To try and understand this a bit better I made some LED-hacks to try things hands on. (couldnt help it..)
Maybe the findings will help someone else understand too, so here it goes:
I disassembled a couple of the leds I bought, those were bayonet base/#44/47s, VDC ones with one SMD chip (Surface mounted Diode). Since they did not work under the PF in the inserts (where weve now learned there is pulsing DC) either the sockets or the leds were wired "wrong". At least their polarity did not match.
Inside the bulbs there are basically the SMD chip on top and then two poles. On one pole there is only a tiny wire that is bent around the plastic "cap" of the led so it touches the outside of the uhm.. bayonet base. The other pole has a small resistor wired to it. According to my assumption, which i do not know is correct, this should be connected to the + side of the DC in the socket because the resistor will do its thing before the current reaches the SMD chip. I dont know if its there to just reduce the current a bit or if it affects the color emitted from the SMD. (I have a led strip driven off 12V and you can change the color of the lights by turning a knob which reduces the voltage to the strip. Again im not sure but i think it works that way).
The resistor is not there to correct the flicker effect that can be seen when using leds in older machines since such a resistor needs to be mounted parallel to the SMD (across the poles) to draw a bit more current and trick the hardware/software in those machines. Absolutely no expert on that but thats what ive read.
Experiment 1 - Getting the DC bulbs to work for inserts (with bayonet base)
My first experiment was to switch the polarity of the bayonet bulbs. So, i took one apart and just flipped the chip. Did that by desoldering the bottom tip and pull out the resistor wire. Next I soldered the other pole straight down to the bottom tip and bent the resistor leg around the LED plastic cap so it had contact with the bayonet case.
When I tried the bulb it now worked! No flicker whatsoever, at least not that I can see, so it now works with the pulsing DC under the PF. Maybe if one would add a capacitor it would look smoother but i dont know what cap to use so i skipped that part. This is the same thing as when you flip a wedge base LED that does not light up, its just that some soldering is required here to rearrange the wires inside the bayonet base.
Experiment 2 - Getting rid of the strobing in the GI LEDs on top of the PF.
Now, since i bought DC LEDs because I was unexperienced, they worked in the sockets above the PF, the GI sockets. But, there was a horrible strobing from them which is most likely at 50Hz since i live in Sweden where we have that Hertz in the electric system.
So, as described above in the thread the strobing occurs because the LED is a DC one and the current in the sockets is AC. The LED is only lighting up during one of the "waves/phases" of current, the one on the right side of the LED, the + side. Then it goes out when the current comes at the other pole since the LED SMD chip is a diode and wont let current through in both directions. Like a standard diode used at switches etc but this is a light emitting one. This is not visible with a standard incandescent bulb since it will let the current through from both directions and just do its thing, glow.
Had I bought AC LEDs there would not be any strobing. And thats because those AC leds contains a rectifier made from diodes (some fancy small ones probably, or some other fancy stuff I dont know the name of). I wont dig into exactly how a rectifier works but the bottom line is, it will direct the two AC waves towards on point and that point would be the + side of the LED.
I digged up 4 diodes and soldered them together to make a rectifier and then hooked up the LED to it. Its not pretty but it worked! The LED worked in the GI circuit and the strobing was gone! yay!
rectifier.jpg
Experiment 3 - Making a home made Flex LED
While was at it I took apart one bayonet base LED and just soldered some harder wires to it (legs from a standard diode), kept them fairly long and soldered them to the bayonet socket together with some shring tubing to make it look better and to keep it from shorting. Now I had my home made Flex. Not really worth the job if you want flex LEDs but since I have allready bought some cheap ones I could at least try it instead of just letting them sit on the shelf.
homeflex.jpg
Experiment 4 - Converting a frosted warm white LED to a wedge based one
I did not buy any white wedges for my inserts since I just got some samples to start the journey of ledding. However I have a shit load of frosted warm whites that strobes like crazy in the GI so I had to try to convert one to a wedge socket so I can at least make use of some of the strobing frosteds.
Again, i ripped one apart and put it in a wedge socket from another LED I trashed. Worked like a sharm too. The polarity issue was easily overcome since it was a wedge base.
wedgie.jpg
Something else to try is to reduce or cure ghosting in the LEDS. Not sure what to add to make that but maybe some resistor could do it. Reduce the threshold of the LED so it wont light up from the small current that is on the circuit even though standard bulbs are off. Ive read its there to keep the incandescent bulbs slightly warm so they can react faster when its time to switch on. They are not as sensitive as the LEDs and will not glow. The LEDs however will light up from the small "warming" current since they are more sensitive.
Understanding how LEDs work really got a lot easier now that I tried some hacks. Of course its better to buy the right stuff from the beginning, it was fun never the less!
Feel free to correct the stuff i probably misunderstood above