchmode supplies into the > same basket. The telecom market in particular is very demanding on low > electrical noise levels. The limits there depend on the frequency band > but wideband noise is typically less than 10mV for a system which may be > capable of 2,000A (thats two thousand amps) at 48V. Psophometric noise > (the most sensitive band for phone gear, about 300Hz to 3kHz) is > typically required to be under 2mV and some customers are asking for less > than 1mV. > > BTW I think I'd rather design analog stuff using a noisy bench supply. > It'll teach you a lesson or two about supply rejection, so you'll have > one less thing to worry about when you put it in a box with a nice quiet > linear supply :) > > Cheers, > > - Dave > > http://www.digistar.com/~dmann/ -- Sincerely etc., Christopher Sawtell
