Getting AC from a laptop CCFL inverter

Discussion in 'Power Electronics' started by seanspotatobusiness, Dec 18, 2017.

  1. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    I would like to power a neon indicator lamp which I heard can be done using a laptop CCFL inverter. The problem is that these inverters output DC which means that only one filament in the indicator lights up. I was wondering, can I just bypass the capacitor/resistor at the end and connect directly to the transformer? Can I limit the current? I think if I just put a resistor there, the inverter will increase the voltage to force the same amount of current through.

    The image below is just a random picture I found and not the design I will necessarily use (I don't know exactly what example I will use yet).

    Edit: I tried again and this time both filaments illuminated even when only one end of the lamp was connected. I think I can use the brightness adjust pin to reduce the current applied to the lamp.


    [​IMG]
     
    Last edited: Dec 18, 2017
  2. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,755
    2,021
  3. ian field

    AAC Fanatic!

    Oct 27, 2012
    6,217
    1,118
    The current is sort of limited anyway - typically CCFL tubes need about 1200V to strike, and ionise at around half that.
     
  4. Dodgydave

    AAC Fanatic!

    Jun 22, 2012
    7,587
    1,250
    Put the neon across the output of the transformer with a series resistor between 220 - 470K to adjust the brightness.
     
  5. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    The CCFL tubes have a thick wire and a thin wire. I read that the thick wire has more insulation because it carries the high voltage whilst the thin wire is the ground. If the output is AC then this cannot be correct, right?
     
  6. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,755
    2,021
    The CCFLs shown here use AC and don't have thick and thin wires.
    It can be correct for AC or DC. If one terminal of a supply is at ground potential then it doesn't need to be insulated from anything else at ground potential; however, the other terminal does need to be insulated.
     
    Last edited: Dec 19, 2017
  7. ian field

    AAC Fanatic!

    Oct 27, 2012
    6,217
    1,118
    So far; I've never seen assymetric insulation on CCFL leads - but I'm not claiming there's no such thing.
     
  8. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Thanks for everyone's responses. I have designed this circuit which reroutes the output of the inverter to a resistor when the lamps are not lit to prevent the inverter producing a very high voltage when the lamps are not lit. I actually have two circuit designs but I don't know whether one is better than the other, other than the first using fewer components.

    [​IMG]
     
    Last edited: Dec 20, 2017
  9. tcmtech

    Distinguished Member

    Nov 4, 2013
    2,809
    2,703
    I'm pretty sure your NPN transistors are not going to do anything good being fed AC power.

    As for snubbing your extra voltage when you're indicators are off you could just stack two neon bulbs in series to work as a voltage snubber that is double the other neons rung voltage.
     
  10. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Thanks for your response!

    What about this change with an inverted transistor to allow the current in the reverse direction?

    [​IMG]

    Would my solution to snub extra voltage when the indicators are off not work (when modified as above to allow the reverse current)?
     
  11. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,755
    2,021
    That arrangement won't work :(. The base-emitter junction of a bjt has a reverse breakdown voltage of only around 5 or 6 Volts.
    I second tcmtech's suggestion to use a couple of neons stacked to act as a voltage clamp. That would also make selection of any switching device less critical.
     
  12. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Behold! Edit: MOC3010 should be maybe MOC3083.

    [​IMG]
     
    Last edited: Dec 20, 2017
  13. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,755
    2,021
    Now we're cooking!
    The first circuit will have problems because the NPN can't control in both half-cycles.
    The second circuit looks more promising but would need a dropper resistor between the transformer and the stacked neons.
    The MOC3010 has an off-state voltage of only 250V. It also seems to be non-stocked by major suppliers so may well be obsolete.
    I don't know what frequency your CCFL inverter operates at, but LTspice is telling me some neons may not like operating at a frequency much higher than mains frequency.
     
  14. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    This version has some changes. The MOC3083 withstands up to 800 V and can close/activate with maybe as little as 5 mA.

    [​IMG]
     
  15. tcmtech

    Distinguished Member

    Nov 4, 2013
    2,809
    2,703
    The opto triac systems could work in theory, however given the high frequency the tiny inverter output operates at there is a good chance it still wont work being most common opto-triacs are designed to work at power line frequencies to maybe a few hundred Hz at best.

    As for the neons, they won't care what frequency the power is at.
     
  16. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,755
    2,021
    The NPN in the latest circuit still can't control the triac in both half-cycles. Also, 5V and 1meg will allow only 5uA through the triac gate.
    Maybe, maybe not. The only Spice model I have implies there is a ~100uS time constant involved in the discharge process, but I've no idea which particular neon the model relates to or if the model is correct.

    If it were me I think I'd just strip the guts from a retro-style panel-mount neon indicator and replace them with a LED :D.
     
    Last edited: Dec 20, 2017
  17. tcmtech

    Distinguished Member

    Nov 4, 2013
    2,809
    2,703
    Neon bulbs will light up at RF frequencies so that says they can take frequencies well beyond what a CCFL inverter puts out. I've never seen a real life neon bulb of any design that wasn't RF sensitive.

    It's also why I don't trust simulator programs one bit. They're too limited on determining what reality can do Vs what they say reality cant do even when it easily can.
     
  18. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,755
    2,021
    I had a feeling you would say that, Tcmtech :).
     
  19. ian field

    AAC Fanatic!

    Oct 27, 2012
    6,217
    1,118
    There was a fad for sticking a neon on the end of CB whips.
     
  20. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Is this because the gate voltage for the triac should be alternating also?
     
Loading...