1
\$\begingroup\$

I found a universal LCD controller in my broken 32" Telefunken TV.

I thought of trying to use it on my broken 15.6" laptop screen.

The interface is LVDS (1 ch, 6-bit), 30-pin connector.

enter image description here

The controlboard is a TP.SK108.PB818

I found the below schematic for a similar model from the same range, so I assume the headers will have the same pinouts. All the VCC and grounds match up to the schematic.

So I rewired the LVDS cable to match; controlboard on the left and LCD connector on the right.

Panel VCC was at 12 V so I rewired the VCC pins through a buck step-down set at 3.3 V and made a plan to turn on the CCFL inverter. I added a 1500 μF capacitor at the output side.

The TV board logo comes up (fits in the screen) and I can see the blue screen and access the menu, but the quality looks off. Then I plugged my laptop HDMI into the input. The image is really bad.

TP.SK108.PB818

This is how I rewired the LVDS cable. Not shown - all grounds connected.

The firmware for the controlboard was already 1366x768 and the LCD is 1366x768 so I did not mess with the firmware.

Looking at the LVDS cable that came from the laptop I noticed the only difference in pins used was 4, 6 and 7 on the LCD side.

I rechecked all wires from the header to the LCD panel for continuity and every pin makes contact to the designated contact.

This is what the screen looks like. The grey scale test is completely the wrong color.

Do you have any advice as to what I am missing here? Except for the fact that the board sellers state it's for 27-32" panels. I can see my desktop, so it sort of works.

enter image description here

enter image description here

enter image description here

enter image description here

\$\endgroup\$
2
  • \$\begingroup\$ When you say the quality is bad, is the image constantly changing/jumping, or is it the same bad image on every frame? There is a possibility that 0/1/2/3 and +/- may be inconsistent in the datasheets or miswired. \$\endgroup\$
    – nanofarad
    Nov 11, 2022 at 16:19
  • \$\begingroup\$ Even if the data lanes are correctly mapped and clock has correct polarity (they look like they do especially if image is stable), there are two standards how to map the bits to lanes. Do you know if the devices use JEIDA or VESA mapping? \$\endgroup\$
    – Justme
    Nov 11, 2022 at 17:28

1 Answer 1

Reset to default
2
\$\begingroup\$

Since each colour slide appears in 4 narrow blocks when it is supposed to be a one long colour slide, the problem is the bit mapping.

It seems your original panel had 4 data lanes and the driver is sending in "vesa-24" format, where last data lane 3 carries the two most significant bits of each of the R, G and B data. Least significant 6 bits of each is sent on the 3 first data lanes.

The laptop panel has just 3 data lanes and it expects the signal in "jeida-18" format, which is compatible with "jeida-24" format. The most significant 6 bits are transferred in the first three data lanes, and the last data lane carries the remaining 2 least significant bits, so when transmitted with 4 data lanes, it is compatible with both 3 and 4 lane displays.

So, as the controller PCB is sending in "vesa-24" format, the two MSBs are missing and you get the data modulo 4, so the least significant 6 bits of 8.

This is why the colour slides appear 4 times.

There might be a way to configure the PCB for different panels, but it might require moving a jumper or soldering a strap, modifying firmware or just going into service menu to change that. Details how to do that are unknown.

Because the colour slides have uneven length the PCB seems to be calibrated for getting the colour balance right on the original panel. More reverse engineering or hacking required to be able to set the colour balance/gains/offsets/gamma/etc to have good looking colour slides.

Edit: Actually opened the linked panel data sheet. Yes, it uses the "vesa-24" format where the last data lane D carries the MSBs. Therefore, no panel with 3 data lanes are compatible with data format the PCB currently sends out.

\$\endgroup\$
2
  • \$\begingroup\$ Justme, thank you for that well laid out answer. It made the stuff I was reading the last couple of days way more clear. The controlboard LVDS cable does have 1 pair of twisted cable not connected. That's likely the 4th lane. I went to search further and found the code "Menu 1147" to get to the service menu of the board. Looking at the Panel info section I noticed the mode is 8bit, 4 channels. It does not have options to change those settings though. Thank you. I will use the board with my 27" HD panel. Firmware is available for 1080p \$\endgroup\$
    – Pieter Bez
    Nov 12, 2022 at 13:24
  • \$\begingroup\$ Turns out searching for the control chip instead, yields way better results. TSUMV53. With the correct firmware and LVDS MAP setting on the screen I will be able to run my laptop screen afterall. \$\endgroup\$
    – Pieter Bez
    Nov 12, 2022 at 15:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.