Warning: Laptop inverters put out high voltage, usually between 500V and
700V, so don't try these tests unless you know your way around electricity.
I'd also strongly advise against my sloppy approach of twisting wires and
not even taping them.Since I recently did a page on testing
a laptop inverter with a multimeter, I thought I'd go over some of the
testing options for a backlight. Laptops employ CCFL tubes for backlights,
due to their bright white light at relatively low power. Like all fluorescent
lamps, CCFL tubes require a high voltage, high frequency input to strike
a plasma and cause the tube coating to fluoresce. Someday not too far into
the future, white LEDs should be available to do the backlight job with even
less power and higher reliability. But in the meantime, a pairing an inverter
with a CCFL is the way to light up a laptop screen.
The picture above shows a 4" blue CCFL tube from a PC modding kit that cost
around $3 by mail order. In fact, I think the shipping cost more than the
whole kit, which included a dual inverter and the tube. Instead of using
the PC connector provided with the kit, I cut the leads off and powered the
inverter with 4 AA batteries in an 8 cell Radio Shack holder that cost $1.89.
Since it was an 8 cell holder and I only wanted around 5V, I had to jumper
the across the top, which I did by just sticking a piece of wire in the
connectors. I also cut the output connector of the inverter since it was
different from the connector laptop backlight I wanted to test. Then I remembered
I hadn't tested the original modding backlight first to prove the inverter
was functioning, which explains the first twist together job.
The picture above shows the modding inverter hooked up to the LCD backlight.
I've folded up the white flap that covers backlight to help keep the light
in the LCD assembly. The inverter did fire up the plasma, but not completely,
which left me wondering if the CCFL tube was bad after all. However, it turned
out that the inverter output was essentially linear with the input, as my
neighbor with a variable linear power supply was able to demonstrate. I came
back and put 8 cells in my battery holder and it fired the whole tube up.
Unfortunately, I got the other four batteries by taking them out of my camera,
so a picture was out of the question:-) Next I decided to try to fire up
the backlight with the original inverter from the laptop. This inverter looked
pretty standard and I guessed it would be happy on 5V, I know some laptop
inverters take a 12 V input but I figured the lower voltage wouldn't hurt.
I'm simply touching postive lead of around 5V DC to the in-board side of
the fuse, and you can see above that the backlight fired up at full brilliancy,
even with the camera flash. I finally zoomed out so you can see that with
the correct inverter, the whole backlight is lit up, and the light is in
fact spread through the LCD substrate and visible at all the edges, as it
should be. I had to take the metal back off the LCD to expose the backlight
this way, which means that all the intermediate filters and layers wanted
to fall out of the LCD if it was tilted. But it's much easier to tell what's
going on from the back than the front, because in the "off" state, an LCD
doesn't transmit much light, all the little crystal cells remain twisted
closed. If you move the whole screen to a shaded area and fire up the backlight,
the screen will visibly light a little, a sort of dull grey that's hard to
capture with a digital camera.
To the right you see the LCD with the backlight lit full on, shot in the
dark. You'll notice that there's actually more light leaking out the back
of the LCD than transmitting through the screen. The lighting gradient you
see is an artifact of the camera angle, the screen was even lit with a dull
grey grid. Keep in mind that an LCD displays white by turning on the red,
green and blue subpixels, which mix the light together for white. But the
fact that some light leaks through when the LCD doesn't have any video input
implies that for real black, the liquid crystals need to be twisted hard
into an opaque state, their unpowered "off" mode allows some light to leak