How does Smartcharge bulbs work?!

[iDeeW]

Hi,
Anyone know how does Smartcharge bulbs work?!

How does it detects the switch position when there is a blackout!!!??

Thanks

[codeboy2k]

I haven't given it much more than 1 second of thought here, but when there IS power, the AC leads coming into the light bulb will have a voltage, so the light knows that AC is on.  It will use the AC power from the grid to power the LED and charge the battery inside. 

When the switch is turned off (or a blackout occurs), the LED will see the AC power go off, and it will not know if it is a real blackout (no power) or the user just turned off the switch and WANTS the light off.  To determine which case it is, the light bulb electronics will need to inject a small (possibly AC?) current into one side of the AC line and see if it can detect that current on the other line side.  If it can detect the current on the other side of the line, then the AC power is off (blackout) but the switch is ON (closed).  So the LED will come on.   If it cannot detect that current, then the switch is OFF (blackout or not, doesn't matter, the LED light will stay off). It probably does this check every couple hundred milliseconds or so while the AC power is OFF (pulse check)

It can do this because the 2 wires from the light bulb through the switch are connected via the transformer at the house or the pole/utility box.  So it can tell if the switch is open or closed. See the diagram below.

Pretty ingenious, actually.  And could indeed be one of the few successful kickstarters.

(noted on their kickstarter page, it doesn't work with multiple bulbs on the same circuit.  They don't explain why, but given my supposition above, I suspect that it cannot detect the switch open/closed position because the filament of the other bulbs on the same circuit interfere and complete the circuit ahead of the switch -- second drawing below)

[amyk]

I suppose that even in the case of multiple (filament) lamps in parallel, it could be made to work if it's able to measure changes in resistance accurately enough, possibly based on calibrating to the circuit's impedance while it's being supplied AC. That would be far too complex though.

Incidentally, there's a thread about (dis)continuity testing nearby at the moment.

[iDeeW]

Thanks for answering.
But if you look carefully (@1.38) he simulate power outage by unplugging the power bar. That's an open circuit (no transformers). But the bulb still detects the switch.
I still don't understand this.

Cheers. 

[codeboy2k]

Yeah, don't know what he's doing with that simulated event. 
There's also a second power strip off-screen that the black table lamp and the power strip on the table is plugged into (you can just barely see it at around 1:16).

So he unplugs (then re-plugs)  the second power strip (which is the supply to the black table lamp and the power strip on the table top supplying the LED lamp).

So we don't know, there might be an isolation transformer in the circuit, or there might be a line-filter in the second power strip (which will have capacitors across the line, thus completing the circuit for an AC pulse)...

There's no way it can detect the switch position without something closing the loop

[sleemanj]

Perhaps this effect:



may lead one to some ideas.

If the power goes totally out, this "phantom power" for want of a better word won't be there any more, so detect the absence and you know the power is gone.

Of course, this effect may be house wiring dependant?

[amyk]

There's no way it can detect the switch position without something closing the loop

That can be done with TDR, but seems a bit too complex for this product. Then again, many Ethernet ICs have one built in, which can detect how long the cable is...

[amyk]

A year later, the Chinese have come up with their version, and the riddle is solved thanks to bigclivedotcom:



The circuitry is extremely simple (and very sensitive), it's just detecting if current can flow out of the lamp. A few uA seem to be enough. When mains is present, it acts against a diode to stop that current flow, and when mains is not present but the circuit is open, the current doesn't flow either. It seems multiple of these could be wired in parallel with the same polarity and they'd all switch simultaneously, as long as their "sense voltages" are the same. I like how they've reused the current-limited LED driver to provide a limited charge current to the cell, so all they need is a voltage regulator to limit the charge voltage.

Yeah, don't know what he's doing with that simulated event. 
There's also a second power strip off-screen that the black table lamp and the power strip on the table is plugged into (you can just barely see it at around 1:16).

So he unplugs (then re-plugs)  the second power strip (which is the supply to the black table lamp and the power strip on the table top supplying the LED lamp).

So we don't know, there might be an isolation transformer in the circuit, or there might be a line-filter in the second power strip (which will have capacitors across the line, thus completing the circuit for an AC pulse)...

There's no way it can detect the switch position without something closing the loop

And that "something closing the loop" would be the resistance across the line of the second traditional table lamp.