Screen Tablet malarky => Monitors => Topic started by: LTN121t on September 20, 2016, 08:50:07 AM

Title: Assorted selection of problems with a universal LVDS controller
Post by: LTN121t on September 20, 2016, 08:50:07 AM
So I've got this random LED LCD and a universal LCD controller I bought from eBay (constraints brought on by cost and location, and I didn't hear of NJYtouch). I didn't think I'd use the connector supplied so I sacrificed it and tried to wire the existing cable to the LCD's LVDS 40-pin cable after consulting the datasheet. Initially I thought this to be merely a simple exercise in counting and stripping fine wire, and went up to connecting one VCC and one GND wire.
I stuck one end of a continuity tester to the board and probed the pins with the other. No result.
But. There was continuity when the probe touched the outer metal body of the connector and this was for both.
I had an important LCD from a tablet laptop, so this LCD is pretty much sacrificial.
The LCD controller is a MT6820-B, one of those universal types common on eBay, with a square PCB and a small square chip.
Also, some users mentioned a place to connect the LED backlight power that is separate on the board to the main LVDS headers. I think I've found it (5V/5V/ADJ/BL/GND/etc?) but for my panel that voltage is too low. Also 5V and GND seem to be inputs, (and maybe the other pins) as I followed some pictures I found on the internet and wired a USB connector to the underside on those points and plugged the whole thing into a USB charger, and it worked - LED on button board lit up, and responds to button presses, and temperature of chip slightly increases.) This is very confusing, as many people and even some additional cables that I bought (that had the LVDS cable branching out into the main LVDS board connector and an extra small white connector), seem to suggest that the (5V/5V/ADJ/BL/GND/I forgot what the other one was) socket on the board was meant to be an output for backlight power yet I'm using it to power the board itself.

Anyway, if all this confuses you feel free to ask for clarification (I'm typing this very hurriedly as I'm outside)

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Ertew on September 27, 2016, 10:59:36 PM
You're right. This board doesn't have dedicated supply input.

By design, Your board should be connected to another board that contains: mains input, 5V supply for your LVDS board and CCFL driver (PWM signal comes from Your board). Without that, You need to create similar circuit.

Please provide more information about Your LCD. I'm guessing that LED driver are inside LCD panel and You have only single connector with 3.3V input for LCD, 8-30V input for LED and LVDS input. In that case You need to have 5V supply for LVDS board (may be stolen from USB) and 12V supply for backlight circuit. LVDS board will create 3.3V for LCD.
So LCD cable need to be split: 3.3V, some GND and LVDS to LVDS output. Backlight controll (enable, PWM) to corresponding pins from [5V/5V/ADJ/BL/GND] connector. Supply for board to other pins from [5V/5V/ADJ/BL/GND] connector. 12V connected to LCD cable - 12V-backlight and GND-backlight.

BTW, I prefer this type of kits:
It have onboard DC/DC converter with 8-20V input and 5V output. Input voltage are connected to 12V pin for backlight, while LCD voltage at LVDS port can be chosen by jumper: 12V, 5V or 3.3V.

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Aerendraca on November 02, 2016, 03:55:10 PM
I actually have one of these boards and I know you're frustration.

For clarification this is the board:

The red socket is meant to be used with backlights of small screen that can run from a 5V supply, typically this is probably a screen size up to 10". The problem here is that your inverter/LED driver (depending on which you have) is likely to require 12V so what to do?

Well here's what I did. As i mentioned in a post I replied to today, I used an LM2596 step down module, a cheap CCFL tester and a 12V power supply.

The cheap CCFL Inverter tester I used:

I ran the 12V straight from the power supply to the tester, and from the tester to - in my case - an LED driver board. I then tapped of the 12V from the power supply to the LM2596 and stepped it down to 5V to driver the MT6820-B, and connected the output of the LM2596 to 5V and GND on the red connector. Problem with this setup is that you have 2 separate circuits, one for driving the LCD and the other for driving the backlight, they are not linked. This means that you manually have to switch the backlight on to see the picture.

03.11.2016 - The following part regarding how I got the backlight to work with the LVDS controller has been shown to me to be, let's say, a little bit pointless. I've kept it in this post for reference, and you can read why it's pointless in the next post by Ertew. All I can say is, we all make mistakes. :)

The way I got around this problem was with a simple additional circuit. The additional circuit was comprised of an IRL540N HEXFET and a resistor, and works by utilizing the BL pin of the LVDS controller.

Here's the circuit:

When the LVDS controller is power on, the BL pin goes HIGH pulling the 'Gate' pin of the IRL540N high, this effectively turns the IRL540N 'ON' allowing current to flow from the 'Drain' to the 'Source'. This is like physically flipping a switch from 12V to the CCFL inverter tester. When the LVDS controller is turned 'OFF' or put in standby, the BL pin goes low and turns off the CCFL tester.

The resistor value should be fairly high, I think I used a 57Kohm as I had one laying around, but any value above 10Kohm will probably be fine. The resistor is just there to hold the 'Gate' of the IRL540N 'LOW' when the LVDS controller is turned off.

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Aerendraca on November 02, 2016, 04:10:35 PM
This is what my setup looks like:

A quick note on the first picture. I've removed the switch and power socket and used jumpers to bridge the gaps as I didn't need these components. This CCFL inverter tester also has a slightly different layout to the one pictured in my previous post but they all work the same.



Additional components that can be seen are an Atmel ATTiny85 (used to display the status of the LVDS controller via an RGB LED), and an Adafruit AR1100 module (used to get the resistive touch layer to communicate with the pc via USB).

This was a 12.1" point of sale screen which I've hacked to be a small monitor for a small computer that I've built to run a CNC machine that I've almost finished building; was trying to keep the overall footprint of the machine small.

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Ertew on November 02, 2016, 09:21:02 PM
You hacked a nice screen.
But why You have chosen this strange way? IMO You can throw away IRL540N MOSFET and connect everything like i described bellow:

1. Connect all grounds together.
2. Connect +12V to input of LM2596 DC/DC module and to input of CCFL inverter. (in typical notebook, CCFL inverter is hooked directly to battery or power input)
3. 5V output from DC/DC to LVDS converter.
4. CCFL inverter can be turned on and off via enable pin (often named EN, ON, POW, BL) - connect it to BL output at LVDS converter.
5. Last thing, brightness. It can be connected to 5V, 0V, pot (as You have on Bl tester) or to brightness (ADJ) output on LVDS converter (if supported by firmware).

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Aerendraca on November 03, 2016, 09:30:12 AM
Haha... :D

Actually that's a very fair point. I did this a few months ago now so I don't remember my reasoning, but I imagine it was just a matter of not seeing the wood for the trees. You're perfectly right that both the resistor and HEXFET aren't really needed. I probably just had the parts laying around and didn't think too much about it, or possibly had intentions to use the BL signal to do something else.

Ertew, thanks. I think I'm going to rewire my setup as per your suggestion since it has the advantage of allowing the backlight to be controlled from the OSD. Nice!

There's an outcome that I hadn't expected when I posted yesterday!

Cheers.  ;)

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Aerendraca on November 03, 2016, 02:32:14 PM
Spot the difference. I'll give you a clue:

The inverter tested has been stripped virtually bare so just the connector, a resistor, and an LED are present. This was so that I could fully disconnect the backlight if I needed to strip the screen down again. Using the connector on the LVDS controller wouldn't allow this due to being connected to 5V power.

Tested and working, check out the refresh rate and resolution of this bad boy! - ahh, and there's the homeless little HEXFET sitting on the bezel:

I left the resistor on the protoboard connected to nothing as I have thousands of them.

2 less components and the OSD now controls brightness.

Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Aerendraca on November 04, 2016, 10:44:12 AM
Hmmmm........ OK, so this could just be coincidence however, today the LED backlight current board burnt out with a huge stink of smoke.  :'(

I suspect that the board - which was from China and very cheap - is at fault rather than the rewiring discussed previously, however I will look into this a bit more and make sure that nothing else got taken out during the event.

Here a pic of the burnt out resistor:


Originally this was an SMD resistor of the same value (of which I forget now), however the original resistor cracked in half when I attempted to pull the board from some double-sided tape. Obviously too much current has passed through the resistor but as yet I have no idea why. I'm not even sure if the LEDs are still working.


Title: Re: Assorted selection of problems with a universal LVDS controller
Post by: Aerendraca on November 07, 2016, 03:13:59 PM
The enable output from the MT6820-B is 5V, I believe the LED driver board requires an enable of 3V maximum. It worked well for a short while until it failed causing 10x more current to pass through the resistor and burn it out. I measured a max of 120mA going through the 100Ohm resistor before I realized that the ME15N10 FET had failed. I'm not completely sure what happened to the ME15N10 transistor as it should've been able to handle +/- 20V at the gate, however it did indeed fail.

The reason I suspect I originally used the IRL540N and inverter tester was so that I had a better control of the enable and dim voltages in case of compatibilty issues. It is still true that the IRL540N was a bit redundant, as the enable switch as I could have utilized the variable resistors on the inverter tester board to adjust the enable voltage. This would have meant hacking the inverter tester a little, which at the time seemed a pain in the backside, so I opted to use the HEXFET to control enabling the inveter tester instead.

Obviously now, hacking the inverter tester seems trivial compared to the problems I caused myself. Doh!


Make sure your parts are compatible; The screen I have was originally CCFL and I converted it using a kit from China - I should have checked the board over properly.

Alternatively, make sure you know what voltages are expected at the LED Driver/CCFL inverter end before you go ahead and plug it in to your LVDS controller; you shouldn't need to worry about this if you purchased the LVDS controller and Inverter/Driver as a kit, the seller will have already matched them.

Here's the modification made to get the board running again - Guess who's back!

I will need to remove the ME15N10 properly and try and mount the resistor a little better, but it's all working fine now, and the IRL540N requires


This worked OK for a bit until the Vcc pin of the LED driver chip came off. Nevermind, I'll just have to order another and learn from my mistakes.

Also, Ertews method is still sound, it's just that you need to make sure that you do not over-volt the driver/inverter somehow before proceeding.