Ian Stirling said:
No, with the input capacitor.
If disconnected at the wrong time, it'll be charged to about 300V.
It won't then discharge on its own.
If someone then touches the plug, a pulse of up to 3A will bite them.
For the faint of heart, this may be a bad idea.
If you read again you will see I do recommend a resistor across the
capacitor or mains input. It gives a small nip in it's present form,
but the discharge resistor is definitely the way to go. I'm surprised I
forgot to include it in the PCB design after having used a similar
circuit many times before.
There should also be a resistor between the zener and the LEDs, so that
the current is smoother, and there is less 50Hz flicker.
Picking the zener and the resistor so that they tend to regulate the
current at the right number is a good idea.
I used a series resistor in an earlier design, but when the circuit is
used without it there is absolutely no 100Hz flicker whatsoever (Even
when the light is "swiped" to see the strobing). The LEDs have a slight
voltage variation depending on their operating current, and it seems to
work despite being a bit naughty. It's easy enough to add, but the
whole concept of the design was a minimalist device. Maybe I'll add it
in to instil good design principles.
The cheap bright LEDs have terrible life even at nominal current.
Feeding them with pulsating current to the same brightness is a
bad idea, as is exceeding their ratings, or allowing them to get hot.
The cheap bright LEDs are indeed crap. The circuit passes a mean
current of 20mA through the LEDs which is within their designed
dissipation.
The 100/1000 (I'd use 1000) ohm resistor should be rated to handle the
A look back at earlier versions reveals that I used a 100 ohm resistor
in series with the LEDs in the first version but found it's function to
be virtually irrelevant. A look at two commercial devices that use this
technique reveals that one has no resistor and one has a 47 ohm
resistor.
surge voltage, and have an appropriate wattage.
Also, consider what happens when someone connects this to a dimmed outlet,
along with a spotlight or other load.
Not going to be pretty.
When run on a dimmer the brightness doesn't vary much once the point of
illumination is reached. I did test an earlier version in this manner
with the intent of destruction. Sadly it failed to do anything
exciting.
At the least, I'd include a thermal fuse, use an appropriate capacitor
(class Y?), due to what happens if the input cap shorts.
The capacitor is an X2 component and if you consider that the whole
circuit is basically a snubber circuit consisting of a resistor in
series with a suppression capacitor, then it's a very stable circuit
that is used in it's more primitive form in dimmers and appliances
around your house.
In the event of catastrophic failure of the capacitor, then the resistor
will open the circuit with a puff of smoke but little else. A typical
1W resistor won't hold 250W for more than a fraction of a second and
won't heat adjacent materials to anywhere near combustion point in that
time. That said, I suppose there's no harm in decreasing the power
rating of the resistor to half watt or even quarter watt, although
quarter is a bit low for a snubber type circuit.
Don't forget the 470K 1W resistor across the input.
I recommend two 1/4W 1M resistors in series for the benefit of the
increased voltage rating.
