ChiMei M270H1 T-con repair

[tom66]

I have this 27" FHD LED-LCD screen from an Acer LED-LCD TV to repair. My second LED TV. Let's hope it's not another bad panel . Fault is the picture is jumpy, unstable, etc. with random (but non static) lines. Some slight (<0.5sec) image retention. No black bars or anything, just unstable picture. I don't see any shorts to ground on the VGL, either. About 20kohm one way, 40kohm the other.

I found some burn damage to the t-con on the part which generates VGL (-5V to -10V normally, I believe it to be approx -7.5V on this panel.) The t-con is permanently mounted to the panel, so can't be replaced unfortunately.

I traced out a part of the circuit, as best I could.

Anyone with a cracked screen panel like this who could read off voltages?

[tom66]

Re: ChiMei M270H1 T-con repair

Picture of burnt diode.

[tom66]

Re: ChiMei M270H1 T-con repair

I tried this ugly bodge... and...

It lives! Those are 1N4148 diodes replacing the dual series diode that I reverse engineered (the diode was totally destroyed and unreadable, but I figure the charge pump circuit couldn't work any other way.)

I am quite surprised myself, was half expecting panel to be bad!

LED TVs & monitors: 2-1 to me.

BTW, this ACER gets a repairability score of negative 5 out of 10. It is so awfully put together. The main board and power board are mounted to the back plastic and the wires are so short, there's barely any space for testing or unplugging cables.

[mmartell]

Re: ChiMei M270H1 T-con repair

geez nice save, tom. How DO those three legged diodes work anyway ?

[tom66]

Re: ChiMei M270H1 T-con repair

Basically two diodes in one. They come in several configurations. See below.

In my case, it was the "BAR43S" type, which is the only type which could work in this application, though I'm not certain it was originally a BAR43S, it would still work. (Dropped voltages would be ~0.3V each with a BAR-type Schottky diode; I used silicon, but as can be seen below, the difference would not matter, as the -14.6V is regulated immediately.)

It's a negative charge pump. It converts a positive AC square wave into negative DC. It has several states.

• At t=0, let's say switch node is at 16V. Cap C18 charges 16V thru D2 3-2. (Minus ~0.7V drop from diode.)
• Then lets say after some finite time t that switch node drops to 0V.
• Cap C18 still has 16V across it (because it can't discharge due to the diodes), but one end is at 0V, therefore other end must be at -16V.
• The diode D2 3-1 transfers this charge on to C20. (Hence name "charge pump". Charge is pumped between two capacitors to create the desired effect.) Minus another 0.7V drop.
• Total output voltage after many cycles = Vsw - 1.4V = -14.6V assuming minimal load.

This -14.6V is then regulated by the transistor-zener circuit to around -7.5V.

[mmartell]

Re: ChiMei M270H1 T-con repair

I've always suspected but now I KNOW I'm a little like that guy from A Beautiful Mind because when you talk like that the logic makes me want to cry. Lol I'm just a little too ignorant as of yet to fully understand it.

Seriously, how did you know to expect a positive square sine wave at that point or did you recognize a charge pump configuration first and deduce the latter ?

[tom66]

Re: ChiMei M270H1 T-con repair

I could recognise the topology of a boost converter (that produces that square wave), plus basically every LCD panel out there uses the same standard design.

The boost converter produces the 16V for the AVDD, which the LCD panel uses about 1 amp from. The other rails, VGL and VGH, require approximately 20mA each. Since their current draw is so minimal, they lend themselves well to a charge pump design (the cost of two caps and a dual diode: <10 cents in bulk volumes. Cost of a dedicated switching converter: ~$1)