CCFL or Inverter?

I have a Gateway Solo 1450 laptop that presenting very dim, and to my
eye, reddish cast, video. Also, there is a bit more flicker in it than
there out to be. Hooking up an external monitor produces bright, clean,
stable video.

Any bets as to if this is a problem with the CCFL(s), inverter, or both?

Thanks for your speculations.

Grinder wrote:
I have a Gateway Solo 1450 laptop that presenting very dim, and to my
eye, reddish cast, video. Also, there is a bit more flicker in it than
there out to be. Hooking up an external monitor produces bright, clean,
stable video.

Any bets as to if this is a problem with the CCFL(s), inverter, or both?

Thanks for your speculations.

If it was a reddish or brownish cast, and still moderately bright, it
would be the CCFL tube itself.

When the light level is weak and discolored, it could be one or both.
If it didn't light up some times (flaky), you'd suspect that was the fault
of the inverter.

When the characteristics of the CCFL tube change enough, that can also cause
a perfectly working inverter to enter protection mode.

Since inverters seem to be the most unreliable part of that subsystem,
I'd start there.

*******

To measure the actual inverter voltage safely, isn't exactly easy.
The tube seems to be capacitively coupled to start with. They
don't use a DC connection. Perhaps the purpose is to prevent any
DC at all from appearing on the tube, to preserve the metal on
the electrodes inside. To get an accurate reading, you
might need to probe on the surface of the inverter, on the
inverter side of the coupling caps.

(See PDF page 8 here. CY1 and CY2 are examples of coupling caps on an inverter.
They're 27pF 2kV rated. Probing on the pin 1 side of the cap, would give
you a reading from the transformer.)

http://www.diodes.com/_files/product...005_AP2001.pdf

The thing is, your multimeter might have a 1000VAC rating, just
barely able to cope with the voltages from the inverter. But
worst case, the inverter may be able to product 4000VAC unloaded,
like if the connection to the tube went open circuit for a moment,
and that would damage your meter. Some inverters can be
damaged if run open circuit (voltage goes too high, piezo element cracks).
As a consequence, if I had to do it, I'd probably want to borrow a
100:1 probe or the like. Picture of one here. I think over the
years, I've seen two of these, one in the physics lab in university,
another one at work for ESD testing.

http://www.qsl.net/kh6grt/page2/test...nt/hvprobe.jpg

In the thread here, one of the participants notes that the inverter
is very load sensitive, and will enter protection mode at the drop
of a hat. Consequently, the design of the probe is quite critical.
The size of the coupling caps, from inverter to CCFL is in the pF
range, and any probe can't present more capacitive load than that.
So both the R and the C value of the probe can be important. The
C value would not be purposeful, but would be a leakage capacitance.

http://www.eevblog.com/forum/general...af6328ef45c0be

They have an example here of how such probes are designed. You can
easily buy 22 megohm resistors, and put a whole bunch of them together
to build a larger resistance. That's what I used to use, for some
high voltage stuff I used to fool around with. If you were working
on really high voltages, such a circuit would need corona protection
(no rough metal edges, corona dope on exposed surfaces, that sort of
thing).

http://www.repairfaq.org/sam/hvprobe.htm

You might not have noticed, but I HATE high voltages. I've only
been thrown across the room once, by high voltage DC, and it left
a lasting impression. When you're on the floor, you can't remember
how you got there... :-) Once you look around the room, and see what
you were working on, then you can figure it out.

To their credit, the inverter making the 700-1000VAC runs at
25KHz. The higher the frequency, the more the current flows over
the surface of your skin. That is what prevents an inverter, from
having the truly nasty wallop DC can have. A danger with high
frequency AC, is it can burn you (burn a hole in your skin),
so even though you might not fly across the room, it can still
damage you. My high voltage device (flyback with voltage multiplier)
could light paper on fire, to give an example of the burning
potential. (The flyback isn't the circuit that threw me
across the room. That was something else I was working on.)
But that flyback thing drew around 36W on input, so about
10x more power was involved, than on an inverter.

Um, have fun ? And make sure there is nothing breakable
behind you, when you rocket across the room :-) I fell on
the cement floor in my basement. I'm just happy my head
didn't hit the floor.

Paul

Thanks, Paul.

Grinder wrote:
Thanks, Paul.

Well, you're still alive. That's a good sign :-)

Paul

speaking about inverters and lamps, i have a faulty notebook with
strange behaviour,

the owner had thought that it's a simple failure like inverter or CCFL
lamp, but i have checked three adittional inverters(two new ones
ordered from USA),
and the same thing. I have attached new screens and a screen from the
same model(with it's own inverter + flat cable), but nothing, the
notebook is showing a symptom of CCFL failure,
the image is showing up but dimed. So the inverter is not, the CCFL
is not a failure, the picture is good-normal when attaching it to an
external LCD monitor!
Then it has to be that chip on the motherboard which converts the
picture to the screen ???

that chip - anybody knows what's the name of that chip in general
terms ?

or somekind of short-circuit somewhere on the board. I have had that
kind of problems, and a very well educated and practiced electrician
can solve that in a matter of a sec.
Is there some possibile solution i could handle without knowing the
electrical stuff as a profession ??

pardon, it's a voltage regulator chip on the motherboard, perhaps.

bymyfault wrote:

speaking about inverters and lamps, i have a faulty notebook with
strange behaviour,

the owner had thought that it's a simple failure like inverter or CCFL
lamp, but i have checked three adittional inverters(two new ones
ordered from USA),
and the same thing. I have attached new screens and a screen from the
same model(with it's own inverter + flat cable), but nothing, the
notebook is showing a symptom of CCFL failure,
the image is showing up but dimed. So the inverter is not, the CCFL
is not a failure, the picture is good-normal when attaching it to an
external LCD monitor!
Then it has to be that chip on the motherboard which converts the
picture to the screen ???

that chip - anybody knows what's the name of that chip in general
terms ?

or somekind of short-circuit somewhere on the board. I have had that
kind of problems, and a very well educated and practiced electrician
can solve that in a matter of a sec.
Is there some possibile solution i could handle without knowing the
electrical stuff as a profession ??


The inverter will have a "contrast" input, coming from the motherboard.
It should be one of the few control signals feeding the inverter.

On really old inverters, the contrast control is a simple analog voltage.
As the voltage varies, the light output of the CCFL tube varies. This
is terms an "analog control" method. One way to determine an analog
control is used, is that analog control does not give a very large
adjustment range. It might vary the light level from 70% to 100%, but
not offer any lower adjustment (as the CCFL tube would turn off if you
tried a lower value).

*******

The more modern control method, is PWM based (pulse width modulation).
A digital square wave is sent on the cable. It would look like this,
at perhaps 200Hz or so.

+---+ +---+ +---+ +-+ +-+
| | | | | | | | | |
---+ +-+ +-+ +- ---+ +----+ +----

High duty cycle = bright Low duty cycle = dim

The PWM signal "modulates" the 25KHz oscillation of the inverter.
The inverter runs in "bursts". Since, on average, the CCFL tube
is maintained in the conducting state, it continues to run with the
lower of the output voltages that the inverter offers (like 700VAC).
If the duty cycle were to go so low as to quench the tube, then
the tube would likely not be able to light up again. So the
duty cycle can be varied over a large range. And the adjustment
range using PWM, is larger than with the analog control method.

If you are seeing a dim output, you'd check to see if the PWM
signal was still working. The PWM frequency has to be higher
than the scan rate of the panel, and preferably a frequency
unrelated. Otherwise, there could be a "beat note" or distracting
visual interference, between the CCFL light source, and the
image provided by the LCD panel itself. What I can't tell
you, is what the correct amplitude of the control signal
should be. Perhaps the amplitude is wrong, instead of just
the pulse width. A wrong amplitude might affect the
operation in unpredictable ways.

This picture, shows the CCFL lamp voltage present, during one
of the pulses in the PWM waveform above. So bursts of 700-1000VAC
sine waves are fed to the CCFL, to control the intensity.

http://www.maxim-ic.com/images/appno.../3997Fig02.gif

And I don't know, what chip on the motherboard, makes that
PWM signal. It would be logical for it to come from the
laptop GPU (graphics processor), but could just as easily
come from some other chip.

You can see in this example, there are five signals feeding
into the inverter (on the ribbon cable). +5V for the logic
gates of the chip. +12V to the MOSFETs pumping the transformer.
Ground signal. And two control signals - an ENABLE signal
to turn the lamp completely ON and OFF, as well as LPWM
to control the intensity while it is ON. Presumably the
allowed logic level on LPWM, is a 5V amplitude signal.

http://circuits.datasheetdir.com/59/OB3302-circuits.jpg

Some inverters will have more signals than that. To
understand inverter operation, you would note the part
number of the IC on the inverter, then dig up a datasheet,
to get a better idea what signals could be on the ribbon cable.

*******

As a "non-electrician", all you can do is verify the ribbon
cable is properly plugged in.

Paul