Mains LED with blocking cap

[Phil Allison]

"John Fields"

I'll probably put something together in the real world today as a
reality check, and when I get the data I'll post it.

** How many LEDs do you have to ruin to be convinced ??



..... Phil

 

[Nico Coesel]

Many
Lack of isolation.
Lack of safety.
No (inrush) current limit.
(You want your LED to emit light, I guess, not to explode?)

Like any diode, an LED can take quite a current overdose.

 

[Phil Allison]

"John Fields"
"Phil Allison"

As many as it takes for me to prove to myself that I'm wrong.

Probably just one.

** I ruined five last evening - all old stock with leads a bit tarnished.

A pair of 3mm red LEDs are now merely 1.2 ohm and 10 ohm resistors after one
dump from a 330nF cap charged to 240 volts. A 5mm red LED is now a 1520 ohm
resistor that gives no light after two dumps.

But a pair of 5m amber LEDs proved tougher - even after multiple dumps
each still produces some light.

BTW: Each application of the charged cap produced a spark similar to
shorting the same cap.

With reverse parallel LEDs, I reckon you will kill two birds with one stone
as the instantaneous voltage across the reverse polarity LED will be so far
over its rated max.


.... Phil

 

[Don Klipstein]

Any problems driving an LED from the mains using a blocking cap as a
constant current source?

There are two to deal with:

1: Inrush current - easy to solve with a series resistor to limit that
to something that does not damage the LED.

2: Reliability of the capacitor and safety of the capacitor failing

The capacitor needs to have a decent AC voltage rating. A DC voltage
rating significantly greater than the peak voltage is not sufficient to
survive prolonged AC, even without any voltage spikes. Things get better
if the capacitor is UL recognized or the like.

One thing that also helps: Have everything enclosed in an enclosure
made of flame retardant plastic or the like, suitable for containing
things going wrong. I would not consider chances of the current
limiting capacitor failing being reduced to zero.

The resistor being used for limiting inrush current should be a
flameproof type, preferably UL recognized or the like.

I would avoid filter capacitors, since they could receive full peak line
voltage plus some of any voltage spikes if the LED(s) fails open.
Electrolytics including tantalums could leak conductive liquid if
overvoltaged. This is a concern especially if things going wrong includes
the current limiting capacitor failing short - even if only temporarily
shorting.

Sounds scary? How about using an LED that gives you enough light at a
current low enough to make practical use of a resistor without a capacitor
for current limiting? There are quite a few LEDs that are plenty bright
for use as indicator lamps at currents anywhere from .2 to 1 milliamp.

 

[Phil Allison]

"Don Klipstein"

There are two to deal with:

1: Inrush current - easy to solve with a series resistor to limit that
to something that does not damage the LED.

2: Reliability of the capacitor and safety of the capacitor failing

The capacitor needs to have a decent AC voltage rating. A DC voltage
rating significantly greater than the peak voltage is not sufficient to
survive prolonged AC, even without any voltage spikes. Things get better
if the capacitor is UL recognized or the like.

** A Y-class cap is the best choice - designed as safety caps for line to
ground use.

Even X-class caps do eventually fail and they mostly fail short with lotsa
smoke.

One thing that also helps: Have everything enclosed in an enclosure
made of flame retardant plastic or the like, suitable for containing
things going wrong.

** One simple thing - make sure that the LED itself is not projecting in
such a way that if it were damaged, metal parts might become exposed to
fingers.


..... Phil