Traveling, stayed in a large international chain motel. The bathroom had a nightlight with a CdS (cadmium sulfide) light sensor, so of course I opened it up. First thing to notice - no high-voltage components. None. Zero. Big fat film capacitor as a voltage dropper? Nope. Hmmm . . .

The light comes from four white LEDs in series, and they act as a shunt regulator for the control circuit. When the ambient light is low, R4 is almost an open circuit, Q1 is turned off by R3, and the LED string limits the max voltage to around 14 V (assuming 3.5 Vf per LED). When the ambient is high, R4 turns on Q1, which shunts current away from the LEDs to turn them off.

Not the most efficient circuit, but the actual part for Q1 is an S8050 that is rated for only 25 V Vce. AND while there are pc traces and holes for thru-hole parts, R1 and R2 are SMT 0805.

Fun fact: There are zero markings of any kind anywhere on the enclosure. Manufacturer, Voltage rating, UL, part number, inside, outside - nothing.

R4:

7.5 K - normal room light
2.5 K - workbench light
150 K - dark room

ak

 

Somewhat surprised they used two resistors and a full bridge. Maybe a half bridge results in too much flicker.

 

120 or 240V?

 

Not sure it would pass EU regulations for standby power.

 

Not sure it would pass EU regulations for standby power.

Definitely it wouldn’t. And I don’t think it would pass a UL listing either.
It is really a brute-force circuit with simplicity and low cost its primary objectives.
The only hope is that the line-side resistors are of the fusible type. But I would not count on that.

 

Definitely it wouldn’t. And I don’t think it would pass a UL listing either.
It is really a brute-force circuit with simplicity and low cost its primary objectives.
The only hope is that the line-side resistors are of the fusible type. But I would not count on that.

I don't know what the official EU requirements for standby power are, but a quick search indicates that it is 0.5 W in standby/off modes.

Two 27 kΩ resistors across 120 VAC would only consume 267 mW of power. So why wouldn't it meet requirements? Is there a requirement that the 'off' power can't be more than some fraction of the 'on' power?

As for UL listing, I would be surprised to, though it might depend on what kind of packaging was used. But given that there were no markings of any kind at all, I would be very doubtful that it would make muster.

 

Big Clive on YouTube has analyzed a large number of LED nightlights. For the most part, they all draw MORE current when off than on – they short out the LEDs when "off".

 

Two 27 kΩ resistors across 120 VAC would only consume 267 mW of power. So why wouldn't it meet requirements?

You are correct that for North American voltages it would not exceed the limit.
However for EU regulations expecting EU level voltages, most likely it would not, or barely meet them.

 

120 volts of pure amurcan power.

The two 27K resistors look like standard 0805 types. I doubt that the layout meets UL creepage requirements.

I don't think standby power rules apply, because the unit does not have a standby mode. No button, no indicator, etc.

ak

 

The two 27K resistors look like standard 0805

Depending on what company makes the resistor but likely the resistor is rated for 150V, each.
This type of product does not have to worry about Line to human spacing because the packaging must be double insulated.

I doubt that the layout meets UL creepage requirements.

The voltage across the Bridge, transistor and LEDs is about 6V to 12V. Depending on what LEDs. There is no real problem with spacing. The 1/2 line voltage across R1 and R2 is the only hot points. If the resistor maker declares the resistor and pads are good to 150V working, and 300V over voltage than UL will believe them. You have Line to Neutral (or L to L) spacing as power enters the board. After R1, R2 the voltages are very low from trace to trace.

I don't see anything wrong here. I probably would want to use fusible resistors for R1,2 but some accountant/manufacturing engineer will see that I wasted 1 cent and have a fit.
Again, I only see problems with the plastic case, which much of a problem.

 

Big Clive on YouTube has analyzed a large number of LED nightlights.

Link?

 

Here's the first link "nightlight" link I came to on Clive's page. It has a nearly identical circuit as that shown here. He shows the schematic starting at about 5:30, and explains the "off" mode starting around 7:00.

If you haven't seen Clive's videos before, he often disects devices and explains the circuit operation in great detail.

 

TS: Do you have a question about this?

 

TS: Do you have a question about this?

Nope. Just sharing what i saw as a different approach to the idea.

ak

 

I often use a capacitor to reduce the voltage. The cap cost more than a resistor.

 

I often use a capacitor to reduce the voltage. The cap cost more than a resistor.

That would pass the ERP regulations, as the output when shorted out would be almost zero power factor. Current flow, but no power!

 

Given that the main purpose of switching a night light OFF is to avoid wasting power, the circuit in post #1 achieves exactly the opposite.

 

Agree. The concept dates back to extending the lifetime of a small incandescent bulb, but it really doesn't apply to an LED fixture.

Still, the idea of using the LEDs as a shunt regulator is kinda cute.

ak

 

Given that the main purpose of switching a night light OFF is to avoid wasting power, the circuit in post #1 achieves exactly the opposite.

i don't get it - why is saving power the main purpose?
if the main purpose is to save power, don't use night light at all. and you save 100%.

the way i look at it, main purpose of a night light is convenience. to have some light when it is dark, just enough so you are not stumbling in the dark and low enough not to disturb others like spouse.

120VAC^2/(2*27k) = 0.267W

the incandescent versions used to be 4-7W.
looks like 16-30x power reduction.

what more saving are needed?

in a year the consumed energy of LED variant is 2.3kWh
so we are talking about $0.25...$0.5 a year.

and yes, this can be made to consume still less power. but at what cost?

 

Agree. The concept dates back to extending the lifetime of a small incandescent bulb, but it really doesn't apply to an LED fixture.

Still, the idea of using the LEDs as a shunt regulator is kinda cute.

ak

yup... i think the circuit is neat, specially considering how low cost it is and how compact it can be, not to mention that it already made massive power savings compared to predecessor.