First off this modification will NOT work with all inverters. Your results may vary... This circuit was made in response to a question e-mailed to me. Many people want to connect a Cold Cathode to their IDE activity indicator, but are reluctant to do so because of the circuitry involved. It is really very simple, with the proper precautions. First off you can get the voltage for the LED (Inside the Opto) From three sources. +12volts (With a 560ohm resistor) +5volts (With a 120ohm resistor) Or Directly from the Motherboard's IDE Header (No resistor needed, Positive side) The negative side of the LED can be wired to two places: The motherboard's IDE Header (Negative side) Or to Pin 39, using my modification guide linked here: Drive-independent Activity LEDs -Guide The output of the OPTO drives a NPN power transistor. A TIP120 is shown here. It is a NPN Darlington transistor, which means it reall has two transistors inside. The transistor then drives the Inverter directly The reason not many people have done Cold Cathodes for IDE is due to the nature of a Inverter. Inverters have lots of coils, and capacitors. When switched on, and off constantly the inverter can cause noise on your +12volt line. To combat this we first use a Diode backwards across the voltage to the inverter. (If you wire the diode in a forwards mode, you will just get nothing on the inverter, and most likely fry the diode, and transistor) The Diode will absobe any voltage pulse when the inverter turns off. (Similar to the type used on Relay Coils) The two capacitors filter the voltage coming into the circuit. Both are needed, and both MUST be rated to handle at least 24 volts. Also note that many capacitors are polarized, and need to be installed correctly. Good luck. NOTE::: I have not personaly tested this circuit, but it should work in it's design.
To light up a lamp, an inverter needs to generate a "strike voltage" that sets up the electric field in side the lamp, then it can back down to a lower voltage to drive the lamp. I haven't spent a lot of time considering how long the startup transient lasts in various inverters, but CCFL's aren't "instant-on" like an LED. This could work, but I'm pretty sure that short little pulses that would light an LED visibly will not trigger the CCFL, due to the effect I discuss above. Further investigation would involve building a one-shot to trigger this circuit, setting up pulses of various widths and finding the minimum pulse width that allowed the lamp to light. Not too big a deal.
oh jeez :/ that doesnt sound to nice but what do u mean by noise? does that mean the +12v line can spike? and if i hook up to a curcuit like this http://www.bit-tech.net/article/58 with a ccfl in it would it cause spikes or noise?
One other comment, you show the inverter connected between emitter and ground, you'd get much nearer 12v with it on the collector side. You could extend the on-time then with a capacitor between base & ground.
Yes, you would get closer to 12 volts. But you would have a LOT of amperage to run through the transistor in order to pull the 12 volt line down. (IE Turn off the CC) But that would also lower the voltage on your ENTIRE 12volt line. Yes, you could but you would be introducting twice the signals noise and spikes, due to the coil in the relay.
I didn't mean use the transistor as a crowbar. Using a BD139 instead of the darlington the drop comes down to about 200mV.
