View Single Post
  #9   Report Post  
Posted to
[email protected] is offline
external usenet poster
Posts: 12,364
Default BenQ G2420HD LCD monitor cannot stay lit up

On Friday, 9 March 2018 18:11:56 UTC, Pimpom wrote:
On 3/9/2018 10:49 PM, tabbypurr wrote:
On Friday, 9 March 2018 15:44:38 UTC, Pimpom wrote:
On 3/9/2018 8:20 PM, John-Del wrote:
On Friday, March 9, 2018 at 8:44:56 AM UTC-5, Pimpom wrote:

I have a BenQ G2420HD monitor that was damaged during
transportation. Two of the four CCFL tubes were broken and I've
solved that problem but it still has a fault. These are the symptoms:

1. When switched on, it lights up and presents a normal display
FOR ONE SECOND, then goes dark for 3 seconds, lights up again for
one second and then goes dark permanently. Power cycling repeats
the same behavior over and over.

2. It detects the video and sync inputs. The pilot light stays
green as long as there's a signal input (even when the screen
goes dark by itself as described) and turns amber when there's no

3. The LCD panel syncs and displays correctly even when the
screen goes dark by itself. I verified this with an improvised
backlight. It is only the backlight that is misbehaving.

4. The light/dark cycle is caused by the 3.3V ENA signal from the
control board to the inverter coming on and off. It's too regular
to be a thermal issue.

5. All power supply voltages (17V, 5V & 3.3V) are OK. The control
board controls the inverter with *one-way* ENA and DIM signals.
There is no feedback from the inverter to the control board.

6. The control board doesn't have obvious cracks, burn marks or
bulging capacitors.

Any help will be greatly appreciated.

If we ignore the shipping damage, it sounds like a typical back light problem. Are you sure you have the correct CCFL tubes in the display?

There's a chance one or more of the inverter transformers was damaged when the display was run with an open tube.

What you can do is place your scope probe on the plastic part of the lamp connector. Where it goes exactly is unimportant as you're going to be looking for *differences* in the waveforms between the original tubes and the replacements. In any case, don't make any electrical connection or you'll likely damage your scope input. But you must be sure to place the probe in the exact same spot on the connector as the waveform will vary greatly with just a small physical movement.

Look at the waveforms for each of the tubes during the time it's actually lit. If one or more waveforms differs from any other, the controller IC will shut down the inverter. Too large a waveform usually indicates a bad, weak, or incorrect CCFL tube. A low or distorted waveform will usually indicate a bad transformer.

Before we go further with your suggestions, please see my point
No.5 again. There is no way the logic board can know about an
inverter problem. It just sends and stops the ENABLE signal as
described without any feedback from the inverter.

The power supply and the inverter are on one pcb while the logic
and display circuits are on another board. The power supply
remains fully ON as long as it's plugged in. The only connections
between the two boards are via a 4-way cable:
1) +5V to the logic board
2) GND
2) ENABLE /from/ the logic board
3) DIM PWM signal /from/ the logic board

The inverter does not turn off by itself. It's turned off when
the ENA signal from the logic board stops.

yes, so overriding the ENA signal will not affect safety. What it would affect is the screen shutting off when pc video card stops sending signal, one would need to use the on/off switch more.

I know. I'd considered that possibility but I hate band-aid
solutions except as a last resort.

It seems clear to me that the problem is on the control board. I
was hoping that someone would know about it and can offer a solution.

I've drawn a block diagram of the boards. Here -

It's not a fault I've ever fixed. I've no clue whether you can trace that line backward to see where it starts to get flakey. I'm guessing it just goes straight into a huge chip though. And if Terry is correct that might not be the solution.