Russell McMahon wrote:

>
> The "direct offline" capacitor fed power supply is especially=20
> susceptible to
> noise spikes.
> At high frequencies the 0.33uF capacitor is very low=20
> impedance and noise
> sees mainly the 220 ohm resistor. The 470 uF capacitor is probably an
> electrolytic and will have an ESR of around 0.1 to 1 ohm. This will
> attenuate noise spikes considerably BUT less so at higher=20
> frequencies due to
> lead inductance and any wiring will add to this.
>=20
> The output ground lead and the input diode should both=20
> terminate on the
> capacitor and it would help to add a 0.1uF ceramic with short=20
> leads across
> the 470 uF. Having a series resistor with a second=20
> electrolytic after it to
> provide a second stage of filtering also helps. We used to=20
> use supplies
> derived from Telecoms 50v battery (very noisy) long ago using=20
> pure resistive
> dividers - using two stages of zeners with a resistor between=20
> them helped a
> lot.
>=20
> Not having the 240V lead in close proximity to the 5v supply will help
> although, for noise spikes, capacitive coupling will be=20
> relatively small
> compared to the direct path via the 0.33 uF cap.
>=20
> As you note - this is a highly dangerous power supply and=20
> must be regarded
> as being at 240 VAC at all times.

Thanks Russell, for your insight and tips.

The 0.1uF ceramic cap that you are suggesting is already there in the
circuit, soldered across the 470uF electrolytic cap. I tried a few minor
layout changes like bypassing certain tracks with short lengths of wire =
and
increasing the value of the elec. cap to 1000uF, but they didn't seem to =
take
me closer to the solution.

The right treatment can be given only when the problem is not correctly
diagnosed! =20

Regards,
Bala


=20
=20

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics