# Why this LVDS transceiver doesn't specify the differential input polarity?

I've been looking to implement a LVDS transceiver in my project where i can hook up to Displayport AUX differential pair.

I've discovered a suitable part, SN65MLVD200A in the following design guideline document (although many alternatives in the TI SN65 series can and might be better to use)

https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/an/an_745.pdf

https://www.ti.com/lit/ds/symlink/sn65mlvd204a.pdf

My question is why there is no polarity specified in the differential inputs marked A and B in the datasheet.

Considering that polarity doesn't matter because of the symmetrical internal logic of the transceiver, wouldn't this mean that the outputs marked D and R in the datasheet will swap when the differential input polarity is swapped?

Kind regards

## 1 Answer

The polarity is specified in a roundabout way. Note that A has no input "bubble" but B has one. That tells you that for differential inputs, A is the non inverting input and B is inverting. This can also be confirmed in figure 22. by looking at the input and output waveforms. Figure 22 is a timing diagram, but it shows that Vid is Va-Vb and it shows that the output will be positive going for positive Vid.

• Oh, wow, there is a truth table, too. Look at table 3 and table 4. It shows right there that Vid is Va-Vb, and exactly what state the output will be in in all cases. Commented Mar 1, 2021 at 9:21
• After i thought about it more before i viewed your answer, it seems A to be the + and B to the - if the output needs to match the input (aka non-inverted); but you can always invert it in software if it allows for better/easier routing of the traces
– TnF
Commented Mar 1, 2021 at 9:36