Long story short...
100 led string of Christmas lights fed from a 3 x AA cell controller that has 8 different flash patterns. It is a two wire set up with red, yellow, green and blue leds. Red and yellow are connected in parallel with one polarity and blue and green are fitted with opposite polarity. The chain is fed from what must be an H bridge driver with one polarity lighting one colour group and the opposite polarity lighting the other group.
Testing a single red led shows a forward voltage of 2.4 v (begins to glow) and 2.8 volts (20 ma flowing). These forward voltages are very high compared to normal red led's we might buy which have voltages around 1.58 and 1.88v under the same conditions.
So how do these led's work so well in that normally any difference in forward voltage would mean one led would be super bright and other dim. There is no visible resistor within the led but in any case that would not account for the red led needing such a high forward voltage. All 50 (one colour group) all illuminate evenly when a DVM on either diode range or low ohms range is connected to the wires. That small and low current test voltage from the DVM is enough. And all are evenly lit.
Are these specially manufactured led's for this type of usage?
100 led string of Christmas lights fed from a 3 x AA cell controller that has 8 different flash patterns. It is a two wire set up with red, yellow, green and blue leds. Red and yellow are connected in parallel with one polarity and blue and green are fitted with opposite polarity. The chain is fed from what must be an H bridge driver with one polarity lighting one colour group and the opposite polarity lighting the other group.
Testing a single red led shows a forward voltage of 2.4 v (begins to glow) and 2.8 volts (20 ma flowing). These forward voltages are very high compared to normal red led's we might buy which have voltages around 1.58 and 1.88v under the same conditions.
So how do these led's work so well in that normally any difference in forward voltage would mean one led would be super bright and other dim. There is no visible resistor within the led but in any case that would not account for the red led needing such a high forward voltage. All 50 (one colour group) all illuminate evenly when a DVM on either diode range or low ohms range is connected to the wires. That small and low current test voltage from the DVM is enough. And all are evenly lit.
Are these specially manufactured led's for this type of usage?
My theory is that all the LEDs used are probably GaN quantum-well/hetero-junction LEDs - using the same bulk semiconductor will mean the externally visible forward voltages are similar, the quantum well structure tunes the actual light produced since the semiconductor layers in which recombination happens are different. The problem with a red GaN LED is its inherently less efficient than a green or blue one due to the high voltage needed, but clearly this isn't seen as an issue here.
Hetero-junctions can be very quantum efficient (most carrier-recombinations produce a photon that escapes the crystal), so this may even overcome the voltage inefficiency. And of course you can tune the light spectrum. Hetero-junctions also have refractive index different from the bulk semiconductor and this is used to good effect in laser diodes by providing a waveguide for the light rather like optical fibre.
This kind of technology is why blue&white LEDs were worthy of a Nobel prize, BTW, as this was the first application of the technology.
https://www.nobelprize.org/uploads/2018/06/nakamura-lecture.pdf
Actually you get bulk made specials, I have friend who makes aviation tower lights (for high rise buildings to warn planes etc.), and he was able to get 22 V LEDs in white.
What is available in China is a vast variety of different colors, shapes, sizes, brightness, focus, diffused....voltages too.
They use different phosphors and laser trimming to get the output.
There is a factory in Hyderabad called Indus Chip or something, they make 50,000 per hour, mostly white.
The circuit is a simple variable frequency AC drive, the controllers and 5 meter string cost like 3 quid here. Even has a remote...
12V/ 1A SMPS, another quid.
What is available in China is a vast variety of different colors, shapes, sizes, brightness, focus, diffused....voltages too.
They use different phosphors and laser trimming to get the output.
There is a factory in Hyderabad called Indus Chip or something, they make 50,000 per hour, mostly white.
The circuit is a simple variable frequency AC drive, the controllers and 5 meter string cost like 3 quid here. Even has a remote...
12V/ 1A SMPS, another quid.
Some leds have an onboard control chip that is too small to see. I suspect these have voltage sensing switch and the excess Vf is the drop across a tiny fet switch.
I'm out of touch with these developments. The 2.8 volts forward voltage mentioned in the pdf would fit perfectly. So these multicoloured leds are essentially all the same. Visually I would say the green and blue look to have a higher visual output. The efficiency is amazing.
These tree lights are simple but sophisticated. One 8 pin chip with no id on it (some kind of custom processor) and a 32.768kHz crystal. The lights are on for 8 hours and off for 16. Eight different flash patterns available. Variable duty cycle and a floating H bridge output. Works very well.
I'd wondered about something like that but thought it would add to much voltage overhead. There is nothing to see with the led's. I tested a single led from the chain and it seems to have quite a soft 'knee' as it comes into conduction. Much softer than a single red led would have for example.