0
\$\begingroup\$

I am currently working on a prototype connecting a SOC-board (Samsung ARTIK530) to a LVDS LCD display (DT070BTFT-PTS1).

The display shows a very (!) distorted picture.

I have tried fixing it using software configuration in the Linux driver with no result. The SOC comes with a predefined LCD configuration which exactly matches the timings of my LCD, so I am rather confident that the software side is ok.

The prototype connects the LVDS output of the SOC and the LCD-display using free-floating enameled copper wire.

My questions:

  1. Can a LVDS generally produce a "rather" correct image using this kind of wiring? Or is a LVDS twisted-pair connection mandatory?
  2. I have tried to visualize the LVDS signals using an oscilloscope, with no success. How can LVDS signals be best measured using an oscilloscope?

UPDATE: I narrowed it down to a hardware/soldering issue. As it showed, the LCD is able to display a correct image even with the free-floating LVDS wiring.

\$\endgroup\$
  • \$\begingroup\$ More info is needed. Do you use 6-bit or 8-bit interface (the amount of data pairs is different)? What is the timing you are using for the display? Have you verified all differential pairs were connected to correct pins with correct polarity? ( I once had clock polarity swapped by accident and there was a horribly distorted picture seen as blue data was understood as sync.) \$\endgroup\$ – Justme Nov 19 '18 at 19:57
0
\$\begingroup\$

Can a LVDS generally produce a "rather" correct image using this kind of wiring? Or is a LVDS twisted-pair connection mandatory?

I don't know your pixel clock, but

The prototype connects the LVDS […] using free-floating enameled copper wire.

sounds bad.

The whole idea of LVDS is that it's differential and hence the routing can take advantage of the electrical field wave traveling between the conductors of a pair.

High-speed LVDS interfaces define a transmission line wave impedance (typical values are around 75 to 100 Ω), and your transmission line has to fulfill that. Your wires certainly don't.

Even without that consideration: Pixel clocks are typically in the high MHz range. You wouldn't ever want to do that with free-floating wire, because that becomes an antenna / filter / otherwise completely undesirable.

I have tried to visualize the LVDS signals using an oscilloscope, with no success. How can LVDS signals be best measured using an oscilloscope?

Rather easily? Simply use two probes and use your scope's ability to calculate the signal difference on the fly, or just use the two probes and do that yourself visually. You should be able with a bit of practice to get a useful eye diagram. Of course, not on the far end of free-flying wire...

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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