This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Parallel Connections of LVDS drivers

Other Parts Discussed in Thread: SN55LVDS31

A question was posed by Alexander Osovets "Can I connect two or four 55LVDS31 outputs together in parallel to open the eye and boost the speed?"

It was answered by Diego Cortes "I am afraid that this is not a recommended option."

Please expand on the answer provided but before doing so, please consider the following background informaion. 

Background:

 TI App note "slaa100" presents examples, test data and a conclussion which indicates that connecting LVDS drivers in parallel directly, specifically the LVDS31, is indeed a valid configuration for improving current drive capability.. 

TI App Note SLLD009, pages 3-7 through 3-9 provide similar guidance for increasing the signals levels and edge rates, although in the examples of SLLD009, the drivers are capacitively connected in parallel.

TI App Note SLLA120 page 11, indicates that the LVDSoutput stages are current driven with current sourced drivers which would indicate that overload should current is limited. The data Seet for SN55LVDS31 (SLLS261M), Section 9.3.4 states that "the SNx5LVDSxx output stage acts as a constant current source",  Hence should not be a concern for device damage from parallel connections.

I would like to pose the same Question as Alex in a slightly different way. Given the App notes listed, What is the tehnical reason of why connecting LVDS Drivers in parallel with their inputs and putputs directly connected, is this not a valid implementation approach to improving the eye pattern or  driving a longer cable?  Will the parts suffer degredation over time or will the signal itself be degraded?.  Is this approach actually frowned upon and in against recommendations, or is only atypical? 

Thanks

Lynn

 

  • Hello Lynn,

    I have to admit that my previous answer was wrong. The reasoning behind my previous answer was because the parallel configuration decreases the output impedance of the driver and over load the receiver as it could happen according to the Figure 3–11 of the LVDS Application and Data Handbook (slld009, which I had no knowledge of its existence), among the skew considerations. so I only have to appreciate your concern and collaboration on the forum.

    Best regards,
    Diego.