Why choose an LED bulb
My previous post describes some of the considerations involved in using LED bulbs, with some detail on a G4 halogen replacement LED. I got interested in this when searching online for some LEDs for a new light fitting I’d bought. This was designed for six 20W 12v halogen bulbs. Reviews such as ‘It’s a pity this only takes 20W bulbs’ made me think that it probably wasn’t all that bright, even though it would use 120 watts. Naturally, I thought I’d fit the brightest LED bulbs that I could get and after some time trawling the internet, it seemed that 6w bulbs were the highest power available in G4 fittings. I ordered 10 cool white dimmable.
Needless to say, I’m wary of stuff from China – they always manage to find a way of exaggerating the specification, so when the bulbs arrived, I felt I should investigate a bit more. This titchy bulb contains some quite sophisticated electronics.
In essence the base contains the circuitry on both sides of a flat board and the top part contains six small LED chips mounted on a rectangular glass substrate with yellow phosphor on both sides of the glass. It is all sealed in a soft, clear plastic resin, so care is needed not to bend it, otherwise the internals will be damaged.
Looking at the electronics board, on one side we have what is essentially the input circuitry – the big yellow rectangular blob is a smoothing capacitor and I assume that the rectangular things are diodes for a bridge rectifier plus some resistors. Together they create a nice smooth 12 v DC input from the 12 v AC that they are supplied with. This is fed to the circuitry on the other side.
Although I can’t read the designations on the small chip in the middle, it is something like the ZXLD1360 LED driver shown in the circuit diagram below.
In essence, it is a current/voltage regulator that reduces the 12 volt input to that needed to drive the LEDs. Since each of the LED chips will need a forward voltage of around 3.4 volts and the supply is nominally 12 volts, I am rather assuming that the six chips are connected in parallel, as the connecting wires are too small for me to see.
How does it work
This chip is in essence a current regulator. The current to the chain of LEDS is sensed from the potential difference across a resistor (shown as Rs in the circuit diagram) The circuitry in the chip pulses the main power transistor (MN in the circuit diagram) on and off at high frequency (kilohertz) to maintain the current at the design level.
The small choke L1 acts to smooth the current curve and also improve the power efficiency of the circuit. When the main transistor MN switches off, the magnetic field in the choke collapses generating a reverse potential across the choke. This feeds back to the input of the circuit through the Schottky diode [don’t confuse the ‘S’ symbol on the cathode with the ‘Z’ symbol of a Zener diode: Schottky diodes have about half the forward voltage drop of an ordinary silicon diode] and thus light the LEDs on the reverse half-cycle.
Since there are six LEDs in the bulb, then each must consume 1 watt if it is a six watt lamp, If the lamps are in parallel, then with a voltage of 3.3 v across each one, the current must be 0.3 amps, requiring 1.8 amps for all six lamps. But the circuit can only provide 1 amp. I suppose they could be connected three in series (as shown in the circuit diagram) and both of the three in series being connected in parallel. It is best if they are all connected in series, as this ensures that the same current flows through each, which evens up their brightness. Sadly, I think it is more likely that they will claim that the ‘W’ doesn’t mean watts at all, but perhaps ‘way’ in other words it has six diodes in it. I suspect the diodes are not likely to take more than about 100 mA each, if that, so it probably consumes 0.3 watts per diode and maybe 1.8 watts all told. In other words, quite a lot less bright than the 20 w halogen. I will measure this when I get access to the 12 v transformer.
Moreover, I’m pretty sure that this will not work with a dimmer – in fact it does its best to maintain the LED current whatever the input voltage, so will resist attempts to dim it. The chip does actually have a separate dimmer input to it, but this needs an external connection, which obviously you don’t have in an ordinary lampholder. However it is not impossible that this is connected to detect changes in the supply voltage and thereby regulate the brightness.
On this point, it is worth noting that you can’t adjust the brightness of an LED by altering the driver voltage. The diode will only conduct (light up) when the input voltage reaches the internal voltage drop of the diode. If you try to turn the voltage higher, the voltage will remain at the internal voltage drop, but the current will increase and the LED will be brighter. But if you increase the current too much the LED will overheat and burn out. The usual way of controlling the brightness is to pulse the LED on and off. If this is done at a high frequency, the light will appear less bright and the flicker will not be perceived. If the pulse rate is too slow, the flicker can become perceptible and cause problems for many people.
So you live and learn. to be honest, the bulbs were remarkably cheap, so I couldn’t have expected more and it is an interesting learning experience.
240v G4 LEDs
As a footnote, I bought some 240 v LEDs as well. These just have a ‘capacitive dropper’ to regulate the input voltage and these probably would work with a dimmer.