Why did grounding the outer LVDS cable shielding make the screen RED ?

[MathWizard]

I'm trying to make this short. I have a Qnix 27" LCD monitor, and as far as I remember, the HDMI worked. But it doesn't now, I get no video, from different videocards/etc. I don't see a no signal message either.

It runs off 12-24V DC in, I used a power brick with good caps (checked just before all this). All the DC-DC converters and regulators are working, and so is the LED backlight. There's only 1 main IC, and some ram chips on this PCB, the only other stuff in the monitor is on the long skinny PCB attached to the top of the panel, T-con stuff I guess.

The HDMI cable works with a newer, faster monitor, so I think it's fine.

Now I'm just trying to find out if the HDMI is getting to the chip , and then to the chips on the panel. The MCU only tries to start the backlight for 100us, then decides against it, and never tries again. I can jumper it on tho, it works. The MCU gets on info back on the state of the backlight either, so thats fine.



Anyways I had the whole thing taken apart, with the panel face down on a table, and the PCB sitting on a piece of wood, on top the panel.

There are LVDS master and slave cables, each with a ground shield that screws to the common grnd on the PCB.

I was scoping the LED backlight PWM chip, for the boost voltage, and had a lot of 60Hz hum , on scope for some reason. And I decided to clip both gnd shields together, and clip them to PCB ground. At which point the whole screen went red. I could only see the red glow from through the cracks in the back of the panel, so I though I had burned them all out and that's why they were red.

Also, 1 of the DC-DC buck converters, or the 24V power brick, has gotten a bit screwed up, because sometimes I swear I was losing the input voltage to the LED backlight section. The PWM enable pin doesn't always work right away either.

But still now, with it unplugged a few times, if I turn it on and enable the backlight, the whole screen is stuck red.

So what do u suppose happened, and did some CMOS get zapped, and made the whole panel stick on RED ? I'm about to play with it a bit more now.



I just found the HDMI input SDA and SCL lines from the MCU/video chip. But I'm in way over my head on all that stuff.

[MathWizard]

The above new issue aside, I forgot that it also doesn't seem to turn off/back to stby anymore. As far as I remember, this was my daily monitor, it's an IPS-AHVA-LED 1440p 60Hz. And I only stopped using it when I got a much faster monitor.

I have the datasheet for the panel. But the IC/MCU name, that does it all, is glued under a heatsink for now.
https://datasheetspdf.com/pdf-file/847631/AUO/M270DAN01.1/1

So it power's up ok, but there's no display at all, no logo, anything, connected or not, but the light works, when enabled.

I soldered some wires to the hotplug and HDMI I2C SCL/SDA, which are attached to an eeprom, the common 24C0xx type. The MCU is wired to control those pins too.

So far I've only seen the MCU take the hotplug low for a moment, I'll have to see if that's right.

Using I2C decode on my scope, I can see my GPU send out it's polling signal, it writes 3Ah, and when the monitor is powered on (without the panel attached tho), the monitor will read 3Ah, then it sends data, but it only sends 5 groups of 00h. Then the GPU writes 3Ah, then the monitor reads 3Ah.

I'm guessing that blank data should be the EDID

So that doesn't sound right. IDK if that block of zero's is in the eeprom at 3A, or if the MCU is doing that. I'm new to eeproms, and I2C, but that's what the scope says it's doing.

So is it looking at the wrong address and finding zero's or is the ram broken? Or the MCU is screwed up ? Agaiun it doesn't seem to turn off anymore. That happened ages ago, but I have no memory of it, as far as I knew, it worked. The pwr button is just on a R2R ladder, and it looks normal.

So far I'm not sure it sends much else, or if it keeps trying the same thing, the polling is about 0.5s apart, and single shot triggering on it and getting a clean signal to decode at 100kHz or so...is proving hard. IDK if I can add some delay when in this mode.

[mclute0]

Normally one of two issues, IMHO...

The LED screens hardly ever go bad, but the control boards do and other parts do.

One is, if to many backlight Leds burn out the led circuit will stop lighting any lcds. In most Smart TV these days that means three or more. The test is to shine a bright flashlight into the screen and see if you can see anything on the screen from teh reflection off the back panel.


Two is the problem I see the most of in anything connected to a computer. I hate to tell you that the majority HDMI chips have little protection, and they can easily be damaged when plugging a live connection in. It is a good practice to turn off or sleep a device before plugging it in, but even just ESD over a period of years can damage the chips.. I have had to replace more 4k Television boards with bad HDMI circuits from this. The best way to test is to use a different port on the device like a VGA or Composite, DVI, or Display port. That really depends on the design of the chipset.  On televisions this is easy to test since the OTA signal does not use the HDMI circuit. Some screens will show a blue screen with the backlight on, but some will go into low power mode without a signal and just have a dark screen with not backlight.

I buy Led backlight strips and control boards from shopjimmy.com, but there are plenty of other sources for parts.

[MathWizard]

Yeah the backlight does work.

I'm going over how I2C and eeprom's work. I pulled it off and put it on a breakout board to talk to with BusPirate
https://www.st.com/resource/en/datasheet/m24c04-f.pdf

On the monitor PCB, the 24C04 was wired with p1,2,3, and p7 , all wired to GND . I'm not sure what brand this SMD version is. If it's some verions, those are it's write-protect-enable, 2 address pins, and maybe it's it's multibyte-mode or pagemode pin. Some makers versions, p1 is n.c.

So the I2C device select code should be 1010000

When I scoped all this the other day, my 6700xt is polling with the device select code 0111010

Last time I remember using this monitor full time, I swear it was over HDMI, and on a rtx2070  that I still have, but it doesn't work with it now either. I really better check what address it put's out too.

So is the GPU addressing the MCU in the monitor, ? I'm just getting into the EDID stuff, the manuals are huge, and google only finds marketing speak articles, not repair pages anymore.


Or is the eeprom wired up to an older standard, and the GPU is looking for something newer (but i swear i used it w/hdmi tho)

I think the hotplug going low for a moment, then back high, is the way it should be. Again I should compare this to a working monitor, but that's my good 1 under warranty, or an old VGA monitor, so that's no help.